The GStreamer team has released a new bug-fix update, 1.26.5, of its cross-platform open-source multimedia framework for the stable 1.26 series.
This release tackles a mix of regressions, crashes, and performance hiccups across different components. Some of the most noticeable fixes include:
- Audioconvert: Patched a caps negotiation regression when using a mix matrix—something that might’ve been tripping up certain audio workflows.
- AWS integrations: The awstranslate element now supports brevity mode, while awstranscriber2 lets users partition speaker outputs in transcriptions—handy for clearer multi-speaker logs.
- Speech-to-text: The speechmatics plugin now exposes a mask-profanities property, giving devs more control over filtered outputs.
- V4L2 & VA encoders: Squashed a memory leak during dynamic resolution changes and tightened up VA encoder stability.
- WebRTC: Added a WHEP client signaller and introduced the whepclientsrc element, expanding streaming possibilities.
- Threadshare & RTP: Multiple improvements to threadshare elements and rtpbin2, making real-time media handling more reliable.
Apart from those, GStreamer 1.26.5 brings some smaller but important fixes—Vulkan integration tweaks, GPU memory buffer support for overlays, and decodebin3/uridecodebin3 crash patches.
For more information, see the announcement. Binaries for Android, iOS, macOS, and Windows are expected to be available soon.
