Add a new algorithm to calculate the true-peak. It implements the
Whittaker- Shannon interpolation from four samples to create 4
intermediate samples (5 x oversampling) inbetween the middle two
samples.
With 4 samples and 4 intermediate samples the algorithm can be
implemented as a 4x4 vector-matrix cross product, which is ideal for
SSE.
I've also replaced the sample-peak algorithm using SSE as well to
improve performance.
Closesobsproject/obs-studio#1189
Adds obs_add_raw_video_callback() and obs_remove_raw_video_callback()
functions which allow the ability to get raw video frames without
necessarily needing to create an output.
Reduces GPU usage when encoding is not active. Does not perform color
conversion, frame staging, or frame downloading unless encoding is
explicitly active.
The filters array should not be accessed without locking the filter
mutex. This locks the filter mutex, grabs a reference to the first
filter, unlocks the mutex, renders the filter, and then releases the
reference.
Closesobsproject/obs-studio#1215
Solves crash when a user tries to paste filters without selecting a
source or when pasting filters from a deleted source. This commit
checks if both sources are still valid before pasting.
This addresses Mantis Issue 1220
(https://obsproject.com/mantis/view.php?id=1220).
If a config file value is set to a blank value, the value would not be
set and the default value would be used. This fixes a bug where the
user would try to remove the replay buffer prefix, only to have the
prefix be added again next time they start the program.
Caches the frequency of the performance counter. It was supposed to be
caching it already, but it seems that code was unintentionally not
included.
Closesjp9000/obs-studio#1185
(This commit also modifies the obs-ffmpeg module)
The default channel layouts from aac spec are implemented in FFmpeg
native aac encoder as follows:
AV_CH_LAYOUT_MONO,
AV_CH_LAYOUT_STEREO,
AV_CH_LAYOUT_SURROUND,
AV_CH_LAYOUT_4POINT0,
AV_CH_LAYOUT_5POINT0_BACK,
AV_CH_LAYOUT_5POINT1_BACK,
AV_CH_LAYOUT_7POINT1,
The correspondence of speaker layouts to AV_CH_LAYOUT from FFmpeg is
changed to reflect the previous table.
Although FFmpeg native aac encoder can now encode all the layouts listed
in avutil channel_layout.h (on master), there might be issues with older
FFmpeg binaries.
Note that 2.1 speaker layout will be encoded as AV_CH_LAYOUT_SURROUND
(FL FR FC) because it is not listed as the default layout for three
channels.
This just means some optimizations for LFE channel will not be used by
the encoder which will treat it as an SCE (single channel element).
Closesjp9000/obs-studio#1182
When using more than two channels, the channel map of pulse-audio is incorrect.
Add an API for getting a speaker map based on OBS speaker layout. Then use the
speaker map when connecting to a pulse-audio device, for both source and
monitor output.
If a filter's implementation (its plugin for example) no longer exists,
it would cause the source to stop rendering if that filter was present
on the source. Instead, just bypass the filter to ensure that the
source continues to render.
The memset in custom_audio_render() did not clear all audio buffers when
the number of output channels was less then 8. This caused wrong audio
output on mixes that did not get cleared.
Closesjp9000/obs-studio#1123
When a scene is added as a scene item with the same audio sources that
are already in the current scene, it would cause the current scene to no
longer output audio due to audio.
To replicate the issue, you would create two separate audio device
captures in scene 1, use add existing in scene 2 and add one of those
audio sources, then go back to scene 1, add scene 2 as a source, then
make scene 1 invisible.
This pull request changes the fallback sample format for pulse-audio
to from PA_SAMPLE_S16LE to PA_SAMPLE_FLOAT32LE.
The pulseaudio plugin can handle the following sample format:
* PA_SAMPlE_U8
* PA_SAMPLE_S16LE
* PA_SAMPLE_S32LE
* PA_SAMPLE_FLOAT32LE
When an audio device advertises itself as another format, the pulseaudio-plugin
will ask pulse audio to convert to the fallback sample format.
The fallback PA_SAMPLE_S16LE is not ideal when your audio interface advertises
as PA_SAMPLE_S24LE since the conversion will lose precision.
With PA_SAMPLE_FLOAT32LE there is no precision loss and it is also equals OBS's
internal format.
(also obs, deps/media-playback, libobs/audio-monitoring, decklink,
linux-alsa, linux-pulseaudio, mac-capture, obs-ffmpeg, win-dshow,
win-wasapi)
Default channel layout for 4 channels is 4.0 in FFmpeg.
Replacing quad with 4.0 will improve compatibility since FFmpeg has
better support of its default channel layouts.
