Fix webrtc multitrack audio#14
Open
FrantaBOT wants to merge 1086 commits intowebrtc-multitrack-audiofrom
Open
Conversation
Collaborator
FrantaBOT
commented
Aug 18, 2025
FrantaBOT
commented
- Rebased to main
- Made changes according to obs mantainer
For the background blur effect, it was found that destroying it and then adding any other NVIDIA effect led to a blank picture. This commit swaps the order of destroying the effect and the associated cudaStream. For other effects the order does not matter; I have no idea why the order matters for blur effect and the SDK gives no indication about any proper order at destruction time. Fixes bug obsproject#11383. Signed-off-by: pkv <pkv@obsproject.com>
This properly allocates the temporal state variable used in AI greenscreen effect, when the latter is reset. Signed-off-by: pkv <pkv@obsproject.com>
During the mapping and unmapping of resources, the CudaStream was not always consistent. This fixes the issue. Thanks to Stephan from NVIDIA Dev support for noticing the mismatch. Signed-off-by: pkv <pkv@obsproject.com>
The reset signal was triggered whenever there was an update in the parent source. This destroys the effect and recreates it, which can be resource heavy. We now let the SDK handle the source update on its own rather than manually resetting the effect on obs side. Signed-off-by: pkv <pkv@obsproject.com>
On initial run of the Blur filter, some parameters are not all set when there is the first call to load the effect. This is actually not an issue, so the log is now silenced at this time. Signed-off-by: pkv <pkv@obsproject.com>
46adc4b - Check source validity before attempting to log renderer crash ac34d8e - Don't loop Cef exit while shutting down if task post fails b56fd78 - Update version to 2.24.6
By adding the workflow_dispatch trigger, the scheduled workflow can also be manually triggered if the scheduled run failed to run due to scheduling issues or because of an error that might have been fixed since then.
To emit frontend events early during OBS' initialization, the disableSaving state variable needs to be "falsy" (as events will not be emitted otherwise). The code path taken for generating a new fallback scene collection has disableSaving decremented to "enable" saving (and thus frontend events) already, so it's just the code paths for existing collections that need this workaround. Emitting the scene changed and preview scene changed events when a fallback scene collection has been created has side-effects when OBS Studio is set to studio mode and can lead to potential crashes, as the studio mode is still in an invalid state at this point.
Fix a possible crash when opening OBS in Studio Mode.
This makes all list widget spacing set to 1.
obs_get_transition_by_name is very problematic because transitions are always private and their names aren't unique. This means that the method iterates over all private sources and then takes the first that both is a transition and matches the name we're looking for. However, this could be from anywhere - it could be a frontend transition, but also a source transition, quick transition, or even one from third-party plugins. This is always never what is intended. As such, this method (which should never have been added in the first place) needs to go. In its place, obs_frontend_get_transitions returns a list of all frontend transitions (which is usually what people are looking for), and alternatively obs_get_source_by_uuid also provides access to private sources. While we're at it, obs_get_transition_by_uuid is basically a wrapper for obs_get_source_by_uuid and not really necessary. UUID's are unique and a source doesn't suddenly change its type, so if you have a transition's UUID you can be pretty sure that when you do obs_get_source_by_uuid, it will still be a transition.
Add graphics APIs to obtain the GPU driver version and renderer strings, as well as GDDR memory sizes. The GDDR memory sizes include the dmem (dedicated memory on the GPU) and smem (shared CPU memory used by the GPU but resides in the CPU DDR). The version and renderer strings are needed for identification purposes, for example enhanced broadcasting used by Twitch, to associate the GPU used by OBS with the PCIe-based identification values such as device_id and vendor_id.
`os_get_free_size()` was simply returning 0. For Linux, implement the free size calculation based on the `sysinfo()` system call.
`get_rc_mode()` compares the incoming rate control mode string to the .name member of rc_mode_t, and the table entries are all upper-case. This caused a crash when the incoming string is set to "cbr" instead of "CBR". Make the string comparison case insensitive.
