Code submissions have continually suffered from formatting
inconsistencies that constantly have to be addressed. Using
clang-format simplifies this by making code formatting more consistent,
and allows automation of the code formatting so that maintainers can
focus more on the code itself instead of code formatting.
Remove KSAUDIO_SPEAKER_2POINT1 because it is included in the Windows
SDK. (I'm guessing it wasn't before.)
Redefine 4POINT1 as KSAUDIO_SPEAKER_SURROUND|SPEAKER_LOW_FREQUENCY
instead of KSAUDIO_SPEAKER_QUAD|SPEAKER_LOW_FREQUENCY. See definitions:
FFmpeg 4.1: FL+FR+FC+LFE+BC
WASAPI QUAD: FL+FR+BL+BR (looks wrong)
WASAPI SURROUND: FL+FR+FC+BC
WASAPI LOW_FREQUENCY: LFE
(This commit also modifies the UI, obs-ffmpeg, and obs-output modules)
Fixes a long-time regression where the program would lock up if an
encode call fails. Shuts down all outputs associated with the failing
encoder and displays an error message to the user.
Ideally, it would be best if a more detailed error could be displayed to
the user about the nature of the error, though the primary problem is
the encoder errors are typically not something the user would be able to
understand. The current message is a bit of a generic error message;
improvement is welcome.
Another suggestion is to try to have the encoder restart seamlessly,
though it would take a significant amount of work to be able to make it
do something like that properly, and it sort of assumes that encoder
failures are sporadic, which may not necessarily be the case with some
hardware encoders on some systems. It may be better just to use another
encoder in that case. For now, seamless restart is ruled out.
This patch fixes QSV failing with new DCH graphics drivers. QSV is not
initializing under certain conditions when the graphics adapter 0 is not
set to iGFX due to outdated MSDK dispatcher. Updating to MSDK with new
dispatcher to enable proper initialization of QSV encoder with DCH
drivers.
There are cases where alpha is multiplied unnecessarily. This change
attempts to use premultiplied alpha blending for composition.
To keep this change simple, The filter chain will continue to use
straight alpha. Otherwise, every source would need to modified to output
premultiplied, and every filter modified for premultiplied input.
"DrawAlphaDivide" shader techniques have been added to convert from
premultiplied alpha to straight alpha for final output. "DrawMatrix"
techniques ignore alpha, so they do not appear to need changing.
One remaining issue is that scale effects are set up here to use the
same shader logic for both scale filters (straight alpha - incorrectly),
and output composition (premultiplied alpha - correctly). A fix could be
made to add additional shaders for straight alpha, but the "real" fix
may be to eliminate the straight alpha path at some point.
For graphics, SrcBlendAlpha and DestBlendAlpha were both ONE, and could
combine together to form alpha values greater than one. This is not as
noticeable of a problem for UNORM targets because the channels are
clamped, but it will likely become a problem in more situations if FLOAT
targets are used.
This change switches DestBlendAlpha to INVSRCALPHA. The blending
behavior of stacked transparents is preserved without overflowing the
alpha channel.
obs-transitions: Use premultiplied alpha blend, and simplify shaders
because both inputs and outputs use premultiplied alpha now.
Fixes https://obsproject.com/mantis/view.php?id=1108
The initial DTS for non-fractional framerates was being incorrectly
calculated. It assumed that the time base was in frames when it was
not.
Closesobsproject/obs-studio#1857
The workarounds were made because of conflicts with running multiple UI
threads at once on macOS, which macOS can't do very well, and would be
susceptible to crashes. This would cause crashes not only on startup
but seemingly at random when using the browser source on macOS. The
original "fix" was a hack to try to minimize UI code and browser UI code
from executing at the same time. The macOS initial scene loading was
deferred until all Qt-related and main window initialization was
completed. Although this worked to some extent to prevent conflicts, it
made it so that there was an initial period on startup where the entire
UI seemed "blank" for users, and it was still possible for the main UI
thread and the browser UI thread to clash, causing crashes seemingly at
random for users.