(also modifies obs-ffmpeg, audio-monitoring, win-wasapi, decklink,
obs-outputs)
Removes speaker layouts which are not exposed in UI. The speaker
layouts selectable by users in the UI are the most common ones. It is
not necessary to keep other layouts. (This basically removes
5POINT1_SURROUND, 7POINT1_SURROUND, SURROUND =3.0).
The following features have been added to the audio-meters:
* Stereo PPM-level meter, with 40 dB/1.7s decay rate.
* Stereo VU-level meter, with 300 ms integration time.
* Stereo Peak-hold meter, with 20 second sustain.
* Input peak level color-squares in front of every meter.
* Minor-ticks for each dB.
* Major-ticks for every 5 dB.
* Meter is divided in sections at -20 dB and -9 dB.
The ballistic parameters chosen here where taken from:
* https://en.wikipedia.org/wiki/Peak_programme_meter (SMPTE RP.0155)
* https://en.wikipedia.org/wiki/VU_meter
In the rework I have removed any ballistic calculations from
libobs/obs-audio-controls.c making the calculations here a lot more
simple doing only MAX and RMS calculations for only the samples in
the current update. The actual ballistics are now done by just
the UI/volume-control.cpp because ballistics need to be updated
based on the repaint-rate of the user-interface.
The dB to pixel conversion has been moved from
libobs/obs-audio-controls.c to UI/volume-control.cpp as well to reduce
coupling between these two objects, especially when implementing the
major- and minor-ticks and the sections.
All colors and ballistic parameters are adjustable via QT style sheets.
There are slight differences in colors for each of the themes.
(Windows only)
3rd party antivirus / firewalls cause a large number of issues in OBS.
Logging which products are installed will help narrow down potential
issues when users ask for help and include their log files.
Closesjp9000/obs-studio#1124
The "obs" global variable can still be accessed by functions during
shutdown. To prevent access to the variable during shutdown, move the
pointer to a temporary function variable, and set the "obs" global
variable to NULL before shutting down.
Individual property objects would store pointers to strings rather than
copies of the strings and assume that memory would continue to exist, so
instead just duplicate the strings.
Because it would be troublesome to add the ability to remove source
types (in case for example a script fails to reload), instead make it so
source types can be temporarily disabled while the program is running.
Convenience function that allows removing the current signal handler
callback within the callback without having to use
signal_handler_disconnect with all of its required parameters.
When splitting a string in to a string list via stringlist_split, it
would previously allocate memory for each sub-string and again for the
list itself. This optimizes that function to use a single contiguous
chunk of memory for the sake of access optimization and memory
allocation efficiency.
(Note: This commit also modifies UI and test)
This makes it so that main preview panes are rendered with the main
output texture rather than re-rendering the main view. The view will
render all objects again, whereas the output texture will be a single
texture render of the same exact thing.
Also fixes some abnormal artifacting when scaling the main preview pane.
The AVCodecParameters weren't introduced until avcodec version 57.48.101
(FFmpeg version 3.1), so this will make sure to still use the older
avcodec_copy_context if the detected FFmpeg version is earlier.
Adds circlebuf_push_front_zero and circlebuf_push_back_zero to
conveniently push zeroed data to the front/back of the buffer without
having to create an intermediary buffer to accomplish the same thing.
(Note: This commit also modifies libobs-d3d11 and libobs-opengl)
Allows the ability to flush data directly without having to use the
buffer's internal data.
Allows the caller to manage his/her own vertex/index buffer data if
desired, working around the design flaw of having to rely on a
vertex/index buffer's internal data.
Prevents from having to pass ownership of buffer data from caller when
using gs_vertexbuffer_create() or gs_indexbuffer_create() (which is a
design flaw).
(This commit also modifies the following modules: UI,
deps/media-playback, coreaudio-encoder, decklink, linux-alsa,
linux-pulseaudio, mac-capture, obs-ffmpeg, obs-filters, obs-libfdk,
obs-outputs, win-dshow, and win-wasapi)
Adds surround sound audio support to the core, core plugins, and user
interface.
Compatible streaming services: Twitch, FB 360 live
Compatible protocols: rtmp / mpeg-ts tcp udp
Compatible file formats: mkv mp4 ts (others untested)
Compatible codecs: ffmpeg aac, fdk_aac, CoreAudio aac,
opus, vorbis, pcm (others untested).
Tested streaming servers: wowza, nginx
HLS, mpeg-dash : surround passthrough
Html5 players tested with live surround:
videojs, mediaelement, viblast (hls+dash), hls.js
Decklink: on win32, swap channels order for 5.1 7.1
(due to different channel mapping on wav, mpeg, ffmpeg)
Audio filters: surround working.
Monitoring: surround working (win macOs linux (pulse-audio)).
