Well, linear SRGB for screen capture. The window capture path failed to
copy between SRGB textures for some reason, so just force nonlinear
formats instead.
For game capture, neither GL nor D3D9 support SRGB shared textures, so
disable linear SRGB support if the texture format doesn't support it.
Similarly, DXGI display capture doesn't work with SRGB at the moment.
Unsure if it will with more work, but disable for now.
Also force linear SRGB off if using GDI-compatible textures.
Both image and color sources have been updated.
Also added alpha support to the color source. Useful for users, and
serves as an easy alpha blending test case.
GS_RGBA, GS_BGRX, and GS_BGRA now use TYPELESS DXGI formats, so we can
alias them between UNORM and UNORM_SRGB as necessary. GS_RGBA_UNORM,
GS_BGRX_UNORM, and GS_BGRA_UNORM have been added to support straight
UNORM types, which Windows requires for sharing textures from D3D9 and
OpenGL. The D3D path aliases via views, and GL aliases via
GL_EXT_texture_sRGB_decode/GL_FRAMEBUFFER_SRGB.
A significant amount of code has changed in the D3D/GL backends, but the
concepts are simple. On the D3D side, we need separate SRVs and RTVs to
support nonlinear/linear reads and writes. On the GL side, we need to
set the proper GL parameters to emulate the same.
Add gs_enable_framebuffer_srgb/gs_framebuffer_srgb_enabled to set/get
the framebuffer as SRGB or not.
Add gs_linear_srgb_active/gs_set_linear_srgb to instruct sources that
they should render as SRGB. Legacy sources can ignore this setting
without regression.
Update obs_source_draw to use linear SRGB as needed.
Update render_filter_tex to use linear SRGB as needed.
Add gs_effect_set_texture_srgb next to gs_effect_set_texture to set
texture with SRGB view instead.
Add SRGB helpers for vec4 struct.
Create GDI-compatible textures without SRGB support. Doesn't seem to
work with SRGB formats.
This function determines if something is an OBS plugin before attempting
to load it. On Windows, many plugins ship their dependent DLLs alongside
the plugin DLL, so OBS would load things like libcef.dll on startup only
to immediately free it. For other platforms, this is less of a concern
so this function is a no-op for now.
This improves startup time and reduces risk from dependent DLLs
potentially running code with unwanted side effects in DllMain.
This is a simple, isolated workflow that generates Flatpak
bundles when running on the master branch, or when a pull
request has the "Seeking Testers" label.
Based on https://github.com/marketplace/actions/flatpak-builder
To support FTL, it needed to be present in-tree to be compiled. This PR
adds support for system-wide installations of libftl. It uses
pkg-config to find the system-wide installation. If pkg-config can't
provide libftl we just fall back to using the in-tree submodule. If
that's also not available it won't be included at all like before.
Functionality is coming in a future Windows update to disable the yellow
border when using WGC. Add code now to opt in. Will require SDK upgrade
later for OBS build.
winrt::apartment_type::multi_threaded is necessary to dodge assert for
calling get() on RequestAccessAsync result. Don't think I will ever
fully understand COM apartments.
This adds functions to the frontend api to start/stop the virtual
camera, to check if it is active and adds function to get the output
reference. It also adds api events for when the virtual camera is
started or stopped.
Fixes#4031
It seems Qt on various flavours of Linux behaves erratically when scaled,
and for the moment macOS doesn't require fractional scaling support.
This commit can be reverted when Qt fixes issues on their side.
When a media source is set to use a URL, always fully shut down the
media when stopped. This ensures that the media will do a full
reconnect after having been stopped.
When connection to FTL ingest is lost, ftl_event() calls
obs_output_signal_stop() to trigger a reconnect. However, during the
reconnect delay, send_thread is still waiting on send_sem semaphore.
After the delay, ftl_stream_start() is called, which in turn resets
the semaphore and creates a new send_thread. Old send_thread now exits
the loop and triggers another reconnect and the whole process repeats
again. The fix resets the semaphore in ftl_event() so the old
send_thread exits immediately.
Also remove use of deprecated [NSProcessInfo operatingSystem]. It always
returns the NSMACHOperatingSystem.
Also stop parsing [NSProcessInfo operatingSystemVersionString]. The
documentation explicitly says "not for parsing".
This commit prevents a crash in, for example, Linux configurations
where PulseAudio is disabled, which leads into 'mod' being 'null',
which in turn leads into a segmentation fault when an ALSA source
is selected.
Closes https://github.com/obsproject/obs-studio/issues/3485
The capture loop errors are now logged as such.
Increased the amount of information logged in debug mode, including
state of all buffers on timeout.
Add device name to all capture loop debug output, especially useful when
debugging issues with multiple v4l2 video streams.
