mirror/obs-studio
0
0
Fork 0
mirror of https://github.com/obsproject/obs-studio.git synced 2024-09-20 13:08:50 +02:00
Code
f2d4de3c03
obs-studio/libobs/obs-source.c

1129 lines
29 KiB
C
Raw Normal View History

first commit
2013-10-01 04:37:13 +02:00
/******************************************************************************
Copyright (C) 2013 by Hugh Bailey <obs.jim@gmail.com>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
With the permission of my fellow contributors, I'm switching obs-studio back to GPL v2+ to prevent issues between this project and the original OBS project, and for personal reasons to avoid legal ambiguity (not political reasons, I admittedly would prefer GPL v3+)
2013-12-03 06:24:38 +01:00
the Free Software Foundation, either version 2 of the License, or
first commit
2013-10-01 04:37:13 +02:00
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
******************************************************************************/
Fix a some more linux/GCC specific warnings
2014-02-14 23:56:01 +01:00
#include <inttypes.h>
not sure how a backslash slipped into an include, but it's gone now
2013-10-30 07:54:43 +01:00
#include "media-io/format-conversion.h"
Implement automatic video scaling (if requested) Add a scaler interface (defaults to swscale), and if a separate output wants to use a different scale or format than the default output format, allow a scaler instance to be created automatically for that output, which will then receive the new scaled output.
2014-02-18 21:37:56 +01:00
#include "media-io/video-frame.h"
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
#include "util/platform.h"
make sure to include function declarations from calldata.h
2013-12-30 18:09:20 +01:00
#include "callback/calldata.h"
changed gs_draw_sprite to allow custom sizes, added output textures to the core, and adjusted the test code to accommodate the changes
2013-11-27 06:26:14 +01:00
#include "graphics/matrix3.h"
#include "graphics/vec3.h"
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
first commit
2013-10-01 04:37:13 +02:00
#include "obs.h"
Rename obs-data.h to obs-internal.h Renaming obs-data.h to avoid confusion about its usage
2014-01-27 02:48:14 +01:00
#include "obs-internal.h"
first commit
2013-10-01 04:37:13 +02:00
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
static void obs_source_destroy(obs_source_t source);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
static inline const struct obs_source_info *find_source(struct darray *list,
add names to sources, use 'id' for the internal source identifier names rather than 'name'
2013-12-21 01:23:19 +01:00
const char *id)
first commit
2013-10-01 04:37:13 +02:00
{
size_t i;
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
struct obs_source_info *array = list->array;
first commit
2013-10-01 04:37:13 +02:00
for (i = 0; i < list->num; i++) {
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
struct obs_source_info *info = array+i;
add names to sources, use 'id' for the internal source identifier names rather than 'name'
2013-12-21 01:23:19 +01:00
if (strcmp(info->id, id) == 0)
first commit
2013-10-01 04:37:13 +02:00
return info;
}
return NULL;
}
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
static const struct obs_source_info *get_source_info(enum obs_source_type type,
update API and implement preliminary ability to add sources to scenes
2013-12-30 14:56:39 +01:00
const char *id)
{
struct darray *list = NULL;
switch (type) {
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
case OBS_SOURCE_TYPE_INPUT:
list = &obs->input_types.da;
break;
case OBS_SOURCE_TYPE_FILTER:
list = &obs->filter_types.da;
break;
case OBS_SOURCE_TYPE_TRANSITION:
list = &obs->transition_types.da;
break;
case OBS_SOURCE_TYPE_SCENE:
update API and implement preliminary ability to add sources to scenes
2013-12-30 14:56:39 +01:00
default:
blog(LOG_WARNING, "get_source_info: invalid source type");
return NULL;
}
return find_source(list, id);
}
Added add/remove signals to scenes Scenes will now signal via their source when an item has been added or removed from them. "add" - Item added to the scene. Parameters: "scene": Scene that the item was added to. "item": Item that was added. "remove" - Item removed from the scene. Parameters: "scene": Scene that the item was removed from. "item": Item that was removed.
2014-01-04 21:38:56 +01:00
bool obs_source_init_handlers(struct obs_source *source)
implement signal/procedure handling into libobs and individual sources
2013-12-27 07:10:15 +01:00
{
source->signals = signal_handler_create();
if (!source->signals)
return false;
source->procs = proc_handler_create();
return (source->procs != NULL);
}
update API and implement preliminary ability to add sources to scenes
2013-12-30 14:56:39 +01:00
const char *obs_source_getdisplayname(enum obs_source_type type,
const char *id, const char *locale)
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
const struct obs_source_info *info = get_source_info(type, id);
update API and implement preliminary ability to add sources to scenes
2013-12-30 14:56:39 +01:00
return (info != NULL) ? info->getname(locale) : NULL;
}
add support for texture_rect texture type in shaders, add new form, clear up a few things in the API
2013-12-20 19:56:01 +01:00
/* internal initialization */
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
bool obs_source_init(struct obs_source *source,
const struct obs_source_info *info)
first commit
2013-10-01 04:37:13 +02:00
{
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
source->refs = 1;
Implement volumes for sources - Added a volume variable to the obs_source structure and implemented functions for manipulating source volume. - Added a volume variable to the audio_data structure so that the volume will be applied when mixing.
2014-01-07 19:03:15 +01:00
source->volume = 1.0f;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_init_value(&source->filter_mutex);
pthread_mutex_init_value(&source->video_mutex);
pthread_mutex_init_value(&source->audio_mutex);
Add a couple more setting data safety measures Prevent null dereferencing if data is null, instead just make the functions return and ignore the get/set requests.
2014-01-28 23:45:30 +01:00
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
memcpy(&source->info, info, sizeof(struct obs_source_info));
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (pthread_mutex_init(&source->filter_mutex, NULL) != 0)
return false;
if (pthread_mutex_init(&source->audio_mutex, NULL) != 0)
return false;
if (pthread_mutex_init(&source->video_mutex, NULL) != 0)
return false;
first commit
2013-10-01 04:37:13 +02:00
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (info->output_flags & OBS_SOURCE_AUDIO) {
Add preliminary audio processing - Mixing still isn't implemented, but the audio system should be able to start up, and mix at least once audio line for the time being. Will have to write some test audio sources to verify things are working properly, and build the rest of the output functionality.
2014-01-10 03:08:20 +01:00
source->audio_line = audio_output_createline(obs->audio.audio,
source->name);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (!source->audio_line) {
blog(LOG_ERROR, "Failed to create audio line for "
Add preliminary audio processing - Mixing still isn't implemented, but the audio system should be able to start up, and mix at least once audio line for the time being. Will have to write some test audio sources to verify things are working properly, and build the rest of the output functionality.
2014-01-10 03:08:20 +01:00
"source '%s'", source->name);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
return false;
}
}
make data access in the threads a bit more safe (note: probably will need some more safety measures later on)
2013-10-19 05:25:13 +02:00
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
return true;
first commit
2013-10-01 04:37:13 +02:00
}
fix a few bugs with callback system, add some test signals
2013-12-27 13:08:58 +01:00
static inline void obs_source_dosignal(struct obs_source *source,
const char *signal)
{
struct calldata data;
calldata_init(&data);
calldata_setptr(&data, "source", source);
signal_handler_signal(obs->signals, signal, &data);
calldata_free(&data);
}
add names to sources, use 'id' for the internal source identifier names rather than 'name'
2013-12-21 01:23:19 +01:00
obs_source_t obs_source_create(enum obs_source_type type, const char *id,
Implement settings interface for plugins Add a fairly easy to use settings interface that can be passed to plugins, and replaced the old character string system that was being used before. The new data interface allows for an easier method of getting/altering settings for plugins, and is built to be serializable to/from JSON. Also, removed another wxFormBuilder file that was no longer in use.
