Remove the update time check from the volume meter since it is no longer
nescessary. The update interval can now be set with the volmeter which has
the added benefit of not loosing data where previously updates had been
skipped.
This was an important change because we were originally using an
hard-coded 709/partial range color matrix for the output, which was
causing problems for people wanting to use different formats or color
spaces. This will now automatically generate the color matrix depending
on the format, color space, and range, or use an identity matrix if the
video format is RGB instead of YUV.
This replaces the old code for the audio meter that was using
calculations in two different places with the new audio meter api.
The source signal will now emit simple levels instead of dB values,
in order to avoid dB conversion and calculation in the source.
The GUI on the other hand now expects simple position values from
the volume meter api with no knowledge about dB calculus either.
That way all code that handles those conversions is in one place,
with the option to easily add new mappings that can be used
everywhere.
This replaces the code for volume control with a newer version that
uses the new fader implementation in order to control the fader/slider
position and source volume.
The volume label will now indicate the gain in dB instead of percent,
mainly because it looks cool and is easy to do.
Due to libobs saving the multiplier option for the source instead of
the slider/fader positon, existing volume levels will (mostly) stay
the same with only the slider beeing at a different position.
This is of course within the resolution of the slider (100 steps).
The remuxer thread was only started if there was an actual remux job,
which resulted in the remuxer thread not being able to call the worker's
destructor (because it wasn't running)
OBS Sparkle feeds have two extensions to vanilla Sparkle feeds:
- There can be two kinds of items per feed: (zipped) .app and .mpkg
via <ce:packageType>app|mpkg</ce:packageType> (default is mpkg)
- Feed items can be disabled via <ce:deployed>false</ce:deployed>; these
items will not be considered for updates unless
"[General] UpdateToUndeployed=1" is set the global config
Unlike other Sparkle implementations the FeedURL cannot be updated via user
preferences because we support multiple app packages with the same package
identifier but different FeedURL settings on the same machine
Typedef pointers are unsafe. If you do:
typedef struct bla *bla_t;
then you cannot use it as a constant, such as: const bla_t, because
that constant will be to the pointer itself rather than to the
underlying data. I admit this was a fundamental mistake that must
be corrected.
All typedefs that were pointer types will now have their pointers
removed from the type itself, and the pointers will be used when they
are actually used as variables/parameters/returns instead.
This does not break ABI though, which is pretty nice.
I feel like people keep mistaking how x264 preset functions. They often
associate it with their own CPU speed, and I feel the name does not help
with that issue. I think it may be for the best to just take it out.
At best, it will keep people away from using it unless they know what
they're doing, and at worst, people at least won't be able to associate
it with their CPU as easily anymore.
Change checkbox with text "Advanced Encoder Settings" to "Use Advanced
Encoder Settings". This way it will hint that the settings will be
disabled when unchecked.
Adds:
ENABLE_UI (on by default) which makes it so that the UI is required, and
will fail if a dependency is not found. This is on by default because
most people are building it with the user interface, and we'll probably
get a lot of issue reports stating "why is there no executable?" if we
don't have this on by default.
DISABLE_UI which forces the UI off.
If neither are set, then the UI will only be built if the dependencies
for it are found, otherwise the UI will be be ignored.
Some devices burst their audio (such as when querying audio from
directshow), and the 250 millisecond threshold that sets the audio meter
back to muted status would erroneously cause the meter to appear bounce
back between muted and unmuted. Instead, a one second test should be
sufficient time to prevent that from happening.
If the properties view changed in the properties window, the viewport of
the properties would not trigger a resize because the size of the window
itself does not change. This creates a signal that allows the parent to
know whether or not to update a viewport, if any.
Sometimes certain drivers do not have the capability required to
initialize the graphics subsystem (mesa currently, for example).
Instead of saying "your hardware is too old", state that the drivers may
also be at fault.
The informative text is meant to be additional information, not the main
information. If you use informative text without using regular text, it
causes the message box to get all squishy.
Do not use the object returned from obs_data_get_obj to set new
settings. It could be the default object, which means you'd just end up
modifying the default data, which won't get saved. Always create a new
object when setting sub-object data. It doesn't particularly hurt
anything.
Multiplication of the matricies was being done in the wrong direction.
This caused source transformations to come out looking incorrect, for
example the linux-xshm source's cursor would not be drawn correctly or
in the right position if the source was moved/scaled/rotated. The
problem just turned out to be that the gs_matrix_* functions were
multiplying in the wrong direction. Reverse the direction of
multiplication, and the problem is solved.
