When obs_get_source_properties is called, it calls the property modified
callbacks without a source instantiation. The callbacks set in
.get_properties for display capture and window capture would then result
in exceptions when anything is dereferenced on the source, such as
wgc_supported or update_mutex, because the source itself is null. Let's
make the callbacks return early if the property param is null.
We shouldn't have to check WGC support for every display capture or
window capture source. Either the system supports it or it doesn't.
Check when the plugin loads and use that result.
The device type (OBS renderer) cannot change without restarting OBS, so
we should only have to check if we're using D3D11 once instead of
checking every time a new display capture or window capture source is
created.
Among the systems we officially support, the oldest Qt version is Qt 5.9
on Ubuntu 18.04. Fractional scaling is supported in Qt 5.6 and newer. We
should be able to safely remove these ifdefs.
Follow up to PR #3988 and commit
5cdd084c7f946e385748e96306223cd1f1d05b3e.
Without this change, displays listed as projector targets on systems
using a fractional scaling factor (e.g., 125%, 150%, 175%) will show an
incorrect width and height for their size. For a display with 125%
scaling, OBS would show 1536x864 instead of 1920x1080. With this change,
it will show 1920x1080.
The cursor bitmap is centered on the hotspot, so not accounting
for it means PipeWire captures were positioning the cursor sprite
slightly off.
Properly account for the hotspot by subtracting it from the cursor
position.
Related: https://github.com/obsproject/obs-studio/issues/4766
The Windows uninstaller was not deleting the desktop shortcut because
the shell context was incorrectly set to "current" for it, causing it to
try to delete in the incorrect location.
This crash was caused by the fact that the SourceToolbar::oldData member
variable was not being released correctly. To release an OBSData object,
it must be either destructed, set to a new value, or set to a null
value.
Before OBS Studio 27, we used our own hard-coded internal encoder IDs on
for the mac-vth264 plugin: vt_h264_hw and vt_h264_sw. As of OBS Studio
27, we changed to using the encoder IDs that Apple's API reports.
Specifically, this behavior occurs due to PR #4105 and commit
6a9f25c8ea1b244b4a439667be33f4a5cc6cdf4b.
Unfortunately, this change in encoder IDs failed to take into account
existing user configurations, resulting in users with broken encoder
settings. This change attempts to gracefully upgrade user configurations
to handle the changed encoder IDs.
Fixes#4799.
Only use lossless encode if the capability is supported.
Set qpPrimeYZeroTransformBypassFlag to 1 for lossless.
Do not set profileGUID for lossless.
For both NVENC implementations, retry with a heavier reset because both
are unable to recover from failure lightly.
For new NVENC, warn if PVT is requested, but unsupported by the GPU.
Also retry without PVT on failure to try to catch bad cominbations.
For old NVENC, if PVT is enabled when FFmpeg failure occurs, retry
without PVT.
When you right-click on a source in the Sources list that does not have
any filters, the "Copy Filters" item is disabled. When you right-click
on a scene in the Scenes list that does not have any filters, the "Copy
Filters" item is enabled. This fixes the Scenes context menu to behave
like the Sources context menu.
When you right-click on a source in the Sources list that does not have
any filters, the "Copy Filters" item is disabled. When you right-click
on an audio source in the Audio Mixer that does not have any filters,
the "Copy Filters" item is enabled. This fixes the Audio Mixer context
menu to behave like the Sources context menu.
Fixes#4790.
Prevents situations where the redistributable is installed and OBS
enables the RTX denoiser but it is non-functional. Changes were tested
on systems with both supported / unsupported GPUs and it adds around
10-20ms to the load time in both cases.
SourceSyncChanged() would call syncOffsetChanged() because it's
connected to the widget. This is fixed by blocking signals
This bug was discovered because undo/redo actions for sync offset
changes were triggering the creation of new undo/redo actions within
their own callbacks.
Fixesobsproject/obs-studio#4760
Because it's not currently possible to guarantee synchronization between
two separate media files (yet), disable separated track matte media
files for now. It'll just result in support requests that can't be
solved.