2014-01-28 07:14:58 +01:00
const char *name, obs_data_t settings)
first commit
2013-10-01 04:37:13 +02:00
{
struct obs_source *source;
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
const struct obs_source_info *info = get_source_info(type, id);
first commit
2013-10-01 04:37:13 +02:00
if (!info) {
add names to sources, use 'id' for the internal source identifier names rather than 'name'
2013-12-21 01:23:19 +01:00
blog(LOG_WARNING, "Source '%s' not found", id);
first commit
2013-10-01 04:37:13 +02:00
return NULL;
}
Use bzalloc instead of bmalloc then memset Reduces needless code repetition and still allows for proper memory alignment. Cleans up the code a bit.
2014-02-09 20:34:07 +01:00
source = bzalloc(sizeof(struct obs_source));
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
implement signal/procedure handling into libobs and individual sources
2013-12-27 07:10:15 +01:00
if (!obs_source_init_handlers(source))
goto fail;
Improve safety for settings usage - Add 'set_default' functions to obs-data.*. These functions ensure that a paramter exists and that the parameter is of a specific type. If not, it will create or overwrite the value with the default setting instead. These functions are meant to be explicitly called before using any of the 'get' functions. The reason why it was designed this way is to encourage defaults to be set in a single place/function. For example, ideal usage is to create one function for your data, "set_my_defaults(obs_data_t data)", set all the default values within that function, and then call that function on create/update, that way all defaults are centralized to a single place. - Ensure that data passed to sources/encoders/outputs/etc is always valid, and not a null value. - While I'm remembering, fix a few defaults of the main program config file data.
2014-01-29 02:41:24 +01:00
source->name = bstrdup(name);
source->type = type;
source->settings = obs_data_newref(settings);
source->data = info->create(source->settings, source);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (!source->data)
goto fail;
Improve safety for settings usage - Add 'set_default' functions to obs-data.*. These functions ensure that a paramter exists and that the parameter is of a specific type. If not, it will create or overwrite the value with the default setting instead. These functions are meant to be explicitly called before using any of the 'get' functions. The reason why it was designed this way is to encourage defaults to be set in a single place/function. For example, ideal usage is to create one function for your data, "set_my_defaults(obs_data_t data)", set all the default values within that function, and then call that function on create/update, that way all defaults are centralized to a single place. - Ensure that data passed to sources/encoders/outputs/etc is always valid, and not a null value. - While I'm remembering, fix a few defaults of the main program config file data.
2014-01-29 02:41:24 +01:00
if (!obs_source_init(source, info))
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
goto fail;
first commit
2013-10-01 04:37:13 +02:00
fixed a typo in a function call
2013-12-27 23:23:42 +01:00
obs_source_dosignal(source, "source-create");
first commit
2013-10-01 04:37:13 +02:00
return source;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
fail:
blog(LOG_ERROR, "obs_source_create failed");
obs_source_destroy(source);
return NULL;
first commit
2013-10-01 04:37:13 +02:00
}
Implement automatic video scaling (if requested) Add a scaler interface (defaults to swscale), and if a separate output wants to use a different scale or format than the default output format, allow a scaler instance to be created automatically for that output, which will then receive the new scaled output.
2014-02-18 21:37:56 +01:00
void source_frame_init(struct source_frame *frame, enum video_format format,
uint32_t width, uint32_t height)
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
{
Implement automatic video scaling (if requested) Add a scaler interface (defaults to swscale), and if a separate output wants to use a different scale or format than the default output format, allow a scaler instance to be created automatically for that output, which will then receive the new scaled output.
2014-02-18 21:37:56 +01:00
struct video_frame vid_frame;
video_frame_init(&vid_frame, format, width, height);
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
frame->format = format;
frame->width = width;
frame->height = height;
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
Implement automatic video scaling (if requested) Add a scaler interface (defaults to swscale), and if a separate output wants to use a different scale or format than the default output format, allow a scaler instance to be created automatically for that output, which will then receive the new scaled output.
2014-02-18 21:37:56 +01:00
for (size_t i = 0; i < MAX_AV_PLANES; i++) {
frame->data[i] = vid_frame.data[i];
frame->linesize[i] = vid_frame.linesize[i];
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
}
}
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
static void obs_source_destroy(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
size_t i;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
fix a few bugs with callback system, add some test signals
2013-12-27 13:08:58 +01:00
obs_source_dosignal(source, "source-destroy");
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
if (source->filter_parent)
obs_source_filter_remove(source->filter_parent, source);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
for (i = 0; i < source->filters.num; i++)
obs_source_release(source->filters.array[i]);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
for (i = 0; i < source->video_frames.num; i++)
source_frame_destroy(source->video_frames.array[i]);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
gs_entercontext(obs->video.graphics);
texture_destroy(source->output_texture);
gs_leavecontext();
first commit
2013-10-01 04:37:13 +02:00
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
if (source->data)
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
source->info.destroy(source->data);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
for (i = 0; i < MAX_AV_PLANES; i++)
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
bfree(source->audio_data.data[i]);
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
audio_line_destroy(source->audio_line);
audio_resampler_destroy(source->resampler);
implement signal/procedure handling into libobs and individual sources
2013-12-27 07:10:15 +01:00
proc_handler_destroy(source->procs);
signal_handler_destroy(source->signals);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
texrender_destroy(source->filter_texrender);
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
da_free(source->video_frames);
da_free(source->filters);
pthread_mutex_destroy(&source->filter_mutex);
pthread_mutex_destroy(&source->audio_mutex);
pthread_mutex_destroy(&source->video_mutex);
Implement settings interface for plugins Add a fairly easy to use settings interface that can be passed to plugins, and replaced the old character string system that was being used before. The new data interface allows for an easier method of getting/altering settings for plugins, and is built to be serializable to/from JSON. Also, removed another wxFormBuilder file that was no longer in use.
2014-01-28 07:14:58 +01:00
obs_data_release(source->settings);
add support for texture_rect texture type in shaders, add new form, clear up a few things in the API
2013-12-20 19:56:01 +01:00
bfree(source->name);
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
bfree(source);
}
Remove external ref count visibility The returned ref count was never used and the returned value doesn't actually confer any meaning in a multithreaded environment
2014-02-03 02:34:06 +01:00
void obs_source_addref(obs_source_t source)
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
{
Remove external ref count visibility The returned ref count was never used and the returned value doesn't actually confer any meaning in a multithreaded environment
2014-02-03 02:34:06 +01:00
if (source)
++source->refs;
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
}
Remove external ref count visibility The returned ref count was never used and the returned value doesn't actually confer any meaning in a multithreaded environment
2014-02-03 02:34:06 +01:00
void obs_source_release(obs_source_t source)
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
{
Remove external ref count visibility The returned ref count was never used and the returned value doesn't actually confer any meaning in a multithreaded environment
2014-02-03 02:34:06 +01:00
if (!source)
return;
adjusted the way source removal is handled to make it a bit more safe
2013-11-21 02:36:46 +01:00
Remove external ref count visibility The returned ref count was never used and the returned value doesn't actually confer any meaning in a multithreaded environment
2014-02-03 02:34:06 +01:00
if (--source->refs == 0)
obs_source_destroy(source);
adjusted the way source removal is handled to make it a bit more safe
2013-11-21 02:36:46 +01:00
}
void obs_source_remove(obs_source_t source)
{
Improve naming scheme of libobs core structures
2014-02-06 05:03:06 +01:00
struct obs_core_data *data = &obs->data;
adjusted the way source removal is handled to make it a bit more safe
2013-11-21 02:36:46 +01:00
size_t id;
pthread_mutex_lock(&data->sources_mutex);
Improve thread safety in UI code - Implemented better C++ classes for handling scenes/sources/items in obs.hpp, allowing them to automatically increment and decrement the references of each, as well as assign them to QVariants. - Because QVariants are now using the C++ classes, remove the pointer QVariant wrapper. - Use the new C++ classes with the QVariant user data of list box items, both for the sake of thread safety and to ensure that the data referenced is not freed until removed. NOTE: still might need some testing. - Implemented a source-remove signal from libobs, and start using that signal instead of the source-destroy signal for signalling item removal.
