+GStreamer 1.18 Release Notes
+GStreamer 1.18.0 was originally released on 7 September 2020.
-GSTREAMER 1.16 RELEASE NOTES
-
-
-GStreamer 1.16.0 was originally released on 19 April 2019.
-
-See https://gstreamer.freedesktop.org/releases/1.16/ for the latest
+See https://gstreamer.freedesktop.org/releases/1.18/ for the latest
version of this document.
-_Last updated: Friday 19 April 2019, 00:00 UTC (log)_
-
+Last updated: Monday 7 September 2020, 10:30 UTC (log)
Introduction
As always, this release is again packed with many new features, bug
fixes and other improvements.
-
Highlights
-- GStreamer WebRTC stack gained support for data channels for
- peer-to-peer communication based on SCTP, BUNDLE support, as well as
- support for multiple TURN servers.
+- GstTranscoder: new high level API for applications to transcode
+ media files from one format to another
-- AV1 video codec support for Matroska and QuickTime/MP4 containers
- and more configuration options and supported input formats for the
- AOMedia AV1 encoder
+- High Dynamic Range (HDR) video information representation and
+ signalling enhancements
-- Support for Closed Captions and other Ancillary Data in video
+- Instant playback rate change support
-- Support for planar (non-interleaved) raw audio
+- Active Format Description (AFD) and Bar Data support
-- GstVideoAggregator, compositor and OpenGL mixer elements are now in
- -base
+- ONVIF trick modes support in both GStreamer RTSP server and client
-- New alternate fields interlace mode where each buffer carries a
- single field
+- Hardware-accelerated video decoding on Windows via DXVA2 /
+ Direct3D11
-- WebM and Matroska ContentEncryption support in the Matroska demuxer
+- Microsoft Media Foundation plugin for video capture and
+ hardware-accelerated video encoding on Windows
-- new WebKit WPE-based web browser source element
+- qmlgloverlay: New overlay element that renders a QtQuick scene over
+ the top of an input video stream
-- Video4Linux: HEVC encoding and decoding, JPEG encoding, and improved
- dmabuf import/export
+- New imagesequencesrc element to easily create a video stream from a
+ sequence of jpeg or png images
-- Hardware-accelerated Nvidia video decoder gained support for VP8/VP9
- decoding, whilst the encoder gained support for H.265/HEVC encoding.
+- dashsink: Add new sink to produce DASH content
-- Many improvements to the Intel Media SDK based hardware-accelerated
- video decoder and encoder plugin (msdk): dmabuf import/export for
- zero-copy integration with other components; VP9 decoding; 10-bit
- HEVC encoding; video post-processing (vpp) support including
- deinterlacing; and the video decoder now handles dynamic resolution
- changes.
+- dvbsubenc: DVB Subtitle encoder element
-- The ASS/SSA subtitle overlay renderer can now handle multiple
- subtitles that overlap in time and will show them on screen
- simultaneously
+- TV broadcast compliant MPEG-TS muxing with constant bitrate muxing
+ and SCTE-35 support
-- The Meson build is now feature-complete (*) and it is now the
- recommended build system on all platforms. The Autotools build is
- scheduled to be removed in the next cycle.
+- rtmp2: new RTMP client source and sink element implementation
-- The GStreamer Rust bindings and Rust plugins module are now
- officially part of upstream GStreamer.
+- svthevcenc: new SVT-HEVC-based H.265 video encoder
-- The GStreamer Editing Services gained a gesdemux element that allows
- directly playing back serialized edit list with playbin or
- (uri)decodebin
+- vaapioverlay compositor element using VA-API
-- Many performance improvements
+- rtpmanager support for Google’s Transport-Wide Congestion Control
+ (twcc) RTP extension
+- splitmuxsink and splitmuxsrc gained support for auxiliary video
+ streams
-Major new features and changes
+- webrtcbin now contains some initial support for renegotiation
+ involving stream addition and removal
-Noteworthy new API
-
-- GstAggregator has a new "min-upstream-latency" property that forces
- a minimum aggregate latency for the input branches of an aggregator.