The external message pump method of CEF is the solution to the problem,
which is the method which allows the main UI thread to share events with
CEF. To do this, all CEF operations need to be performed in the UI
thread (Qt's main thread), and CefDoMessageLoopWork() needs to be called
when CefApp::OnScheduleMessagePumpWork callback is triggered. A number
of other issues had to be solved as well, such as CefBrowser references
getting "stuck" in the Qt event queue.
With this, macOS no longer needs to do the "deferred load" hack, and
browsers are now much more stable and no longer as susceptible to
seemingly random crashes, improving overall program stability when
browsers are used.
Instead of having ffmpeg-mux stored in a data directory, install it to
the primary binary directory. On windows, this fixes ffmpeg-mux
potentially accessing the wrong FFmpeg libraries (some programs install
them to system32, foolishly), and instead ensures that it uses the ones
that come with the program. On Linux, ensures that a binary is in its
appropriate directory (/usr/bin or /usr/local/bin rather than a subset
of /usr/share or /usr/local/share).
libobs: Add support for limited to full color range conversions when
using RGB or Y800 formats, and move RGB converison for Y800 formats to
the GPU.
decklink: Stop hiding color space/range properties for RGB formats, and
remove "YUV" from "YUV Color Space" and "YUV Color Range".
win-dshow: Remove "YUV" from "YUV Color Space" and "YUV Color Range".
UI: Remove "YUV" from "YUV Color Space" and "YUV Color Range".
Allows the ability to override and use partial range RGB with the
DirectShow and Decklink device sources when partial range RGB is
implemented. Fixes certain cases where devices could capture RGB in
limited range via HDMI (per the HDMI specs).
This commit fixes a bug where the user would have to restart OBS when
certain types of devices are unplugged and plugged back in again
(particularly USB devices). A wait interval is required
(RECONNECT_INTERVAL) because Windows does not send any events to
IAudioCaptureClient on device invalidation.
Currently SrcBlendAlpha and DestBlendAlpha are both ONE, and can
combine together to form two. This is not a noticeable problem for
UNORM targets because the channels are clamped, but it will likely
become a problem if FLOAT targets are more widely used.
This change switches DestBlendAlpha to INVSRCALPHA, and starts
backgrounds as opaque black instead of transparent black. The blending
behavior of stacked transparents is preserved without overflowing the
alpha channel.
The quadro P5000 would incorrectly be considered blacklisted because it
used a string search for the P500, which is an earlier quadro variant
that does not have NVENC support.
To fix this, instead of just doing a string search, additionally check
to make sure that there preceding or trailing numbers on the adapter
name.
As per mantis issue 1403, the current wording for ClearOnMediaEnd is
ambiguous. Since it says "Hide source", some may believe it to mean the
scene item in the sources list will have it's visibility set to hidden.
This wording makes it clearer what the checkbox actually does.
This new scale filter computes pixels by weighing the coverage area of
source pixels over the target pixel. This algorithm works well for both
upsampling and downsampling, but was mainly designed to upscale
high-quality low-resolution sources like RGB/HDMI retro consoles. I've
heard of people using odd workarounds like scaling up to very high
resolutions before scaling back down to preserve pixel shartpness. This
algorithm directly addresses this use-case in a much more direct
fashion.
The Area scale filter does a better job of preserving the thickness of
thin features than the Point filter.
The Area scale filter does not look at source pixels that lie outside
of the target pixel, leading to a much sharper image than Bilinear,
Bicubic, and Lanczos filters.
This filter should interpolate pixels in linear space, but OBS is not
equipped to do that at the moment.
libobs: Add GPU effect, and wire up scene serialization.
obs-filters: Add Area as an option for scale_filter.
UI: Add Area as an option for both scene items, and canvas downscaling.