2014-02-02 22:26:23 +01:00
if (!source || source->removed)
Add module UI export capability Add the ability to be able to call and use toolkit-specific or program-specific user interface in modules. User interface code can be either bundled with the module, or 'split' out in to separate libraries (recommended). There are three reasons why splitting is recommended: 1.) It allows plugins to be able to create custom user interface for each toolkit if desired. 2.) Often, UI will be programmed in one language (the language of the toolkit), and core logic may be programmed in another. This allows plugins to keep the languages separated if necessary. 3.) It prevents direct linkage of UI toolkits libraries with core module logic. Splitting is not required, though is recommended if you want your plugin to be more flexible with other user interface toolkits or programs. Will implement a generic properties lookup next, which will be used for automatic UI handling so that plugin UI isn't necessarily required.
2014-02-01 08:49:50 +01:00
return;
Improve thread safety in UI code - Implemented better C++ classes for handling scenes/sources/items in obs.hpp, allowing them to automatically increment and decrement the references of each, as well as assign them to QVariants. - Because QVariants are now using the C++ classes, remove the pointer QVariant wrapper. - Use the new C++ classes with the QVariant user data of list box items, both for the sake of thread safety and to ensure that the data referenced is not freed until removed. NOTE: still might need some testing. - Implemented a source-remove signal from libobs, and start using that signal instead of the source-destroy signal for signalling item removal.
2014-02-02 22:26:23 +01:00
source->removed = true;
Add module UI export capability Add the ability to be able to call and use toolkit-specific or program-specific user interface in modules. User interface code can be either bundled with the module, or 'split' out in to separate libraries (recommended). There are three reasons why splitting is recommended: 1.) It allows plugins to be able to create custom user interface for each toolkit if desired. 2.) Often, UI will be programmed in one language (the language of the toolkit), and core logic may be programmed in another. This allows plugins to keep the languages separated if necessary. 3.) It prevents direct linkage of UI toolkits libraries with core module logic. Splitting is not required, though is recommended if you want your plugin to be more flexible with other user interface toolkits or programs. Will implement a generic properties lookup next, which will be used for automatic UI handling so that plugin UI isn't necessarily required.
2014-02-01 08:49:50 +01:00
Improve thread safety in UI code - Implemented better C++ classes for handling scenes/sources/items in obs.hpp, allowing them to automatically increment and decrement the references of each, as well as assign them to QVariants. - Because QVariants are now using the C++ classes, remove the pointer QVariant wrapper. - Use the new C++ classes with the QVariant user data of list box items, both for the sake of thread safety and to ensure that the data referenced is not freed until removed. NOTE: still might need some testing. - Implemented a source-remove signal from libobs, and start using that signal instead of the source-destroy signal for signalling item removal.
2014-02-02 22:26:23 +01:00
obs_source_addref(source);
id = da_find(data->sources, &source, 0);
if (id != DARRAY_INVALID) {
da_erase_item(data->sources, &source);
obs_source_release(source);
adjusted the way source removal is handled to make it a bit more safe
2013-11-21 02:36:46 +01:00
}
pthread_mutex_unlock(&data->sources_mutex);
Improve thread safety in UI code - Implemented better C++ classes for handling scenes/sources/items in obs.hpp, allowing them to automatically increment and decrement the references of each, as well as assign them to QVariants. - Because QVariants are now using the C++ classes, remove the pointer QVariant wrapper. - Use the new C++ classes with the QVariant user data of list box items, both for the sake of thread safety and to ensure that the data referenced is not freed until removed. NOTE: still might need some testing. - Implemented a source-remove signal from libobs, and start using that signal instead of the source-destroy signal for signalling item removal.
2014-02-02 22:26:23 +01:00
obs_source_dosignal(source, "source-remove");
obs_source_release(source);
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
}
bool obs_source_removed(obs_source_t source)
{
return source->removed;
first commit
2013-10-01 04:37:13 +02:00
}
Add property list callbacks - Add property list callbacks to sources/outputs/encoders so that if necessary user interface can be automatically generated or perhaps a property list widget can be used for them. - Change some of the property API names. obs_property_list_t felt a bit awkward when actually using it, so I just renamed it to obs_properties_t. - Removed the getdata/setdata nad getparam/setparam functions from sources/services, they will be superseded by the dynamic procedure call API.
2014-02-02 06:46:13 +01:00
obs_properties_t obs_source_properties(enum obs_source_type type,
const char *id, const char *locale)
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
const struct obs_source_info *info = get_source_info(type, id);
if (info && info->get_properties)
return info->get_properties(locale);
Add property list callbacks - Add property list callbacks to sources/outputs/encoders so that if necessary user interface can be automatically generated or perhaps a property list widget can be used for them. - Change some of the property API names. obs_property_list_t felt a bit awkward when actually using it, so I just renamed it to obs_properties_t. - Removed the getdata/setdata nad getparam/setparam functions from sources/services, they will be superseded by the dynamic procedure call API.
2014-02-02 06:46:13 +01:00
return NULL;
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
uint32_t obs_source_get_output_flags(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
return source->info.output_flags;
first commit
2013-10-01 04:37:13 +02:00
}
Implement settings interface for plugins Add a fairly easy to use settings interface that can be passed to plugins, and replaced the old character string system that was being used before. The new data interface allows for an easier method of getting/altering settings for plugins, and is built to be serializable to/from JSON. Also, removed another wxFormBuilder file that was no longer in use.
2014-01-28 07:14:58 +01:00
void obs_source_update(obs_source_t source, obs_data_t settings)
first commit
2013-10-01 04:37:13 +02:00
{
Add a couple more setting data safety measures Prevent null dereferencing if data is null, instead just make the functions return and ignore the get/set requests.
2014-01-28 23:45:30 +01:00
obs_data_replace(&source->settings, settings);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.update)
source->info.update(source->data, source->settings);
first commit
2013-10-01 04:37:13 +02:00
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_activate(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.activate)
source->info.activate(source->data);
first commit
2013-10-01 04:37:13 +02:00
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_deactivate(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.deactivate)
source->info.deactivate(source->data);
first commit
2013-10-01 04:37:13 +02:00
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_video_tick(obs_source_t source, float seconds)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
/* reset the filter render texture information once every frame */
if (source->filter_texrender)
texrender_reset(source->filter_texrender);
if (source->info.video_tick)
source->info.video_tick(source->data, seconds);
first commit
2013-10-01 04:37:13 +02:00
}
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
/* unless the value is 3+ hours worth of frames, this won't overflow */
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
static inline uint64_t conv_frames_to_time(size_t frames)
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
{
Simplify media i/o interfaces Completely revamped the entire media i/o data and handlers. The original idea was to have a system that would have connecting media inputs and outputs, but at a certain point I realized that this was an unnecessary complexity for what we wanted to do. (Also, it reminded me of directshow filters, and I HATE directshow with a passion, and wouldn't wish it upon my greatest enemy) Now, audio/video outputs are connected to directly, with better callback handlers, and will eventually have the ability to automatically handle conversions such as 4:4:4 to 4:2:0 when connecting to an input that uses them. Doing this will allow the video/audio i/o handlers to also prevent duplicate conversion, as well as make it easier/simple to use. My true goal for this is to make output and encoder plugins as simple to create as possible. I want to be able to be able to create an output plugin with almost no real hassle of having to worry about image conversions, media inputs/outputs, etc. A plugin developer shouldn't have to handle that sort of stuff when he/she doesn't really need to. Plugins will be able to simply create a callback via obs_video() and/or obs_audio(), and they will automatically receive the audio/video data in the formats requested via a simple callback, without needing to do almost anything else at all.