- This is useful for dynamic pipelines where branches with a higher
- latency might be added later after the pipeline is already up and
- running and where a change in the latency would be disruptive. This
- only applies to the case where at least one of the input branches is
- live though, it won’t force the aggregator into live mode in the
- absence of any live inputs.
-
-- GstBaseSink gained a "processing-deadline" property and
- setter/getter API to configure a processing deadline for live
- pipelines. The processing deadline is the acceptable amount of time
- to process the media in a live pipeline before it reaches the sink.
- This is on top of the systemic latency that is normally reported by
- the latency query. This defaults to 20ms and should make pipelines
- such as v4l2src ! xvimagesink not claim that all frames are late in
- the QoS events. Ideally, this should replace the "max-lateness"
- property for most applications.
-
-- RTCP Extended Reports (XR) parsing according to RFC 3611:
- Loss/Duplicate RLE, Packet Receipt Times, Receiver Reference Time,
- Delay since the last Receiver (DLRR), Statistics Summary, and VoIP
- Metrics reports. This only provides the ability to parse such
- packets, generation of XR packets is not supported yet and XR
- packets are not automatically parsed by rtpbin / rtpsession but must
- be actively handled by the application.
-
-- a new mode for interlaced video was added where each buffer carries
- a single field of interlaced video, with buffer flags indicating
- whether the field is the top field or bottom field. Top and bottom
- fields are expected to alternate in this mode. Caps for this
- interlace mode must also carry a format:Interlaced caps feature to
- ensure backwards compatibility.
-
-- The video library has gained support for three new raw pixel
- formats:
-
- - Y410: packed 4:4:4 YUV, 10 bits per channel
- - Y210: packed 4:2:2 YUV, 10 bits per channel
- - NV12_10LE40: fully-packed 10-bit variant of NV12_10LE32,
- i.e. without the padding bits
-
-- GstRTPSourceMeta is a new meta that can be used to transport
- information about the origin of depayloaded or decoded RTP buffers,
- e.g. when mixing audio from multiple sources into a single stream. A
- new "source-info" property on the RTP depayloader base class
- determines whether depayloaders should put this meta on outgoing
- buffers. Similarly, the same property on RTP payloaders determines
- whether they should use the information from this meta to construct
- the CSRCs list on outgoing RTP buffers.
-
-- gst_sdp_message_from_text() is a convenience constructor to parse
- SDPs from a string which is particularly useful for language
- bindings.
-
-Support for Planar (Non-Interleaved) Raw Audio
-
-Raw audio samples are usually passed around in interleaved form in
-GStreamer, which means that if there are multiple audio channels the
-samples for each channel are interleaved in memory, e.g.
-|LEFT|RIGHT|LEFT|RIGHT|LEFT|RIGHT| for stereo audio. A non-interleaved
-or planar arrangement in memory would look like
-|LEFT|LEFT|LEFT|RIGHT|RIGHT|RIGHT| instead, possibly with
-|LEFT|LEFT|LEFT| and |RIGHT|RIGHT|RIGHT| residing in separate memory
-chunks or separated by some padding.
-
-GStreamer has always had signalling for non-interleaved audio since
-version 1.0, but it was never actually properly implemented in any
-elements. audioconvert would advertise support for it, but wasn’t
-actually able to handle it correctly.
-
-With this release we now have full support for non-interleaved audio as
-well, which means more efficient integration with external APIs that
-handle audio this way, but also more efficient processing of certain
-operations like interleaving multiple 1-channel streams into a
-multi-channel stream which can be done without memory copies now.
-
-New API to support this has been added to the GStreamer Audio support
-library: There is now a new GstAudioMeta which describes how data is
-laid out inside the buffer, and buffers with non-interleaved audio must
-always carry this meta. To access the non-interleaved audio samples you
-must map such buffers with gst_audio_buffer_map() which works much like
-gst_buffer_map() or gst_video_frame_map() in that it will populate a
-little GstAudioBuffer helper structure passed to it with the number of
-samples, the number of planes and pointers to the start of each plane in
-memory. This function can also be used to map interleaved audio buffers
-in which case there will be only one plane of interleaved samples.