For the sake of consistency, renamed these two functions to include
_value at the end so they are consistent.
Renamed: To:
-------------------------------------------------------
obs_data_has_default obs_data_has_default_value
obs_data_has_autoselect obs_data_has_autoselect_value
obs_data_item_has_default obs_data_item_has_default_value
obs_data_item_has_autoselect obs_data_item_has_autoselect_value
Instead of having functions like obs_signal_handler() that can fail to
properly specify their actual intent in the name (does it signal a
handler, or does it return a signal handler?), always prefix functions
that are meant to get information with 'get' to make its functionality
more explicit.
Previous names: New names:
-----------------------------------------------------------
obs_audio obs_get_audio
obs_video obs_get_video
obs_signalhandler obs_get_signal_handler
obs_prochandler obs_get_proc_handler
obs_source_signalhandler obs_source_get_signal_handler
obs_source_prochandler obs_source_get_proc_handler
obs_output_signalhandler obs_output_get_signal_handler
obs_output_prochandler obs_output_get_proc_handler
obs_service_signalhandler obs_service_get_signal_handler
obs_service_prochandler obs_service_get_proc_handler
API Removed:
- graphics_t obs_graphics();
Replaced With:
- void obs_enter_graphics();
- void obs_leave_graphics();
Description:
obs_graphics() was somewhat of a pointless function. The only time
that it was ever necessary was to pass it as a parameter to
gs_entercontext() followed by a subsequent gs_leavecontext() call after
that. So, I felt that it made a bit more sense just to implement
obs_enter_graphics() and obs_leave_graphics() functions to do the exact
same thing without having to repeat that code. There's really no need
to ever "hold" the graphics pointer, though I suppose that could change
in the future so having a similar function come back isn't out of the
question.
Still, this at least reduces the amount of unnecessary repeated code for
the time being.
Changed:
- obs_source_gettype
To:
- enum obs_source_type obs_source_get_type(obs_source_t source);
- const char *obs_source_get_id(obs_source_t source);
This function was inconsistent for a number of reasons. First, it
returns both the ID and the type of source (input/transition/filter),
which is inconsistent with the name of "get type". Secondly, the
'squishy' naming convention which has just turned out to be bad
practice and causes inconsistencies. So it's now replaced with two
functions that just return the type and the ID.
Prefix with obs_ for the sake of consistency
Renamed enums:
- order_movement (now obs_order_movement)
Affected functions:
- obs_source_filter_setorder
- obs_sceneitem_setorder
The naming here is a poor choice, a source type determines if a source
is either a regular input, a filter, or a transition. The ID is the
actual unique identifier of a source.
The OBSBasic UI will now allow the use of a subdirectory of the user
application data directory for third-party plugins. Though I'm not
entirely sure if this ideal or not. Regardless, this is one of the
first (of many) steps towards a plugin manager.
On windows, this is %appdata%/obs-studio/plugins
On linux, this is ~/.obs-studio/plugins
On mac, this is ~/Library/Application Support/obs-sudio/plugins
Changed API:
- char *obs_find_plugin_file(const char *sub_path);
Changed to: char *obs_module_file(const char *file);
Cahnge it so you no longer need to specify a sub-path such as:
obs_find_plugin_file("module_name/file.ext")
Instead, now automatically handle the module data path so all you need
to do is:
obs_module_file("file.ext")
- int obs_load_module(const char *name);
Changed to: int obs_open_module(obs_module_t *module,
const char *path,
const char *data_path);
bool obs_init_module(obs_module_t module);
Change the module loading API so that if the front-end chooses, it can
load modules directly from a specified path, and associate a data
directory with it on the spot.
The module will not be initialized immediately; obs_init_module must
be called on the module pointer in order to fully initialize the
module. This is done so a module can be disabled by the front-end if
the it so chooses.
New API:
- void obs_add_module_path(const char *bin, const char *data);
These functions allow you to specify new module search paths to add,
and allow you to search through them, or optionally just load all
modules from them. If the string %module% is included, it will
replace it with the module's name when that string is used as a
lookup. Data paths are now directly added to the module's internal
storage structure, and when obs_find_module_file is used, it will look
up the pointer to the obs_module structure and get its data directory
that way.
Example:
obs_add_module_path("/opt/obs/my-modules/%module%/bin",
"/opt/obs/my-modules/%module%/data");
This would cause it to additionally look for the binary of a
hypthetical module named "foo" at /opt/obs/my-modules/foo/bin/foo.so
(or libfoo.so), and then look for the data in
/opt/obs/my-modules/foo/data.