2014-01-14 09:58:47 +01:00
const struct audio_output_info *info;
info = audio_output_getinfo(obs->audio.audio);
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
return (uint64_t)frames * 1000000000ULL /
(uint64_t)info->samples_per_sec;
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
}
add support for texture_rect texture type in shaders, add new form, clear up a few things in the API
2013-12-20 19:56:01 +01:00
/* maximum "direct" timestamp variance in nanoseconds */
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
#define MAX_TS_VAR 5000000000ULL
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
/* maximum time that timestamp can jump in nanoseconds */
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
#define MAX_TIMESTAMP_JUMP 2000000000ULL
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
static inline void reset_audio_timing(obs_source_t source, uint64_t timetamp)
{
source->timing_set = true;
source->timing_adjust = os_gettime_ns() - timetamp;
}
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
static inline void handle_ts_jump(obs_source_t source, uint64_t ts,
uint64_t diff)
{
Fix inttypes.h usage ...I neglected to put a '%' character before using the PRI* macros.
2014-02-15 00:05:52 +01:00
blog(LOG_DEBUG, "Timestamp for source '%s' jumped by '%"PRIu64"', "
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
"resetting audio timing", source->name, diff);
/* if has video, ignore audio data until reset */
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.output_flags & OBS_SOURCE_ASYNC_VIDEO)
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
source->audio_reset_ref--;
else
reset_audio_timing(source, ts);
}
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
static void source_output_audio_line(obs_source_t source,
const struct audio_data *data)
{
struct audio_data in = *data;
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
uint64_t diff;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (!source->timing_set) {
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
reset_audio_timing(source, in.timestamp);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
/* detects 'directly' set timestamps as long as they're within
add support for texture_rect texture type in shaders, add new form, clear up a few things in the API
2013-12-20 19:56:01 +01:00
* a certain threshold */
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
if ((source->timing_adjust + MAX_TS_VAR) < MAX_TS_VAR * 2)
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
source->timing_adjust = 0;
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
} else {
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
diff = in.timestamp - source->next_audio_ts_min;
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
/* don't need signed because negative will trigger it
* regardless, which is what we want */
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
if (diff > MAX_TIMESTAMP_JUMP)
handle_ts_jump(source, in.timestamp, diff);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
}
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
source->next_audio_ts_min = in.timestamp +
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
conv_frames_to_time(in.frames);
Detect audio timestamp jumps and reset timing - Often, timestamps will go "back" in time with certain.. terrible devices that no one should use. When this occurs, timing is now reset so that the new audio comes in directly after the old audio seamlessly.
2014-01-10 20:45:53 +01:00
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
if (source->audio_reset_ref != 0)
return;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
in.timestamp += source->timing_adjust;
Implement volumes for sources - Added a volume variable to the obs_source structure and implemented functions for manipulating source volume. - Added a volume variable to the audio_data structure so that the volume will be applied when mixing.
2014-01-07 19:03:15 +01:00
in.volume = source->volume;
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
audio_line_output(source->audio_line, &in);
}
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
static bool set_texture_size(obs_source_t source, struct source_frame *frame)
{
if (source->output_texture) {
uint32_t width = texture_getwidth(source->output_texture);
uint32_t height = texture_getheight(source->output_texture);
if (width == frame->width && height == frame->height)
return true;
}
texture_destroy(source->output_texture);
source->output_texture = gs_create_texture(frame->width, frame->height,
GS_RGBA, 1, NULL, GS_DYNAMIC);
return source->output_texture != NULL;
}
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
enum convert_type {
CONVERT_NONE,
CONVERT_NV12,
CONVERT_420,
CONVERT_422_U,
CONVERT_422_Y,
};
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
static inline enum convert_type get_convert_type(enum video_format format)
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
{
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
switch (format) {
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
case VIDEO_FORMAT_I420:
return CONVERT_420;
case VIDEO_FORMAT_NV12:
return CONVERT_NV12;
case VIDEO_FORMAT_YVYU:
case VIDEO_FORMAT_YUY2:
return CONVERT_422_Y;
case VIDEO_FORMAT_UYVY:
return CONVERT_422_U;
Add preliminary ffmpeg plugin (still testing) - Added some code for FFmpeg output that I'm still playing around with. Right now I'm just trying to get it to output to file and try to understand the FFmpeg/libav APIs. Hopefully in the future this plugin can be used for any sort of output to FFmpeg. - Fixed a cast warning in audio-io.c with size_t -> uint32_t - Renamed the 'video_info' and 'audio_info' structures to 'video_conver_info' and 'audio_convert_info' to better represent their actual purpose, and to avoid confusion with 'audio_output_info' and 'video_output_info' structures. - Removed a few macros from obs-def.h that were at one point going to be used but no longer going to be used (at least for now)
2014-01-19 11:16:41 +01:00
case VIDEO_FORMAT_NONE:
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
case VIDEO_FORMAT_RGBA:
case VIDEO_FORMAT_BGRA:
case VIDEO_FORMAT_BGRX:
return CONVERT_NONE;
}
return CONVERT_NONE;
}
static bool upload_frame(texture_t tex, const struct source_frame *frame)
{
void *ptr;
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
uint32_t linesize;
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
enum convert_type type = get_convert_type(frame->format);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
if (type == CONVERT_NONE) {
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
texture_setimage(tex, frame->data[0], frame->linesize[0],
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
false);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
return true;
}
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
if (!texture_map(tex, &ptr, &linesize))
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
return false;
if (type == CONVERT_420)
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
decompress_420((const uint8_t* const*)frame->data,
frame->linesize,
0, frame->height, ptr, linesize);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
else if (type == CONVERT_NV12)
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
decompress_nv12((const uint8_t* const*)frame->data,
frame->linesize,
0, frame->height, ptr, linesize);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
else if (type == CONVERT_422_Y)
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
decompress_422(frame->data[0], frame->linesize[0],
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
0, frame->height, ptr, linesize, true);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
else if (type == CONVERT_422_U)
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
decompress_422(frame->data[0], frame->linesize[0],
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
0, frame->height, ptr, linesize, false);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
texture_unmap(tex);
return true;
}
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
static void obs_source_draw_texture(texture_t tex, struct source_frame *frame)
{
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
effect_t effect = obs->video.default_effect;
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
bool yuv = format_is_yuv(frame->format);
Implement output scaling/conversion/downloading - Implement texture scaling/conversion/downloading for the main view so we can finally start getting data to output. Also, redesign how it works a bit, it will now properly wait one full frame for each step in the process: rendering the main texture, scaling the main texture to an output texture, staging/downloading the ouput texture, and then outputting that staged data. This way, the GPU will have more than enough time to fully complete each step. - Fix a bug with OpenGL plugin's texture staging function. Was using glBindBuffer instead of what should have been used: glBindTexture. - Change the naming scheme of the variables in default.effect. It's now named with the idea of just "color matrix" in mind instead of "yuv matrix", and instead of DrawRGBToYUV, it's now just DrawMatrix.