VST: stereo plugins keep in general only the first two channels.
surround plugins should work (e.g. mcfx does).
OS: win, macOs, linux (alsa, pulse-audio).
Misc: larger audio bitrates unlocked to accommodate more channels
NB: mf-aac only supports mono and stereo + 5.1 on win 10
(not implemented due to lack of usefulness)
Closesjp9000/obs-studio#968
(This commit also modifies the deps/media-playback, obs-ffmpeg, and
win-dshow modules)
More fixes due to ffmpeg renaming some constants and deprecating
AVFMT_RAWPICTURE and AV_PIX_FMT_VDA_VLD.
Latter replaced by AV_PIX_FMT_VIDEOTOOLBOX per ffmpeg dev advice.
Closesjp9000/obs-studio#1061
This is to prevent confusion with video_thread in
libobs/media-io/video-io.c, which is used exclusively for video
encoding/output. Also prevents confusion in the profiler log data.
This commit adds logging for Windows 10 Gaming features that may
negatively impact the performance of game capture, recording, or
streaming. This doesn't check Group Policy settings.
Decoupling the audio from the video causes the audio to be played right
when it's received rather than attempt to sync up to the video frames.
This is useful with certain async sources/devices when the audio/video
timestamps are not reliable.
Naturally because it plays audio right when it's received, this should
only be used when the async source is operating in unbuffered mode,
otherwise the video frame timing will be out of sync by the amount of
buffering the video currently has.
Follow-up to 47aa56b (PR #620). Windows 10 revision detection broke in
Build 15036 (Creators Update) after Revision 296. This aims to further
fix revision detection on Windows 10.
The cmake script code used to find and depend on pulseaudio for audio
monitoring on linux was unintentionally being triggered on mac. This
patch makes it so that the pulseaudio cmake code is ignored on mac.
This function had a number of bugs and just wasn't working properly at
all. This function is currently not used in public builds because GPU
is used for color conversion instead (hence why it had probably not
really been tested), but a need came up where CPU conversion was useful
for the sake of testing something else in the back-end, and it needed to
be fixed before CPU conversion could be used.
When an async video frame comes in and it sets the timing_adjust value
(used to sync audio to video based upon their timestamps), it would use
os_gettime_ns as a base. Instead, it should use OBS' current video
frame time so that the audio and video playback is as accurate as
possible relative to the actual exact timestamp of the video frame.
(Results are almost insignificant, but it's nice to be as precise as
possible)
During packet interleaving (for outputs), ensure that if two packets
coincide with the same timestamp, that the video packet always comes
first instead of the audio packet.
This fix is required to make FLV demux properly with certain demuxers;
some FLV demuxers expect the video packet before the audio packet when
two packets coincide with the same timestamps.
(This commit also modifies the obs-outputs module)
The first video packet video offset (the value used to set the starting
point of video data) would be set to the DTS value of the first video
packet. However, when b-frames are used, the first DTS value will be
negative. This was originally done because FLV muxing requires that the
first packet's DTS start from 0. Unfortunately, this would also
effectively cause the first packet's PTS/DTS value to be shifted forward
by the negative amount, which would cause video sync to be off by a
video frame or two.
This fixes it to start at the PTS value instead and preserve any
negative offsets. Additionally, the FLV muxing code has been fixed to
ensure that it adjusts the starting video DTS to 0, and now correctly
adjusts the first audio packet's timestamp according to that DTS as well
(which it didn't do before).
On audio encoder startup, audio encoders paired with a video encoder
would unintentionally discard a single audio data segment, causing it to
be 1024 audio frames out of sync.
Certain PSAPI functions moved to KERNEL32 after Vista, setting
PSAPI_VERSION 1 uses the old PSAPI exports for compatibility. Vista
support was unintentionally broken by a6318ff.
With certain combinations of CMake (<3.9) and ImageMagick (>=7),
configure will silently fail setting up support for ImageMagick,
even if the user requested for it, and the build will move on
(as long as FFMPEG_AVCODEC is found). This commit adds a check
that will stop configure in such cases.
Closesjp9000/obs-studio#1018
For those resolutions the last two chroma samples of every other
line would be overwritten by the last chroma samples of the previous
line (depending on sampler used), producing artifacts on the left
edge of the resulting image (e.g. any color present on the right
edge of the image would "bleed" to every other line on
the left edge)
Fixes https://obsproject.com/mantis/view.php?id=508
The toggle_visibility set only by init_hotkeys() call, but used each
time in obs_sceneitem_destroy() call. Because zero hotkey_pair_id is
valid - we need to set item property other than zero here.
Closesjp9000/obs-studio#1000
Instead of logging the relative encoded count (which is susceptible to
integer overflow), log the output frame count instead. If there's an
issue with encoding, it'll show up when all encoding stops regardless.