This reverts commit 3c22cf35c9, reversing
changes made to c7dab6c92b.
This is being reverted due to many people being unable to start up the
QSV encoder with these changes.
Fixes mantis issue https://obsproject.com/mantis/view.php?id=1379
For some devices with hdmi input, the Front Center channel and the LFE
channel are swapped. For some others they are not.
To solve the issue a new swap setting is added so that the user can
swap the two channels if needed.
MatcherClosestFrameRateSelector updates best_match as a side effect of
visiting every VideoInfo instance, but CapsMatch uses std::any_of,
which will stop iterating prematurely on first match. This means that
the highest FPS is not selected.
This change switches to a for loop that doesn't exit early.
In the Twitch chat dock window, the "Moderation Tools" link in the chat
settings should not navigate the window to the new URL; it should pop up
the URL in the user's system browser instead.
According to nvidia's documentation using b-frames as reference results in
a slight quality improvement with no/negligible performance impact and is
recommended to be enabled by default when using multiple b-frames.
When window capture is not capturing a window, don't try to find the
window every frame. Instead, just check once per frame. The process of
finding a window has a lot of checks and requires a surprising amount of
processing.
Defers the update of the slideshow source until the user has pressed OK
or exited the properties dialog. This ensures images aren't reloaded
every time the user adjusts a setting.
Because the new NVENC implementation requires texture sharing, if the
user chooses to use a secondary GPU, fall back to the old implementation
instead.
Fixes the Profile Level problem with integrated AMD Vega GPUs.
(Jim: Also fixes an issue where the amf-test executable would not build)
Closesobsproject/obs-studio#1619
Adds a texture-based NVENC implementation which passes OBS NV12 output
textures directly to NVENC without downloading them off of the GPU,
increasing NVENC performance by a significant margin.
If NV12 textures are unavailable or the new encoder fails to initialize
for whatever reason, it will fall back to the FFmpeg NVENC
implementation safely.
Adds VBR rate control mode, adds a lookahead option, adds psycho visual
tuning option, removes level property (now always set to 'auto'),
removes "2pass" option (replaced by new "Max Quality" preset which uses
2pass by default), modifies a few defaults, and updates their locale
text.
This implements a blacklist of devices that ship with NVENC libraries,
but do not support NVENC, which would cause the plugin to mistakenly
think that NVENC was available when it was not. If these devices are
the only NVIDIA devices on the system, consider NVENC unavailable.
This is much more efficient than spawning an encoder on startup to see
if NVENC is available, which would incur a ~500 millisecond hit on
startup. Additionally, also much less crash-prone.
Operating systems don't report monitors from 0, so OBS shouldn't
either. This avoids user confusion when display capture doesn't work.
This does not change monitor count internally.
They now provide ingest URLs on demand in their panel.
The ones in our list point to random OVH IPs that are likely assigned to
new customers by now and refuse the connection on TCP port 1935.
The GDI+ based Text Source actually uses Premultiplied Alpha. The edges
of the fonts are therefore incorrectly blended, causing ugly artifacts
especially if bright text above a bright background is used. Here's an
image comparing the new text blending (left) to before (right):
![https://i.imgur.com/VhhkQcZ.png](https://i.imgur.com/VhhkQcZ.png)
Additionally, the game capture has the same problem, so premultiplied
alpha is used there as well now.
A bit of a hack, but this is one of the more common errors that users
are encountering, so showing useful information will help cut down on
the number of support issues.
Input buffer by default is set to 2MB - for some sources, for example streaming
RTSP input over UDP, this is not enough and causes significant playback issues
that are not present while playing back the same source under ffplay/mpv.
It looks like someone actually started working on this feature as there are
translated strings, properties and everything ready - only the control was
missing from the UI. This commit adds that control.
Currently, the range is set to 1-16MB, step 1MB. This is somewhat arbitrary,
so suggestions to tweak this range with more real-world use cases are welcome.