2014-02-06 04:36:21 +01:00
const char *type = yuv ? "DrawMatrix" : "Draw";
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
technique_t tech;
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
eparam_t param;
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
if (!upload_frame(tex, frame))
return;
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
tech = effect_gettechnique(effect, type);
technique_begin(tech);
technique_beginpass(tech, 0);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
if (yuv) {
Implement output scaling/conversion/downloading - Implement texture scaling/conversion/downloading for the main view so we can finally start getting data to output. Also, redesign how it works a bit, it will now properly wait one full frame for each step in the process: rendering the main texture, scaling the main texture to an output texture, staging/downloading the ouput texture, and then outputting that staged data. This way, the GPU will have more than enough time to fully complete each step. - Fix a bug with OpenGL plugin's texture staging function. Was using glBindBuffer instead of what should have been used: glBindTexture. - Change the naming scheme of the variables in default.effect. It's now named with the idea of just "color matrix" in mind instead of "yuv matrix", and instead of DrawRGBToYUV, it's now just DrawMatrix.
2014-02-06 04:36:21 +01:00
param = effect_getparambyname(effect, "color_matrix");
effect_setval(effect, param, frame->color_matrix,
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
sizeof(float) * 16);
}
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
param = effect_getparambyname(effect, "image");
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
effect_settexture(effect, param, tex);
changed gs_draw_sprite to allow custom sizes, added output textures to the core, and adjusted the test code to accommodate the changes
2013-11-27 06:26:14 +01:00
gs_draw_sprite(tex, frame->flip ? GS_FLIP_V : 0, 0, 0);
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
technique_endpass(tech);
technique_end(tech);
}
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
static void obs_source_render_async_video(obs_source_t source)
{
struct source_frame *frame = obs_source_getframe(source);
if (!frame)
return;
add frame flipping code
2013-10-27 00:05:51 +02:00
if (set_texture_size(source, frame))
fill in the texture_setimage function, fill in a few other functions, and change certain names to be a little more consistent
2013-10-25 19:25:28 +02:00
obs_source_draw_texture(source->output_texture, frame);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
obs_source_releaseframe(source, frame);
}
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
static inline void obs_source_render_filters(obs_source_t source)
{
source->rendering_filter = true;
obs_source_video_render(source->filters.array[0]);
source->rendering_filter = false;
}
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
static inline void obs_source_default_render(obs_source_t source,
bool color_matrix)
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
{
effect_t effect = obs->video.default_effect;
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
const char *tech_name = color_matrix ? "DrawMatrix" : "Draw";
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
technique_t tech = effect_gettechnique(effect, tech_name);
size_t passes, i;
passes = technique_begin(tech);
for (i = 0; i < passes; i++) {
technique_beginpass(tech, i);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
source->info.video_render(source->data, effect);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
technique_endpass(tech);
}
technique_end(tech);
}
static inline void obs_source_main_render(obs_source_t source)
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
uint32_t flags = source->info.output_flags;
bool color_matrix = (flags & OBS_SOURCE_COLOR_MATRIX) != 0;
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
bool default_effect = !source->filter_parent &&
source->filters.num == 0 &&
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
(flags & OBS_SOURCE_CUSTOM_DRAW) == 0;
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
if (default_effect)
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
obs_source_default_render(source, color_matrix);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
else
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
source->info.video_render(source->data, NULL);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_video_render(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.video_render) {
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
if (source->filters.num && !source->rendering_filter)
obs_source_render_filters(source);
else
obs_source_main_render(source);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
} else if (source->filter_target) {
obs_source_video_render(source->filter_target);
} else {
obs_source_render_async_video(source);
first commit
2013-10-01 04:37:13 +02:00
}
}
some static analysis cleanup
2013-10-18 02:21:42 +02:00
uint32_t obs_source_getwidth(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.getwidth)
return source->info.getwidth(source->data);
first commit
2013-10-01 04:37:13 +02:00
return 0;
}
some static analysis cleanup
2013-10-18 02:21:42 +02:00
uint32_t obs_source_getheight(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->info.getheight)
return source->info.getheight(source->data);
first commit
2013-10-01 04:37:13 +02:00
return 0;
}
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
obs_source_t obs_filter_getparent(obs_source_t filter)
{
return filter->filter_parent;
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
obs_source_t obs_filter_gettarget(obs_source_t filter)
first commit
2013-10-01 04:37:13 +02:00
{
return filter->filter_target;
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_filter_add(obs_source_t source, obs_source_t filter)
first commit
2013-10-01 04:37:13 +02:00
{
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_lock(&source->filter_mutex);
some static analysis cleanup
2013-10-18 02:21:42 +02:00
if (da_find(source->filters, &filter, 0) != DARRAY_INVALID) {
first commit
2013-10-01 04:37:13 +02:00
blog(LOG_WARNING, "Tried to add a filter that was already "
"present on the source");
return;
}
if (source->filters.num) {
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
obs_source_t *back = da_end(source->filters);
first commit
2013-10-01 04:37:13 +02:00
(*back)->filter_target = filter;
}
da_push_back(source->filters, &filter);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_unlock(&source->filter_mutex);
filter->filter_parent = source;
first commit
2013-10-01 04:37:13 +02:00
filter->filter_target = source;
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_filter_remove(obs_source_t source, obs_source_t filter)
first commit
2013-10-01 04:37:13 +02:00
{
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
size_t idx;
pthread_mutex_lock(&source->filter_mutex);
idx = da_find(source->filters, &filter, 0);
some static analysis cleanup
2013-10-18 02:21:42 +02:00
if (idx == DARRAY_INVALID)
first commit
2013-10-01 04:37:13 +02:00
return;
if (idx > 0) {
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
obs_source_t prev = source->filters.array[idx-1];
first commit
2013-10-01 04:37:13 +02:00
prev->filter_target = filter->filter_target;
}
da_erase(source->filters, idx);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_unlock(&source->filter_mutex);
filter->filter_parent = NULL;
first commit
2013-10-01 04:37:13 +02:00
filter->filter_target = NULL;
}
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
void obs_source_filter_setorder(obs_source_t source, obs_source_t filter,
first commit
2013-10-01 04:37:13 +02:00
enum order_movement movement)
{
size_t idx = da_find(source->filters, &filter, 0);
size_t i;
some static analysis cleanup
2013-10-18 02:21:42 +02:00
if (idx == DARRAY_INVALID)
first commit
2013-10-01 04:37:13 +02:00
return;
if (movement == ORDER_MOVE_UP) {
if (idx == source->filters.num-1)
return;
da_move_item(source->filters, idx, idx+1);
} else if (movement == ORDER_MOVE_DOWN) {
if (idx == 0)
return;
da_move_item(source->filters, idx, idx-1);
} else if (movement == ORDER_MOVE_TOP) {
if (idx == source->filters.num-1)
return;
da_move_item(source->filters, idx, source->filters.num-1);
} else if (movement == ORDER_MOVE_BOTTOM) {
if (idx == 0)
return;
da_move_item(source->filters, idx, 0);
}
add support for texture_rect texture type in shaders, add new form, clear up a few things in the API
2013-12-20 19:56:01 +01:00
/* reorder filter targets, not the nicest way of dealing with things */
first commit
2013-10-01 04:37:13 +02:00
for (i = 0; i < source->filters.num; i++) {
change names, fix some bugs, minor GL/D3D fixes, update tests, fix effect files, output a little more debug information
2013-10-14 21:37:52 +02:00
obs_source_t next_filter = (i == source->filters.num-1) ?
first commit
2013-10-01 04:37:13 +02:00
source : source->filters.array[idx+1];
source->filters.array[i]->filter_target = next_filter;
}
}
Implement settings interface for plugins Add a fairly easy to use settings interface that can be passed to plugins, and replaced the old character string system that was being used before. The new data interface allows for an easier method of getting/altering settings for plugins, and is built to be serializable to/from JSON. Also, removed another wxFormBuilder file that was no longer in use.
2014-01-28 07:14:58 +01:00
obs_data_t obs_source_getsettings(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Implement settings interface for plugins Add a fairly easy to use settings interface that can be passed to plugins, and replaced the old character string system that was being used before. The new data interface allows for an easier method of getting/altering settings for plugins, and is built to be serializable to/from JSON. Also, removed another wxFormBuilder file that was no longer in use.