-
-Of course support for this has also been implemented in the various
-audio helper and conversion APIs, base classes, and in elements such as
-audioconvert, audioresample, audiotestsrc, audiorate.
-
-Support for Closed Captions and Other Ancillary Data in Video
-
-The video support library has gained support for detecting and
-extracting Ancillary Data from videos as per the SMPTE S291M
-specification, including:
-
-- a VBI (Vertical Blanking Interval) parser that can detect and
- extract Ancillary Data from Vertical Blanking Interval lines of
- component signals. This is currently supported for videos in v210
- and UYVY format.
-
-- a new GstMeta for closed captions: GstVideoCaptionMeta. This
- supports the two types of closed captions, CEA-608 and CEA-708,
- along with the four different ways they can be transported (other
- systems are a superset of those).
-
-- a VBI (Vertical Blanking Interval) encoder for writing ancillary
- data to the Vertical Blanking Interval lines of component signals.
-
-The new closedcaption plugin in gst-plugins-bad then makes use of all
-this new infrastructure and provides the following elements:
-
-- cccombiner: a closed caption combiner that takes a closed captions
- stream and another stream and adds the closed captions as
- GstVideoCaptionMeta to the buffers of the other stream.
-
-- ccextractor: a closed caption extractor which will take
- GstVideoCaptionMeta from input buffers and output them as a separate
- closed captions stream.
-
-- ccconverter: a closed caption converter that can convert between
- different formats
-
-- line21encoder, line21decoder: inject/extract line21 closed captions
- to/from SD video streams
-
-- cc708overlay: decodes CEA 608/708 captions and overlays them on
- video
-
-Additionally, the following elements have also gained Closed Caption
-support:
-
-- qtdemux and qtmux support CEA 608/708 Closed Caption tracks
-
-- mpegvideoparse, h264parse extracts Closed Captions from MPEG-2/H.264
- video streams
-
-- avviddec, avvidenc, x264enc got support for extracting/injecting
- Closed Captions
-
-- decklinkvideosink can output closed captions and decklinkvideosrc
- can extract closed captions
-
-- playbin and playbin3 learned how to autoplug CEA 608/708 CC overlay
- elements
-
-- the externally maintained ajavideosrc element for AJA capture cards
- has support for extracting closed captions
+- New RTP source and sink elements to easily set up RTP streaming via
+ rtp:// URIs
-The rsclosedcaption plugin in the Rust plugins collection includes a
-MacCaption (MCC) file parser and encoder.
+- New Audio Video Transport Protocol (AVTP) plugin for Time-Sensitive
+ Applications
-New Elements
-
-- overlaycomposition: New element that allows applications to draw
- GstVideoOverlayCompositions on a stream. The element will emit the
- "draw" signal for each video buffer, and the application then
- generates an overlay for that frame (or not). This is much more
- performant than e.g. cairooverlay for many use cases, e.g. because
- pixel format conversions can be avoided or the blitting of the
- overlay can be delegated to downstream elements (such as
- gloverlaycompositor). It’s particularly useful for cases where only
- a small section of the video frame should be drawn on.
+- Support for the Video Services Forum’s Reliable Internet Stream
+ Transport (RIST) TR-06-1 Simple Profile
-- gloverlaycompositor: New OpenGL-based compositor element that
- flattens any overlays from GstVideoOverlayCompositionMetas into the
- video stream. This element is also always part of glimagesink.
+- Universal Windows Platform (UWP) support
-- glalpha: New element that adds an alpha channel to a video stream.
- The values of the alpha channel can either be set to a constant or
- can be dynamically calculated via chroma keying. It is similar to
- the existing alpha element but based on OpenGL. Calculations are
- done in floating point so results may not be identical to the output
- of the existing alpha element.
+- rpicamsrc element for capturing from the Raspberry Pi camera
-- rtpfunnel funnels together RTP streams into a single session. Use
- cases include multiplexing and bundle. webrtcbin uses it to
- implement BUNDLE support.