This reverts commit 4e3e67bb8c.
The way this is handled will erroneously report 0 frames encoded when
frames have actually been properly encoded, which is best avoided.
Additionally, and overflow would be generated for drawn frames where
none occurred before. The encoded value should probably not even be
present in the log for the output due to the way it's handled.
Some plugins are using the OBS_VERSION macro obs-config.h to specify the
current version, which will bake the string in to the plugin, making it
so that if the plugin is not replaced for a patch, it could potentially
have the incorrect version. This makes it so that a plugin/frontend can
get the current version string that's baked in to libobs itself.
When item in the old scene had only Scale Filter, texture renderer
was not created for an item in the new scene. However if the item also
had Crop set, the texture renderer was created.
Now after copying the data, texture renderer is created for the item in
the new scene when needed.
Closesjp9000/obs-studio#969
Scene items would incorrectly have pre-multiplied alpha when they were
scaled with different scale filtering, when they were cropped, or when
the item itself was a scene. This happens because the scene renders the
items to a texture in those cases, and when they are rendered to a
texture the blend function would be the default srcalpha/invsrcalpha
blend function, which would cause alpha to become pre-multiplied in the
texture's result rather than straight alpha.
This changes the behavior to directly copy the color/alpha to the
texture using the one/zero blend function instead, which makes the
resulting texture straight alpha. (Note that you do not want to turn
off the blend mode for the same result because certain sources can have
custom drawing that may rely on blending being available)
Related Issue: https://obsproject.com/mantis/view.php?id=954Closesjp9000/obs-studio#966
This allows the ability for certain types of modules (particularly
scripting-related modules) to initialize extra data when all other
modules have loaded. Because front-ends may wish to have custom
handling for loading modules, the front-end must manually call
obs_post_load_modules after it has completed loading all plug-in
modules.
Closesjp9000/obs-studio#965
For specific types of transitions (stingers in this case), there is no
blending done of the two targets, so wasting GPU resources rendering
them to textures is unnecessary.
Allows the ability to change the output type in case one service
requires a different output type.
NOTE: This should be considered a temporarily yet simple solution to a
specific problem: support for RTMP-like outputs. This will allows
seamless integration of supporting different RTMP-like output types
within the same service. This should probably be replaced with a more
ideal solution later, such as implementing a completely different
service type instead, when time permits.
If size is 0 after popping data from the front or back, set the
start/end points to 0 as well to ensure that any subsequent buffer
pushes start from the beginning of the buffer rather than the middle of
the buffer. Reduces potential unnecessary operations in that case.
Additionally, this fixes a bug with circulebuf_pop_back where if start
position was 0, and all the data was popped off the buffer (equal to the
capacity), the end position would be equal to the original size. As an
example to replicate the bug, push 5, pop 5, then push 10. The
start/end points will be invalid.
Closesjp9000/obs-studio#954
If a source is created with settings, default values would not be set on
data items that already had values set. The check here was supposed to
be a not equal rather than an equal because you're not supposed to be
able to set a default value of one type on an item that already exists
but is of a different type.
This bug would make it so that if a particular setting was removed,
there would be no default value to fall back on, and it would always be
0 or NULL for all values.
It's likely this bug wasn't encountered until now because before now
there had been no reason to clear or remove settings on most things.
I believe the issue with the next to impossible frame count to be an integer underflow, as in order to achieve those you'd have to have recorded for at least 345 days with 144 fps. So this commit fixes them by using a normal integer first and then deciding on the result if it should be used or be replaced with a 0.
Rename "Processor" to "CPU Name" and place it before the number of
cores. Log free RAM for *nix. Log CPU Speed on Linux and FreeBSD. Unify
processor logging on Linux and FreeBSD. Get physical and logical cores
on *nix.
Use the new os_get_logical_cores and os_get_physical_cores function from
commits 6fc74d6 and e4a64f0 to get CPU core counts on *nix systems.
When an output fails to connect and it's already been prematurely
stopped, the event to mark the output as stopped would not be signaled,
causing obs_output_destroy to lock up indefinitely while waiting for the
event to be signaled.
Rather than have the back-end try to determine whether the output can or
cannot stop, allow the stop callback to continue in the plugin either
way and let the plugin itself make that determination.
This fixes a bug where the back-end wouldn't have data active while
connecting, therefore the stop callback wouldn't be called, and once
connected it wouldn't know that it was supposed to stop. In other words
trying to call obs_output_stop on an output that was in a state of
connecting would do nothing and the output would never stop.
Fixes a potential issue where copying filters from one source to another
might add filters from the old source that are not compatible with the
new source.