2014-01-28 07:14:58 +01:00
obs_data_addref(source->settings);
return source->settings;
first commit
2013-10-01 04:37:13 +02:00
}
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
static inline struct source_frame *filter_async_video(obs_source_t source,
struct source_frame *in)
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
{
size_t i;
for (i = source->filters.num; i > 0; i--) {
struct obs_source *filter = source->filters.array[i-1];
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (filter->info.filter_video) {
in = filter->info.filter_video(filter->data, in);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (!in)
return NULL;
}
}
return in;
}
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
static inline void copy_frame_data_line(struct source_frame *dst,
const struct source_frame *src, uint32_t plane, uint32_t y)
{
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
uint32_t pos_src = y * src->linesize[plane];
uint32_t pos_dst = y * dst->linesize[plane];
uint32_t bytes = dst->linesize[plane] < src->linesize[plane] ?
dst->linesize[plane] : src->linesize[plane];
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
memcpy(dst->data[plane] + pos_dst, src->data[plane] + pos_src, bytes);
}
static inline void copy_frame_data_plane(struct source_frame *dst,
const struct source_frame *src, uint32_t plane, uint32_t lines)
{
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
if (dst->linesize[plane] != src->linesize[plane])
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
for (uint32_t y = 0; y < lines; y++)
copy_frame_data_line(dst, src, plane, y);
else
memcpy(dst->data[plane], src->data[plane],
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
dst->linesize[plane] * lines);
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
}
static void copy_frame_data(struct source_frame *dst,
const struct source_frame *src)
{
dst->flip = src->flip;
dst->timestamp = src->timestamp;
memcpy(dst->color_matrix, src->color_matrix, sizeof(float) * 16);
switch (dst->format) {
case VIDEO_FORMAT_I420:
copy_frame_data_plane(dst, src, 0, dst->height);
copy_frame_data_plane(dst, src, 1, dst->height/2);
copy_frame_data_plane(dst, src, 2, dst->height/2);
break;
case VIDEO_FORMAT_NV12:
copy_frame_data_plane(dst, src, 0, dst->height);
copy_frame_data_plane(dst, src, 1, dst->height/2);
break;
case VIDEO_FORMAT_YVYU:
case VIDEO_FORMAT_YUY2:
case VIDEO_FORMAT_UYVY:
case VIDEO_FORMAT_NONE:
case VIDEO_FORMAT_RGBA:
case VIDEO_FORMAT_BGRA:
case VIDEO_FORMAT_BGRX:
copy_frame_data_plane(dst, src, 0, dst->height);
}
}
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
static inline struct source_frame *cache_video(const struct source_frame *frame)
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
{
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
/* TODO: use an actual cache */
Implement output, improve video/audio subsystems - Fill in the rest of the FFmpeg test output code for testing so it actually properly outputs data. - Improve the main video subsystem to be a bit more optimal and automatically output I420 or NV12 if needed. - Fix audio subsystem insertation and byte calculation. Now it will seamlessly insert new audio data in to the audio stream based upon its timestamp value. (Be extremely cautious when using floating point calculations for important things like this, and always round your values and check your values) - Use 32 byte alignment in case of future optimizations and export a function to get the current alignment. - Make os_sleepto_ns return true if slept, false if the time has already been passed before the call. - Fix sinewave output so that it actually properly calculates a middle C sinewave. - Change the use of row_bytes to linesize (also makes it a bit more consistent with FFmpeg's naming as well)
2014-02-09 13:51:06 +01:00
struct source_frame *new_frame = source_frame_create(frame->format,
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
frame->width, frame->height);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
copy_frame_data(new_frame, frame);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
return new_frame;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
}
void obs_source_output_video(obs_source_t source,
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
const struct source_frame *frame)
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
{
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
struct source_frame *output = cache_video(frame);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_lock(&source->filter_mutex);
output = filter_async_video(source, output);
pthread_mutex_unlock(&source->filter_mutex);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
if (output) {
pthread_mutex_lock(&source->video_mutex);
da_push_back(source->video_frames, &output);
pthread_mutex_unlock(&source->video_mutex);
}
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
}
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
static inline struct filtered_audio *filter_async_audio(obs_source_t source,
struct filtered_audio *in)
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
{
size_t i;
for (i = source->filters.num; i > 0; i--) {
struct obs_source *filter = source->filters.array[i-1];
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (filter->info.filter_audio) {
in = filter->info.filter_audio(filter->data, in);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (!in)
return NULL;
}
}
return in;
}
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
static inline void reset_resampler(obs_source_t source,
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
const struct source_audio *audio)
{
Simplify media i/o interfaces Completely revamped the entire media i/o data and handlers. The original idea was to have a system that would have connecting media inputs and outputs, but at a certain point I realized that this was an unnecessary complexity for what we wanted to do. (Also, it reminded me of directshow filters, and I HATE directshow with a passion, and wouldn't wish it upon my greatest enemy) Now, audio/video outputs are connected to directly, with better callback handlers, and will eventually have the ability to automatically handle conversions such as 4:4:4 to 4:2:0 when connecting to an input that uses them. Doing this will allow the video/audio i/o handlers to also prevent duplicate conversion, as well as make it easier/simple to use. My true goal for this is to make output and encoder plugins as simple to create as possible. I want to be able to be able to create an output plugin with almost no real hassle of having to worry about image conversions, media inputs/outputs, etc. A plugin developer shouldn't have to handle that sort of stuff when he/she doesn't really need to. Plugins will be able to simply create a callback via obs_video() and/or obs_audio(), and they will automatically receive the audio/video data in the formats requested via a simple callback, without needing to do almost anything else at all.