VAAPI encoders deviate from other encoders (e.g. AMF, NVENC) with the "profile" setting being an integer instead of a string. With enhanced broadcasting, "profile" is signalled as a string. Convert the string-based profile to the appropriate integer-based profile for VAAPI encoders as a workaround, until VAAPI supports string-based "profile" (if ever).
Enhanced broadcasting requires system information to be gathered on the client and submitted to the GetClientConfiguration request in order to obtain a valid response from the server. This commit adds support for gathering the required information on Linux-based systems.
Allos creating additional views that act like a MAIN_VIEW instead of AUX_VIEW.
Signals when a video reset has completed. This is mainly useful when canvases or views are used, whose mixes will have been freed or recreated during the reset.
This adds a new obs_canvas object that acts as a shareable (reference-counted) owner of views, mixes, and (optionally) scenes. This is a step towards faciliatating multi-canvas and multi-output features in OBS Studio. It solves a number of complications that exist with the manual approach of using views, such as audio mixing, source active-state tracking, and scenes not havinga reliable way of identifying the actual available canvas size.
The gh CLI utility will fail without a git checkout.
Canvas is a move-only type without a copy constructor. Since C++17 requires types stored in std::optional to be copy-constructible unless explicitly allowed, Clang 20+ emits an error when attempting to instantiate std::optional<Canvas>. While GCC allows this as an extension, Clang enforces the standard more strictly. This PR replaces std::optional<Canvas> with std::unique_ptr<Canvas> to resolve the build error with Clang 20+ while keeping functional behavior identical. Tested with GCC 15.1.0 and Clang 20.1.7. Co-Authored-By: Ryan Foster <ryan@obsproject.com>
The generate-docs action started failing, presumably because it is using Debian Buster. Update the pinned commit for the third-party totaldebug/sphinx-publish-action to hopefully fix this.
v4l2-decoder: continue sending frame to codec if avcodec_receive_frame() returns AVERROR(EAGAIN)
Move 3 headers out from the Linux-specific section. Without these headers functions like open, close, drmXx aren't found. * open requires <fcntl.h> * close requres <unistd.h> * drmXx requires <xf86drm.h>
Subsequent commits use g_clear_fd() added in GLib 2.76.
PipeWire format renegotiation runs in parallel with video rendering. When the stream format is renegotiated, PipeWire removes the existing buffers and closes the syncobj file descriptors. At the same time, the video rendering thread may try to import the (already closed) syncobj acquire fd, and hang on waiting for the fence to become available. This is not a problem for the dmabuf fd because it's imported into a texture right away, which doesn't disappear when PipeWire closes the fd. This commit adds duping to the syncobj fds so that they too remain open as long as the rendering thread may access them.
Since video rendering happens on a separate thread from PipeWire buffer ingestion, it may happen that two buffers are ingested in quick succession without the rendering thread ever getting to them. Since, prior to this commit, the release sync point is only "primed" (set to signal in the future) only on the video rendering thread, this results in the buffer being returned to PipeWire's pool without anything ever signaling the release point, effectively blocking it from ever getting reused in the future. This quickly clogs up the buffer pool and leaves only one buffer to circulate between the screencast source and OBS. This commit adds a flag tracking whether the release point had been primed. If, when ingesting a new PW buffer, the old buffer's release point hadn't been primed, that means the video rendering thread never got to that buffer, so the release point is immediately signaled, marking the buffer reusable by the screencast source.
GitHub introduced a policy change on 2024-05-06 which deprecated support for SARIF files that contain multiple runs for the same tool.
This reverts commit 4dfd0b3. Also adds a FIXME comment.
This fixes a change in PR 11605 that breaks Decklink video outputs. get_mix_for_video() returns NULL for those outputs, causing start_raw_video() to return before connecting the output.
Collaborator
Author
|
Hello, |
Co-authored-by: Sean DuBois <sean@siobud.com>
Co-authored-by: Sean DuBois <sean@siobud.com>
FrantaBOT
force-pushed
the
webrtc-multitrack-audio-2
branch
from
19432ce to
33be0b5
Compare
Collaborator
Author
|
@Sean-Der Hello, sorry for pinging you. Did you get a chance to review this? |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.