The arrival de GStreamer 1.28 This marks a new step forward for one of the most widely used open-source multimedia frameworks in desktop and server environments, especially on GNU/Linux systems. The new stable release introduces significant changes in audio and video processing, analytics, and integration with other graphics and hardware technologies.
In this version, the project reinforces its commitment to performance, security and support for new architecturesWith an increasingly prominent role for the Rust language and improvements designed for professional workflows in streaming, encoding, and content analysis, GStreamer has a global reach. Many of these new features are particularly interesting for European developers and companies that base their multimedia platforms on Linux.
GStreamer 1.28: New stable series and focus on the open ecosystem
GStreamer 1.28 is presented as the latest stable branch of the multimedia frameworkFocused on managing audio and video workflows, from capture to playback and advanced processing, this edition consolidates the work of recent versions and continues to refine support for standards and codecs used in streaming, videoconferencing, and audiovisual production platforms.
The source code for GStreamer 1.28 is available through GitLab from FreeDesktop.org and the project's official websiteThis facilitates their integration into European Linux distributions and enterprise solutions that require custom builds. As usual, the pre-compiled packages will gradually be added to the stable repositories of major distributions.
Boost to Rust to improve security and reliability from GStreamer 1.28
One of the key aspects of this version is the expansion of the use of Rust to implement critical elements of the framework. The goal is to leverage the memory safety guarantees of this language, reducing the risk of classic errors such as out-of-range accesses or memory leaks in high-performance scenarios.
Among the new components in Rust, one stands out. audio source separation element based on demucswhich allows you to isolate voices and instruments in a mix, which can be useful for both musical applications and audio processing in research and analysis.
A YOLOX inference element and a YOLOX tensor decoderThese components, also written in Rust, are geared towards computer vision tasks such as real-time object detection within video streams. They align with the growing use of GStreamer as a platform for channeling data into artificial intelligence models.
This block is completed by a New GIF decoder with loop support and an element icecastsink with AAC compatibility, both implemented in Rust. This strengthens GStreamer's role in audio streaming and lightweight web content scenarios.
Video improvements: Vulkan, AMD HIP and next-generation codecs
In the video section, GStreamer 1.28 incorporates a H.264 encoder over VulkanThis opens the door to better leveraging the graphics capabilities of compatible GPUs for accelerated encoding. This type of advancement is especially useful for applications that require low latency, such as live streaming or video conferencing.
The update also adds a plugin based on AMD HIP (Heterogeneous-compute Interface for Portability)AMD's computing platform aims to offer an open alternative to other heterogeneous computing ecosystems. This benefits those in Europe who use AMD hardware in data centers or workstations for intensive multimedia processing.
Another key aspect is the arrival of Decoding and encoding support for LCEVC over H.265 and H.266GStreamer 1.28 can automatically configure the decoding of LCEVC H.265/H.266 video streams and adds specific encoders for these formats. LCEVC (Low Complexity Enhancement Video Coding) aims to improve compression efficiency while maintaining moderate resource consumption, a key factor for large-scale streaming services.
Along with these new features, the Matroska demultiplexer is now capable of manage uncompressed 4K video, and several problems in the MP4 demultiplexer have been corrected (qtdemux), which should translate into more robust reproduction and processing of high-resolution content.
Integration with Wayland, WebKit and modern graphics
For modern desktop environments, GStreamer 1.28 incorporates basic colorimetry support in WaylandThis is an important step towards ensuring more accurate color reproduction on systems that have already migrated from X11. This is especially relevant in European distributions that have adopted Wayland as their default graphics server.
The version also introduces a new plugin wpe2 for WebKit It utilizes the "WPE Platform API," designed for integrations with embedded browsers and applications that need to render web content alongside multimedia streams. This type of combination is common in connected devices, digital signage, and OTT platforms.
In the field of 3D graphics, an element appears d3d12remap for Direct3D 12, primarily geared towards Windows environments, and adds the ability to directly pass a QQuickItem to the QML rendering tree, which improves integration with Qt-based applications that rely on GStreamer for content playback.
Advanced audio, ALSA and new MPEG-H plugin
In addition to source separation with demucs, the project introduces a New MPEG-H audio decoding plugin with MP4 demultiplexing supportMPEG-H is gaining traction in immersive audio and broadcast scenarios, so its presence in GStreamer can facilitate testing and deployments in European production chains.
For Linux systems, support is added for List PCM virtual devices in ALSAThis simplifies the management of multiple output destinations and complex audio configurations. Icecast support has also been improved with the new element. icecastsink and AAC compatibility, useful for online radio stations and audio streaming platforms.
Analytics, AI, and new inference capabilities
GStreamer 1.28 strengthens its analytics capabilities by adding a generic classification tensor-decoder and a new face detector, geared towards AI scenarios where it is necessary to interpret the results of machine learning models in real time.
The version also includes a LiteRT inference element and a more user-friendly analytics API, designed to facilitate the development of applications that combine video streams with object detection, facial recognition, or statistical content analysis.
As a complement, a Deepgram speech-to-text transcription pluginThis allows you to transform spoken audio into text using speech recognition services. This type of integration may be of interest to companies that offer automatic subtitling, call analysis, or indexing of audiovisual content.
GStreamer 1.28 introduces improvements in playback, formats and interoperability
The playback section has been added GstPlay support for gapless loopingThis is especially appreciated in musical contexts, interactive installations, and looped playback where skipping between tracks or repetitions are annoying.
The JPEG parser now corrects the image processing with HDR gain mapsThis facilitates more accurate handling of photographs and content with high dynamic range. Additionally, a property has been introduced. unifxfdsink which allows copying data to make this element compatible with a larger number of upstream elements in the pipeline.
GStreamer 1.28 also adds support for GstReferenceTimestampMeta, capable of carrying additional information by timestamp, facilitating use cases where precise metadata associated with each frame or audio block is needed.
Debugging tools, traces, and bindings
In the area of developer tools, a new tracer hook which helps to track when buffers are queued and dequeued in pools, something crucial for diagnosing bottlenecks and performance problems in complex pipelines.
Pipeline graphics files (pipeline graph dot filesThey can now include information about active tracers, offering a more complete view of the application's internal state during execution. This makes it easier to understand system behavior, especially in demanding production environments.
Utility gst-inspect-1.0 gains the ability to Print type information for caps fieldsThis helps to streamline capability negotiations between elements. Additionally, a wrapper is introduced. Gst.Float for Python bindings, improving the framework's interoperability with projects written in this language so widely used in the European data science and prototyping ecosystem.
GStreamer 1.28 introduces fixes, revised properties, and general polishing
Beyond the new features, the team has recovered the property "new-pref" to provide better control when cap changes occur, reducing unexpected behavior in format transitions within the same pipeline.
Work has also been done on the implementation of a new API GstLogContextDesigned to mitigate log "spam" across various components and provide a cleaner, more structured logging system, something administrators and developers will appreciate in large-scale deployments.
Windows systems have been updated WASAPI2 with IMMDevice-based device selectionoffering finer control over which audio devices are used. While not a Europe-specific improvement, it does help GStreamer maintain a strong and consistent presence across different platforms.
Overall, GStreamer 1.28 consolidates the project as a versatile multimedia platform ready for modern workloadsCombining internal optimizations, new AI capabilities, improved support for current graphics hardware, and more refined integration with technologies like Wayland, WebKit, and ALSA, this update is particularly attractive for development teams and companies operating multimedia services in Spain and the rest of Europe. Its focus on security through Rust and enhanced debugging and analytics tools make it a particularly compelling upgrade.