2014-01-14 09:58:47 +01:00
const struct audio_output_info *obs_info;
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
struct resample_info output_info;
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
obs_info = audio_output_getinfo(obs->audio.audio);
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
output_info.format = obs_info->format;
output_info.samples_per_sec = obs_info->samples_per_sec;
output_info.speakers = obs_info->speakers;
source->sample_info.format = audio->format;
source->sample_info.samples_per_sec = audio->samples_per_sec;
source->sample_info.speakers = audio->speakers;
if (source->sample_info.samples_per_sec == obs_info->samples_per_sec &&
source->sample_info.format == obs_info->format &&
source->sample_info.speakers == obs_info->speakers) {
source->audio_failed = false;
return;
}
audio_resampler_destroy(source->resampler);
source->resampler = audio_resampler_create(&output_info,
&source->sample_info);
source->audio_failed = source->resampler == NULL;
if (source->resampler == NULL)
blog(LOG_ERROR, "creation of resampler failed");
}
static inline void copy_audio_data(obs_source_t source,
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
const uint8_t *const data[], uint32_t frames, uint64_t ts)
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
{
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
size_t planes = audio_output_planes(obs->audio.audio);
cleaned up main internal data structure design, changed to reference counting for sources to ensure safe destruction of source objects from all parts of the system, added some service-related stuff for testing
2013-11-20 23:00:16 +01:00
size_t blocksize = audio_output_blocksize(obs->audio.audio);
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
size_t size = (size_t)frames * blocksize;
bool resize = source->audio_storage_size < size;
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
source->audio_data.frames = frames;
source->audio_data.timestamp = ts;
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
for (size_t i = 0; i < planes; i++) {
/* ensure audio storage capacity */
if (resize) {
bfree(source->audio_data.data[i]);
source->audio_data.data[i] = bmalloc(size);
}
memcpy(source->audio_data.data[i], data[i], size);
}
if (resize)
source->audio_storage_size = size;
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
}
/* resamples/remixes new audio to the designated main audio output format */
static void process_audio(obs_source_t source, const struct source_audio *audio)
{
if (source->sample_info.samples_per_sec != audio->samples_per_sec ||
source->sample_info.format != audio->format ||
source->sample_info.speakers != audio->speakers)
reset_resampler(source, audio);
if (source->audio_failed)
return;
if (source->resampler) {
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
uint8_t *output[MAX_AV_PLANES];
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
uint32_t frames;
uint64_t offset;
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
memset(output, 0, sizeof(output));
audio_resampler_resample(source->resampler,
output, &frames, &offset,
audio->data, audio->frames);
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
copy_audio_data(source, (const uint8_t *const *)output, frames,
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
audio->timestamp - offset);
} else {
copy_audio_data(source, audio->data, audio->frames,
audio->timestamp);
}
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
}
void obs_source_output_audio(obs_source_t source,
const struct source_audio *audio)
{
uint32_t flags = obs_source_get_output_flags(source);
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
struct filtered_audio *output;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
process_audio(source, audio);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_lock(&source->filter_mutex);
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
output = filter_async_audio(source, &source->audio_data);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
if (output) {
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
bool async = (flags & OBS_SOURCE_ASYNC_VIDEO) == 0;
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
pthread_mutex_lock(&source->audio_mutex);
add support for texture_rect texture type in shaders, add new form, clear up a few things in the API
2013-12-20 19:56:01 +01:00
/* wait for video to start before outputting any audio so we
* have a base for sync */
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (source->timing_set || async) {
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
struct audio_data data;
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
for (int i = 0; i < MAX_AV_PLANES; i++)
Add planar audio support, improve test output - Add planar audio support. FFmpeg and libav use planar audio for many encoders, so it was somewhat necessary to add support in libobs itself. - Improve/adjust FFmpeg test output plugin. The exports were somewhat messed up (making me rethink how exports should be done). Not yet functional; it handles video properly, but it still does not handle audio properly. - Improve planar video code. The planar video code was not properly accounting for row sizes for each plane. Specifying row sizes for each plane has now been added. This will also make it more compatible with FFmpeg/libav. - Fixed a bug where callbacks wouldn't create properly in audio-io and video-io code. - Implement 'blogva' function to allow for va_list usage with libobs logging.
2014-02-07 11:03:54 +01:00
data.data[i] = output->data[i];
finish up most of the source audio stuff and rename some variables/structs/enum to be a bit more consistent
2013-10-31 18:28:47 +01:00
data.frames = output->frames;
data.timestamp = output->timestamp;
source_output_audio_line(source, &data);
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
}
pthread_mutex_unlock(&source->audio_mutex);
}
pthread_mutex_unlock(&source->filter_mutex);
}
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
static inline bool frame_out_of_bounds(obs_source_t source, uint64_t ts)
{
return ((ts - source->last_frame_ts) > MAX_TIMESTAMP_JUMP);
}
static inline struct source_frame *get_closest_frame(obs_source_t source,
Fix one minor bug with new timing code When the first async video frame is used it would not set audio timing, moved that code into obs_source_getframe. Also, might consider renaming obs_source_getframe. "Query frame" instead perhaps? Will sleep on it, might not even bother.
2014-01-12 11:24:01 +01:00
uint64_t sys_time, int *audio_time_refs)
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
{
struct source_frame *next_frame = source->video_frames.array[0];
struct source_frame *frame = NULL;
uint64_t sys_offset = sys_time - source->last_sys_timestamp;
uint64_t frame_time = next_frame->timestamp;
uint64_t frame_offset = 0;
/* account for timestamp invalidation */
if (frame_out_of_bounds(source, frame_time)) {
source->last_frame_ts = next_frame->timestamp;
Fix one minor bug with new timing code When the first async video frame is used it would not set audio timing, moved that code into obs_source_getframe. Also, might consider renaming obs_source_getframe. "Query frame" instead perhaps? Will sleep on it, might not even bother.
2014-01-12 11:24:01 +01:00
(*audio_time_refs)++;
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
} else {
frame_offset = frame_time - source->last_frame_ts;
source->last_frame_ts += sys_offset;
}
while (frame_offset <= sys_offset) {
source_frame_destroy(frame);
frame = next_frame;
da_erase(source->video_frames, 0);
if (!source->video_frames.num)
break;
next_frame = source->video_frames.array[0];
/* more timestamp checking and compensating */
if ((next_frame->timestamp - frame_time) > MAX_TIMESTAMP_JUMP) {
source->last_frame_ts =
next_frame->timestamp - frame_offset;
Fix one minor bug with new timing code When the first async video frame is used it would not set audio timing, moved that code into obs_source_getframe. Also, might consider renaming obs_source_getframe. "Query frame" instead perhaps? Will sleep on it, might not even bother.
2014-01-12 11:24:01 +01:00
(*audio_time_refs)++;
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
}
frame_time = next_frame->timestamp;
frame_offset = frame_time - source->last_frame_ts;
}
return frame;
}
make text lookup use a simple linked lists rather than an unnecessary array of pointers
2013-12-07 19:23:49 +01:00
/*
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
* Ensures that cached frames are displayed on time. If multiple frames
* were cached between renders, then releases the unnecessary frames and uses
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
* the frame with the closest timing to ensure sync. Also ensures that timing
* with audio is synchronized.
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
*/
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
struct source_frame *obs_source_getframe(obs_source_t source)
first commit
2013-10-01 04:37:13 +02:00
{
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
struct source_frame *frame = NULL;
Fix one minor bug with new timing code When the first async video frame is used it would not set audio timing, moved that code into obs_source_getframe. Also, might consider renaming obs_source_getframe. "Query frame" instead perhaps? Will sleep on it, might not even bother.
2014-01-12 11:24:01 +01:00
int audio_time_refs = 0;
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
uint64_t sys_time;
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
pthread_mutex_lock(&source->video_mutex);
if (!source->video_frames.num)
goto unlock;
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
sys_time = os_gettime_ns();
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
if (!source->last_frame_ts) {
frame = source->video_frames.array[0];
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
da_erase(source->video_frames, 0);
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
source->last_frame_ts = frame->timestamp;
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
} else {
Fix one minor bug with new timing code When the first async video frame is used it would not set audio timing, moved that code into obs_source_getframe. Also, might consider renaming obs_source_getframe. "Query frame" instead perhaps? Will sleep on it, might not even bother.