This gives the front-end more flexibility for handling third-party
plugin modules, or handling all plugin modules in a custom way.
- void obs_find_modules(obs_find_module_callback_t callback, void
*param);
This searches the existing paths for modules and calls the callback
function when any are found. Useful for plugin management and custom
handling of the paths by the front-end if desired.
- void obs_load_all_modules(void);
Search through the paths and both loads and initializes all modules
automatically without custom handling.
- void obs_enum_modules(obs_enum_module_callback_t callback,
void *param);
Enumerates currently opened modules.
This refactors the sub-window code a bit so that instead of deleting the
window pointers, it calls QWidget::close() on them to safely trigger a
normal close on them instead (which will also delete them).
Moves setting the DeleteOnClose flag from inside of the Dialog classes
into the OBSBasic class, to make that behaviour more obvious.
This causes the main window to signal the application to exit and delete
its own pointer on close. This fixes an issue where apparently some
windows that aren't explicitly connected to the main window would be
left open when the main window was closed because by default Qt will
only exit when all windows have been closed.
Because it deletes its own pointer, instead of storing it in a
std::unique_ptr, use a QPointer because it has an internal mechanism for
automatically tracking QObject deletion even if the deletion was not
done on the QPointer itself, where as unique_ptr does not have that
functionality. In other words, if the pointer is deleted elsewhere for
whatever reason, the QPointer will still set that internal pointer value
to null.
(message and minor modificiations by Jim)
Changed API functions:
libobs: obs_reset_video
Before, video initialization returned a boolean, but "failed" is too
little information, if it fails due to lack of device capabilities or
bad video device parameters, the front-end needs to know that.
The OBS Basic UI has also been updated to reflect this API change.
There's no need to initialize the map value to 0. What was happening is
that obs_scene_add was adding a ref to a non-existent value, which
simply created it and added 1, which is perfectly fine. Then,
obs_add_source would set the ref to 0, overwriting the existing value.
So this meant that if you didn't call them in the right order, it
wouldn't work properly, and would break, which was pretty stupid.
Turns out that if you access a map value that doesn't exist, it'll
create one with the default constructor of that type. In this case, int
will initialize to 0, which was exactly what we wanted in the first
place, and defeats the purpose of even needing to initialize the value
to 0. So, there's no need to manually set it to 0 in
OBSBasic::SourceAdded, or worry about the order in which the functions
are called.
Just goes to show you have to be careful with reference counting.
These functions match the known obs locales with the system supplied
locales and return a vector of possible locales with the highest
priority locale first
Because we're using .ini format, the translation servies spit out files
with .ini extensions, so this makes it so we don't necessarily have to
rename those files from .ini to .txt before merging.
The status bar now displays:
- Auto-reconnect information (reconnecting and reconnect success)
- Dropped frames (as well as percentage of total video frames)
- Duration of session
- CPU usage of the program
- Kbp/s
The OBSBasic class is getting a bit big, so I separated out the
status bar code to its own class derived from QStatusBar.
Contains Move Up, Move Down, Move to Top, Move to Bottom. Also assigns
Ctrl-Up, Ctrl-Down, Ctrl-Home, Ctrl-End to each action.
This was also added to the right-click context menu popup for sources.
The removeItemAction just for a keyboard shortcut was unnecessary.
Instead, use the toolbar button to associate a shortcut with, and remove
the removeItemAction object.
I screwed it up a bit originally, using && instead of ||.
Use 'trimmed' function to prevent sources with leading or trailing
whitespace.
Also, do not allow an empty value.
Similar to the shader functions, the effect parameter functions take
the effect as a parameter. However, the effect parameter is pretty
pointless, because the effect parameter.. parameter stores the effect
pointer interally.
The locale parameter was a mistake, because it puts extra needless
burden upon the module developer to have to handle this variable for
each and every single callback function. The parameter is being removed
in favor of a single centralized module callback function that
specifically updates locale information for a module only when needed.
This API is used to set the current locale for libobs, which it will set
for all modules when a module is loaded or specifically when the locale
is manually changed.
Currently, if a user presses 'OK' or 'Apply' on the settings window, it
will save all data from all controls on a settings pane, regardless of
whether of not they were changed. The major issue with this is that
setting the data will overwrite all default values, making it impossible
for default values to be used if a user didn't actually change a value.