+- RTSP Server TCP interleaved backpressure handling improvements as
+ well as support for Scale/Speed headers
-- testsrcbin is a source element that provides an audio and/or video
- stream and also announces them using the recently-introduced
- GstStream API. This is useful for testing elements such as playbin3
- or uridecodebin3 etc.
+- GStreamer Editing Services gained support for nested timelines,
+ per-clip speed rate control and the OpenTimelineIO format.
-- New closed caption elements: cccombiner, ccextractor, ccconverter,
- line21encoder, line21decoder and cc708overlay (see above)
+- Autotools build system has been removed in favour of Meson
-- wpesrc: new source element acting as a Web Browser based on WebKit
- WPE
+Major new features and changes
-- Two new OpenCV-based elements: cameracalibrate and cameraundistort
- that can communicate to figure out distortion correction parameters
- for a camera and correct for the distortion.
+Noteworthy new features and API
+
+Instant playback rate changes
+
+Changing the playback rate as quickly as possible so far always required
+a flushing seek. This generally works, but has the disadvantage of
+flushing all data from the playback pipeline and requiring the demuxer
+or parser to do a full-blown seek including resetting its internal state
+and resetting the position of the data source. It might also require
+considerable decoding effort to get to the right position to resume
+playback from at the higher rate.
+
+This release adds a new mechanism to achieve quasi-instant rate changes
+in certain playback pipelines without interrupting the flow of data in
+the pipeline. This is activated by sending a seek with the
+GST_SEEK_FLAG_INSTANT_RATE_CHANGE flag and start_type = stop_type =
+GST_SEEK_TYPE_NONE. This flag does not work for all pipelines, in which
+case it is necessary to fall back to sending a full flushing seek to
+change the playback rate. When using this flag, the seek event is only
+allowed to change the current rate and can modify the trickmode flags
+(e.g. keyframe only or not), but it is not possible to change the
+current playback position, playback direction or do a flush.
+
+This is particularly useful for streaming use cases like HLS or DASH
+where the streaming download should not be interrupted when changing
+rate.
+
+Instant rate changing is handled in the pipeline in a specific sequence
+which is detailed in the seeking design docs. Most elements don’t need
+to worry about this, only elements that sync to the clock need some
+special handling which is implemented in the GstBaseSink base class, so
+should be taken care of automatically in most normal playback pipelines
+and sink elements.
+
+See Jan’s GStreamer Conference 2019 talk “Changing Playback Rate
+Instantly” for more information.
+
+You can try this feature by passing the -i command line option to
+gst-play-1.0. It is supported at least by qtdemux, tsdemux, hlsdemux,
+and dashdemux.
+
+Google Transport-Wide Congestion Control
+
+rtpmanager now supports the parsing and generating of RTCP messages for
+the Google Transport-Wide Congestion Control RTP Extension, as described
+in:
+https://tools.ietf.org/html/draft-holmer-rmcat-transport-wide-cc-extensions-01.
+
+This “just” provides the required plumbing/infrastructure, it does not
+actually make effect any actual congestion control on the sender side,
+but rather provides information for applications to use to make such
+decisions.
+
+See Håvard’s “Google Transport-Wide Congestion Control” talk for more
+information about this feature.
+
+GstTranscoder: a new high-level transcoding API for applications
+
+The new GstTranscoder library, along with transcodebin and
+uritranscodebin elements, provides high level API for applications to
+transcode media files from one format to another. Watch Thibault’s talk
+“GstTranscoder: A High Level API to Quickly Implement Transcoding
+Capabilities in your Applications” for more information.
+
+This also comes with a gst-transcoder-1.0 command line utility to
+transcode one URI into another URI based on the specified encoding
+profile.
+
+Active Format Description (AFD) and Bar Data support
+
+The GstVideo Ancillary Data API has gained support for Active Format
+Description (AFD) and Bar data.
+
+This includes various two new buffer metas: GstVideoAFDMeta and
+GstVideoBarMeta.
+
+GStreamer now also parses and extracts AFD/Bar data in the h264/h265
+video parsers, and supports both capturing them and outputting them in
+the decklink elements. See Aaron’s lightning talk at the GStreamer
+Conference for more background.