2014-01-12 11:24:01 +01:00
frame = get_closest_frame(source, sys_time, &audio_time_refs);
}
/* reset timing to current system time */
if (frame) {
source->audio_reset_ref += audio_time_refs;
source->timing_adjust = sys_time - frame->timestamp;
source->timing_set = true;
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
}
source->last_sys_timestamp = sys_time;
unlock:
pthread_mutex_unlock(&source->video_mutex);
made the getframe function a bit more safe and modified a few names for clarity
2013-11-23 00:18:31 +01:00
Add preliminary handling of timestamp invalidation - Add preliminary (yet to be tested) handling of timestamp invalidation issues that can happen with specific devices, where timestamps can reset or go backward/forward in time with no rhyme or reason. Spent the entire day just trying to figure out the best way to handle this. If both audio and video are present, it will increment a reference counter if video timestamps invalidate, and decrement the reference counter when the audio timestamps invalidate. When the reference counter is not 0, it will not send audio as the audio will have invalid timing. What this does is it ensures audio data will never go out of bounds in relation to the video, and waits for both audio and video timestamps to "jump" together before resuming audio. - Moved async video frame timing adjustment code into obs_source_getframe instead so it's automatically handled whenever called. - Removed the 'audio wait buffer' as it was an unnecessary complexity that could have had problems in the future. Instead, audio will not be added until video starts for sources that have both async audio/video. Audio could have buffered for too long of a time anyway, who knows what devices are going to do. - Fixed a minor conversion warning in audio-io.c
2014-01-12 10:40:51 +01:00
if (frame)
made the getframe function a bit more safe and modified a few names for clarity
2013-11-23 00:18:31 +01:00
obs_source_addref(source);
fill out the rest of the source video frame functions, added nv12 decompression function, and cleaned up the design of the source video frame stuff
2013-10-26 23:32:06 +02:00
return frame;
first commit
2013-10-01 04:37:13 +02:00
}
added intial async audio/video code, fixed a few bugs, improved thread safety, and made a few other minor adjustments
2013-10-24 09:57:55 +02:00
void obs_source_releaseframe(obs_source_t source, struct source_frame *frame)
first commit
2013-10-01 04:37:13 +02:00
{
made the getframe function a bit more safe and modified a few names for clarity
2013-11-23 00:18:31 +01:00
if (frame) {
source_frame_destroy(frame);
obs_source_release(source);
}
first commit
2013-10-01 04:37:13 +02:00
}
add obs_source functions to allow name manipulation and lookup based upon name, and a few other functions related to getting/setting source information
2013-12-21 03:35:12 +01:00
const char *obs_source_getname(obs_source_t source)
{
return source->name;
}
void obs_source_setname(obs_source_t source, const char *name)
{
bfree(source->name);
source->name = bstrdup(name);
}
implement scene adding callbacks, make a few API tweaks
2013-12-28 13:33:16 +01:00
void obs_source_gettype(obs_source_t source, enum obs_source_type *type,
add obs_source functions to allow name manipulation and lookup based upon name, and a few other functions related to getting/setting source information
2013-12-21 03:35:12 +01:00
const char **id)
{
implement scene adding callbacks, make a few API tweaks
2013-12-28 13:33:16 +01:00
if (type) *type = source->type;
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (id) *id = source->info.id;
add obs_source functions to allow name manipulation and lookup based upon name, and a few other functions related to getting/setting source information
2013-12-21 03:35:12 +01:00
}
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
static inline void render_filter_bypass(obs_source_t target, effect_t effect,
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
bool use_matrix)
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
const char *tech_name = use_matrix ? "DrawMatrix" : "Draw";
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
technique_t tech = effect_gettechnique(effect, tech_name);
size_t passes, i;
passes = technique_begin(tech);
for (i = 0; i < passes; i++) {
technique_beginpass(tech, i);
obs_source_video_render(target);
technique_endpass(tech);
}
technique_end(tech);
}
static inline void render_filter_tex(texture_t tex, effect_t effect,
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
uint32_t width, uint32_t height, bool use_matrix)
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
{
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
const char *tech_name = use_matrix ? "DrawMatrix" : "Draw";
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
technique_t tech = effect_gettechnique(effect, tech_name);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
eparam_t image = effect_getparambyname(effect, "image");
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
size_t passes, i;
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
effect_settexture(effect, image, tex);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
passes = technique_begin(tech);
for (i = 0; i < passes; i++) {
technique_beginpass(tech, i);
gs_draw_sprite(tex, width, height, 0);
technique_endpass(tech);
}
technique_end(tech);
}
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
void obs_source_process_filter(obs_source_t filter, effect_t effect,
uint32_t width, uint32_t height, enum gs_color_format format,
make direct filter rendering optional (can be useful for certain cases)
2013-12-22 10:03:40 +01:00
enum allow_direct_render allow_direct)
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
{
obs_source_t target = obs_filter_gettarget(filter);
obs_source_t parent = obs_filter_getparent(filter);
uint32_t target_flags = obs_source_get_output_flags(target);
uint32_t parent_flags = obs_source_get_output_flags(parent);
int cx = obs_source_getwidth(target);
int cy = obs_source_getheight(target);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
bool use_matrix = !!(target_flags & OBS_SOURCE_COLOR_MATRIX);
bool expects_def = !(parent_flags & OBS_SOURCE_CUSTOM_DRAW);
make direct filter rendering optional (can be useful for certain cases)
2013-12-22 10:03:40 +01:00
bool can_directly = allow_direct == ALLOW_DIRECT_RENDERING;
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
/* if the parent does not use any custom effects, and this is the last
* filter in the chain for the parent, then render the parent directly
* using the filter effect instead of rendering to texture to reduce
* the total number of passes */
make direct filter rendering optional (can be useful for certain cases)
2013-12-22 10:03:40 +01:00
if (can_directly && expects_def && target == parent) {
Remove majority of warnings There were a *lot* of warnings, managed to remove most of them. Also, put warning flags before C_FLAGS and CXX_FLAGS, rather than after, as -Wall -Wextra was overwriting flags that came before it.
2014-02-14 23:13:36 +01:00
render_filter_bypass(target, effect, use_matrix);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
return;
}
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
if (!filter->filter_texrender)
filter->filter_texrender = texrender_create(format,
GS_ZS_NONE);
if (texrender_begin(filter->filter_texrender, cx, cy)) {
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
gs_ortho(0.0f, (float)cx, 0.0f, (float)cy, -100.0f, 100.0f);
make direct filter rendering optional (can be useful for certain cases)
2013-12-22 10:03:40 +01:00
if (expects_def && parent == target)
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
obs_source_default_render(parent, use_matrix);
make direct filter rendering optional (can be useful for certain cases)
2013-12-22 10:03:40 +01:00
else
obs_source_video_render(target);
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
texrender_end(filter->filter_texrender);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
}
/* --------------------------- */
Revamp API and start using doxygen The API used to be designed in such a way to where it would expect exports for each individual source/output/encoder/etc. You would export functions for each and it would automatically load those functions based on a specific naming scheme from the module. The idea behind this was that I wanted to limit the usage of structures in the API so only functions could be used. It was an interesting idea in theory, but this idea turned out to be flawed in a number of ways: 1.) Requiring exports to create sources/outputs/encoders/etc meant that you could not create them by any other means, which meant that things like faruton's .net plugin would become difficult. 2.) Export function declarations could not be checked, therefore if you created a function with the wrong parameters and parameter types, the compiler wouldn't know how to check for that. 3.) Required overly complex load functions in libobs just to handle it. It makes much more sense to just have a load function that you call manually. Complexity is the bane of all good programs. 4.) It required that you have functions of specific names, which looked and felt somewhat unsightly. So, to fix these issues, I replaced it with a more commonly used API scheme, seen commonly in places like kernels and typical C libraries with abstraction. You simply create a structure that contains the callback definitions, and you pass it to a function to register that definition (such as obs_register_source), which you call in the obs_module_load of the module. It will also automatically check the structure size and ensure that it only loads the required values if the structure happened to add new values in an API change. The "main" source file for each module must include obs-module.h, and must use OBS_DECLARE_MODULE() within that source file. Also, started writing some doxygen documentation in to the main library headers. Will add more detailed documentation as I go.
2014-02-12 16:04:50 +01:00
render_filter_tex(texrender_gettexture(filter->filter_texrender),
effect, width, height, use_matrix);
add major optimization to filter processing, and as a nice side effect, make it easier to create new filters and sources
2013-12-22 09:30:18 +01:00
}
implement signal/procedure handling into libobs and individual sources
2013-12-27 07:10:15 +01:00
signal_handler_t obs_source_signalhandler(obs_source_t source)
{
return source->signals;
}
proc_handler_t obs_source_prochandler(obs_source_t source)
{
return source->procs;
}
Implement volumes for sources - Added a volume variable to the obs_source structure and implemented functions for manipulating source volume. - Added a volume variable to the audio_data structure so that the volume will be applied when mixing.
2014-01-07 19:03:15 +01:00
void obs_source_setvolume(obs_source_t source, float volume)
{
source->volume = volume;
}
float obs_source_getvolume(obs_source_t source)
{
return source->volume;
}
Copy Permalink