(Thanks to palana for pointing this fact out)
So instead, mark a control as 'changed' using QObject::property() and
QObject::sender(), and add a few helper functions to controls to ensure
that they are checked to see whether they were actually changed directly
by the user before saving the value to the config.
If the scene item has a bounding box set up for it, do not make it use
aspect ratio clamping by default.
Instead, make it so that shift will turn on aspect ratio, and make it
also apply to all types of bounding box modes.
The only time where aspect ratio clamping should apply by default is
when bounds are not in use (i.e. when a source is just created). Some
will disagree with me just because that's how photoshop does it, but
we're not photoshop, and I feel that the majority of users will have
more trouble with it disabled by default than enabled by default.
So to sum it up:
If bounds inactive, clamp aspect ratio by default, because scene items
start out with it inactive, and it directly affects the scale.
If bounds active, do not clamp aspect ratio by default, because clamping
to aspect ratio doesn't particularly have an effect for mode bounds
modes except for "stretch to bounds".
Instead of using bounds to force the source to be centered (and
unrotated), just center the source positionally on the screen. Will
also preserve rotation.
There's no reason to represent this value in terms of scale. Scale is a
useless value for users to use. What are they going to enter, 0.5?
2.0? 0.25?
Even if it can be subject to change by the source itself, and even if
it's still converted to scale internally, having it display the base
source size value is much more ideal for the user.
Every time I created a source I found myself in need to actually open up
the properties. It was getting somewhat on my nerves, so I decided to
just make it automatically pop up when the source is created.
Structures with anonymous unions would a warning when you do a brace
assignment on them.
Also fixed some unused parameters and removed some unused variables.
Add a 10 pixel padding to the sides and remove viewport cutting to
ensure that the editing rectangles are visible even when in the upper
corners.
Also, add a black background for the actual 'scene' in the preview
window so that the scene boundries are actually visible in relation to
the rest of the preview space.
So, scene editing was interesting (and by interesting I mean
excruciating). I almost implemented 'manipulator' visuals (ala 3dsmax
for example), and used 3 modes for controlling position/rotation/size,
but in a 2D editing, it felt clunky, so I defaulted back to simply
click-and-drag for movement, and then took a similar though slightly
different looking approach for handling scaling and reszing.
I also added a number of menu item helpers related to positioning,
scaling, rotating, flipping, and resetting the transform back to
default.
There is also a new 'transform' dialog (accessible via menu) which will
allow you to manually edit every single transform variable of a scene
item directly if desired.
If a scene item does not have bounds active, pulling on the sides of a
source will cause it to resize it via base scale rather than by the
bounding box system (if the source resizes that scale will apply). If
bounds are active, it will modify the bounding box only instead.
How a source scales when a bounding box is active depends on the type of
bounds being used. You can set it to scale to the inner bounds, the
outer bounds, scale to bounds width only, scale to bounds height only,
and a setting to stretch to bounds (which forces a source to always draw
at the bounding box size rather than be affected by its internal size).
You can also set it to be used as a 'maximum' size, so that the source
doesn't necessarily get scaled unless it extends beyond the bounds.
Like in OBS1, objects will snap to the edges unless the control key is
pressed. However, this will now happen even if the object is rotated or
oriented in any strange way. Snapping will also occur when stretching
or changing the bounding box size.
This adds support for the more smooth volume levels that accounts for
both level and magnitude. Currently, it just averages level and
magnitude, later on a full audio meter control can be made that would
properly utilize level, magnitude, and peak.
Also cleaned up the code a bit and removed some trailing whitespace.
Character conversion functions did not previously ask for a maximum
buffer size for their 'dst' parameter, it's unsafe to assume some given
destination buffer may have enough size to accommodate a conversion.
Implement the 'file path' in output settings, and implement the 'start
recording' button, though for the time being I'm just going to make it
use a directory rather than allow custom file names.
This file output will actually share the video and audio encoder with
the stream.
I don't really know what to do about MP4 -- I don't really like the idea
of saving directly in the program, if you do and the program crashes,
that MP4 file is lost. I'm contemplating making some sort of mp4 output
process stub. So no MP4 file output for the time being.
If you need MP4, just remux it with FFmpeg:
ffmpeg -i flv_file.flv -acodec copy -vcodec copy mp4_file.mp4
On windows this will return the documents\video directory, but on
linux/mac it'll just return $HOME for the time being because I don't
know if there really are any other appropriate adequate paths to use.
Perhaps someone else can be willing to fill this in if they wish.
