Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning, specifically the variant used by Rust.

[0.10.3] - 2023-03-02

tracers: queue_levels tracer now also supports printing the appsrc levels.
webrtc: webrtcsink can use nvvidconv if nvvideoconvert does not exist on an NVIDIA platform.

gtk4: Set the sync point on the video frame after mapping it as otherwise the frame might not be ready yet for further usage.
livesync: Correctly calculate the fallback buffer duration from the video framerate.
ndi: Handle caps changes correctly in ndisinkcombiner.

hlssink3: Allow signal handlers to return None
gtk4: Make GL context sharing more reliable in pipelines with multiple gtk4paintablesinks
gtk4: Attach channel receiver to the main context from the correct thread to make it possible to start the sink from a different thread than the main thread without having retrieved the paintable from the main thread before.
fmp4mux/mp4mux: Ignore caps changes if only the framerate changes.

gtk4: Simplify and refactor GL context sharing. Apart from being more reliable this reduces GL resource usage.

rtpav1pay: Fix calculation of Leb128 size size to work correctly with streams from certain encoders.

audiornnoise: Use correct value range for the samples
awss3sink: Treat stopping without EOS as an error for multipart upload
awss3hlssink: Fix the name of the hlssink child element
awss3hlssink: Fix deadlock on EOS
dav1d: Various fixes to improve performance, to handle decoding errors more gracefully and to make sure all frames are output in the end
fmp4mux: Various fixes to fragment splitting behaviour, output formatting and header generation
gtk4: Various stability and rendering fixes
meson: Various fixes and improvements to the meson-based build system
ndi: provide non-Linux/macOS UNIX fallback for the soname
ndisrc: Use default channel mask for audio output to allow >2 channels to work better
rav1e: Correctly enable threading support
rtpav1: Various fixes to the payloader and depayloader to handle streams more correctly and to handle errors more cleanly
rtpav1depay: Set caps on the source pad
spotify: fix "start a runtime from within a runtime" with static link
textahead: fix previous buffers
textwrap: Don't panic on empty buffers
tttocea608: Don't fail if a GAP event contains no duration
webrtchttp: whipsink: construct TURN URL correctly
webrtcsink: fix panic on pre-bwe request error
whipsink: Send ICE candidates together with the offer
whipsink: Various cleanups and minor fixes

audiornnoise: Add voice detection threshold property
awss3hlssink: Add stats property
awss3sink: Add properties to set Content-Type and Content-Disposition
fmp4mux: add 'offset-to-zero' property
fmp4mux/mp4mux: add support for muxing Opus, VP8, VP9 and AV1 streams
fmp4mux/mp4mux: Make media/track timescales configurable
fmp4mux: Add support for CMAF-style chunking, e.g. low-latency / LL HLS and DASH
gtk4: Support for rendering GL textures on X11/EGL, X11/GLX, Wayland and macOS
hlssink3: Allow generating i-frame-only playlist
livesync: New element that alllows maintaining a contiguous live stream without gaps from a potentially unstable source.
mp4mux: New non-fragmented MP4 muxer element
spotifyaudiosrc: Support configurable bitrate
textahead: add settings to display previous buffers
threadshare: Introduce new ts-audiotestsrc
webrtcsink: Support nvv4l2vp9enc
whepsource: Add a WebRTC WHEP source element

audiofx: Derive from AudioFilter where possible
dav1ddec: Lower rank to primary to allow usage of hardware decoders with higher ranks
fmp4mux: Only push fragment_offset if write-mfra is true to reduce memory usage
webrtcsink: Make the turn-server property a turn-servers list
webrtcsink: Move from async-std to tokio