The GStreamer team has released a major update for its cross-platform open-source multimedia framework, v1.26, in which codec support has taken center stage.
Notably, GStreamer 1.26 rolls out H.266 Versatile Video Coding (VVC) capabilities, enabling both software—and hardware-based encoding and decoding for next-generation video content.
This is complemented by JPEG XS (a low-latency, visually lossless codec) support and Low Complexity Enhancement Video Coding (LCEVC) for improving quality on top of existing codecs like H.264.
Simultaneously, GStreamer has significantly expanded closed captioning workflows. From H.264/H.265 caption extraction elements that align captions with re-ordered frames to the brand-new cea708overlay and cea708mux components, these enhancements ensure that GStreamer-based applications can offer more robust accessibility options.
Meanwhile, specialized converters like tttocea708 allow transformations from timed text into closed captions, further broadening caption support for broadcast and online streaming scenarios.
Moreover, users working with adaptive streaming will be delighted by the HLS and DASH improvements. New elements such as hlscmafsink, hlssink3, and hlsmultivariantsink make producing CMAF (fMP4) or MPEG-TS fragments for different variants easier.
The dashsink element also benefits from memory leak fixes and better period handling for more accurate dynamic MPD generation.
For continuous, on-the-fly recordings, splitmuxsrc and splitmuxsink now support lazy loading and dynamic fragment addition, enabling real-time appending of new segments to an active pipeline—especially helpful if you are archiving a live feed while simultaneously playing it back.
Another major focus in this release is analytics. GStreamer’s analytics API gains a new GstTensorMeta for carrying inference data between elements, along with N-to-N relationship support in GstAnalyticsRelationMeta. Furthermore, the new originalbuffer plugin allows developers to restore the original video buffer after performing transformations for analysis.
For teams working with ONVIF metadata, GStreamer now offers dedicated elements to extract and convert ONVIF metadata, ensuring users can integrate standard-compliant metadata flows into their pipelines.
Regarding hardware-accelerated video, the Vulkan integration sees significant gains: improved buffer management, expanded color space selection, and refined encoders for formats like H.264 and H.265.
Additionally, OpenGL receives more versatile color-conversion features, and Qt5/Qt6 QML sinks now facilitate direct DMABuf import from hardware decoders—an especially neat touch for developers aiming to optimize display performance on embedded devices. CUDA users benefit from a brand-new nvav1enc encoder for AV1, plus enhancements like a cudacompositor element and Jetson NVMM memory support.
Microsoft Windows developers can tap into fresh Direct3D12 integration, with new elements such as d3d12swapchainsink and d3d12deinterlace ensuring zero-copy memory-sharing and smoother playback.
Lastly, HDR support, advanced scheduling for Decklink cards, better signal loss handling for AJA capture sources, and new RTP/RTSP synchronization modes. Lastly, GStreamer 1.26 broadens WebRTC functionality through improved rollbacks, more encoder options, and built-in web/signaling servers in the webrtcsink plugin.
For more information, see the announcement.