+
+ONVIF trick modes support in both GStreamer RTSP server and client
+
+- Support for the various trick modes described in section 6 of the
+ ONVIF streaming spec has been implemented in both gst-rtsp-server
+ and rtspsrc.
+- Various new properties in rtspsrc must be set to take advantage of
+ the ONVIF support
+- Examples are available here: test-onvif-server.c and
+ test-onvif-client.c
+- Watch Mathieu Duponchelle’s talk “Implementing a Trickmode Player
+ with ONVIF, RTSP and GStreamer” for more information and a live
+ demo.
+
+GStreamer Codecs library with decoder base classes
+
+This introduces a new library in gst-plugins-bad which contains a set of
+base classes that handle bitstream parsing and state tracking for the
+purpose of decoding different codecs. Currently H264, H265, VP8 and VP9
+are supported. These bases classes are meant primarily for internal use
+in GStreamer and are used in various decoder elements in connection with
+low level decoding APIs like DXVA, NVDEC, VAAPI and V4L2 State Less
+decoders. The new library is named gstreamer-codecs-1.0 /
+libgstcodecs-1.0 and is not yet guaranteed to be API stable across major
+versions.
+
+MPEG-TS muxing improvements
+
+The GStreamer MPEG-TS muxer has seen major improvements on various
+fronts in this cycle:
+
+- It has been ported to the GstAggregator base class which means it
+ can work in defined-latency mode with live input sources and
+ continue streaming if one of the inputs stops producing data.
+
+- atscmux, a new ATSC-specific tsmux subclass
+
+- Constant Bit Rate (CBR) muxing support via the new bitrate property
+ which allows setting the target bitrate in bps. If this is set the
+ muxer will insert null packets as padding to achieve the desired
+ multiplex-wide constant bitrate.
+
+- compliance fixes for TV broadcasting use cases (esp. ATSC). See
+ Jan’s talk “TV Broadcast compliant MPEG-TS” for details.
+
+- Streams can now be added and removed at runtime: Until now, any
+ streams in tsmux had to be present when the element started
+ outputting its first buffer. Now they can appear at any point during
+ the stream, or even disappear and reappear later using the same PID.
+
+- new pcr-interval property allows applications to configure the
+ desired interval instead of hardcoding it
+
+- basic SCTE-35 support. This is enabled by setting the scte-35-pid
+ property on the muxer. Sending SCTE-35 commands is then done by
+ creating the appropriate SCTE-35 GstMpegtsSection and sending them
+ on the muxer.
+
+- MPEG-2 AAC handling improvements
+
+New elements
+
+- New qmlgloverlay element for rendering a QtQuick scene over the top
+ of a video stream. qmlgloverlay requires that Qt support adopting an
+ external OpenGL context and is known to work on X11 and Windows.
+ Wayland is known not to work due to limitations within Qt. Check out
+ the example to see how it works.
-- New sctp plugin based on usrsctp with sctpenc and sctpdec elements.
- These elements are used inside webrtcbin for implementing data
- channels.
+- The clocksync element is a generic element that can be placed in a
+ pipeline to synchronise passing buffers to the clock at that point.
+ This is similar to identity sync=true, but because it isn’t
+ GstBaseTransform-based, it can process GstBufferLists without
+ breaking them into separate GstBuffers. It is also more discoverable
+ than the identity option. Note that you do not need to insert this
+ element into your pipeline to make GStreamer sync to the pipeline
+ clock, this is usually handled automatically by the elements in the
+ pipeline (sources and sinks mostly). This element is useful to feed
+ non-live input such as local files into elements that expect live
+ input such as webrtcbin.`
+
+- New imagesequencesrc element to easily create a video stream from a
+ sequence of JPEG or PNG images (or any other encoding where the type
+ can be detected), basically a multifilesrc made specifically for
+ image sequences.
+
+- rpicamsrc element for capturing raw or encoded video (H.264, MJPEG)
+ from the Raspberry Pi camera. This works much like the popular
+ raspivid command line utility but outputs data nicely timestamped
+ and formatted in order to integrate nicely with other GStreamer
+ elements. Also comes with a device provider so applications can
+ discover the camera if available.
