Due to an oversight in enc-amf with strcmp, the code currently causes
certain weirdness to appear. This fixes it so it will use "High"
Profile instead of attempting to use "Constrained Baseline" and also
changes it to actually set things up for "Constant Bitrate" instead of
"Variable Bitrate (Latency Constrained)". Finally I've added a slight
change that always makes it set the B-Frame Pattern to 0 in order to
work better with VCE2.0 and VCE3.0 cards.
Closesjp9000/obs-studio#978
If the service settings change and the output type changes, the previous
signals would attempt to disconnect after the output has been destroyed,
and subsequently that would cause a crash.
Allows an output to automatically specify which audio codec it requires
and use it instead of AAC. This change is intended to be
behind-the-scenes and seamless to the user.
Allows a service to specify the output it needs in order to function as
configured.
NOTE: This functionality should be considered temporary as a seamless
means of implementing support for different output types within the same
service. Ideally, different services should be used for this
functionality.
Due to the default value assigned to AMF.H264.VBVBuffer.Strictness and AMF.H264.Version, the config upgrade code ran every time someone tried to use simple mode. That caused the AMF.H264.VBVBuffer.Strictness value to always be stuck at 100, which is wrong for streaming and recording.
This is mainly to give visual feedback to those affected by NVENC not
working with older driver versions. Currenlty obs fails silently which
could go unnoticed for users who are using hotkeys as well as confuse
users who are not trained to read their logs when issues occur.
Closesjp9000/obs-studio#788
Adds the ability to use a specific prefix and/or suffix on replay buffer
filenames to distinguish them from normal recordings. Defaults to
having the prefix "Replay".
Replay buffer and recording should be separate in case the user wants to
start recording from a specific point rather being forced to reconfigure
for regular recording.
Creates a new button on the main window below the recording button for
turning on/off the replay buffer.
If the user hits the "stop stream" button it'll transition in to
"stopping stream..." and will continue to output until the stream
reaches the timing in which "stop" was pressed.
However, if there is significant congestion, stopping the stream can
take far longer than the user may like. So there needs to be an option
to forcibly stop the stream in that case; pushing the "stop" button a
second time should allow the user to tell the stream/recording to stop
right away instead of waiting for the precise stop timing.
Richard Stanway
Michael Fabian Dirks
jp9000
derrod
Michael Fabian Dirks