+
+- aatv and cacatv video filters that transform video ASCII art style
+
+- avtp: new Audio Video Transport Protocol (AVTP) plugin for Linux.
+ See Andre Guedes’ talk “Audio/Video Bridging (AVB) support in
+ GStreamer” for more details.
+
+- clockselect: a pipeline element that enables clock selection/forcing
+ via gst-launch pipeline syntax.
+
+- dashsink: Add new sink to produce DASH content. See Stéphane’s talk
+ or blog post for details.
+
+- dvbsubenc: a DVB subtitle encoder element
+
+- microdns: a libmicrodns-based mdns device provider to discover RTSP
+ cameras on the local network
+
+- mlaudiosink: new audio sink element for the Magic Leap platform,
+ accompanied by an MLSDK implementation in the amc plugin
-New element features and additions
-
-- playbin3, playbin and playsink have gained a new "text-offset"
- property to adjust the positioning of the selected subtitle stream
- vis-a-vis the audio and video streams. This uses subtitleoverlay’s
- new "subtitle-ts-offset" property. GstPlayer has gained matching API
- for this, namely gst_player_get_text_video_offset().
-
-- playbin3 buffering improvements: in network playback scenarios there
- may be multiple inputs to decodebin3, and buffering will be done
- before decodebin3 using queue2 or downloadbuffer elements inside
- urisourcebin. Since this is before any parsers or demuxers there may
- not be any bitrate information available for the various streams, so
- it was difficult to configure the buffering there smartly within
- global constraints. This was improved now: The queue2 elements
- inside urisourcebin will now use the new bitrate query to figure out
- a bitrate estimate for the stream if no bitrate was provided by
- upstream, and urisourcebin will use the bitrates of the individual
- queues to distribute the globally-set "buffer-size" budget in bytes
- to the various queues. urisourcebin also gained "low-watermark" and
- "high-watermark" properties which will be proxied to the internal
- queues, as well as a read-only "statistics" property which allows
- querying of the minimum/maximum/average byte and time levels of the
- queues inside the urisourcebin in question.
-
-- splitmuxsink has gained a couple of new features:
-
- - new "async-finalize" mode: This mode is useful for muxers or
- outputs that can take a long time to finalize a file. Instead of
- blocking the whole upstream pipeline while the muxer is doing
- its stuff, we can unlink it and spawn a new muxer + sink
- combination to continue running normally. This requires us to
- receive the muxer and sink (if needed) as factories via the new
- "muxer-factory" and "sink-factory" properties, optionally
- accompanied by their respective properties structures (set via
- the new "muxer-properties" and "sink-properties" properties).
- There are also new "muxer-added" and "sink-added" signals in
- case custom code has to be called for them to configure them.
-
- - "split-at-running-time" action signal: When called by the user,
- this action signal ends the current file (and starts a new one)
- as soon as the given running time is reached. If called multiple
- times, running times are queued up and processed in the order
- they were given.
-
- - "split-after" action signal to finish outputting the current GOP
- to the current file and then start a new file as soon as the GOP
- is finished and a new GOP is opened (unlike the existing
- "split-now" which immediately finishes the current file and
- writes the current GOP into the next newly-started file).
-
- - "reset-muxer" property: when unset, the muxer is reset using
- flush events instead of setting its state to NULL and back. This
- means the muxer can keep state across resets, e.g. mpegtsmux
- will keep the continuity counter continuous across segments as
- required by hlssink2.
-
-- qtdemux gained PIFF track encryption box support in addition to the
- already-existing PIFF sample encryption support, and also allows
- applications to select which encryption system to use via a
- "drm-preferred-decryption-system-id" context in case there are
- multiple options.
-
-- qtmux: the "start-gap-threshold" property determines now whether an
- edit list will be created to account for small gaps or offsets at
- the beginning of a stream in case the start timestamps of tracks
- don’t line up perfectly. Previously the threshold was hard-coded to
- 1% of the (video) frame duration, now it is 0 by default (so edit
- list will be created even for small differences), but fully
- configurable.
-
-- rtpjitterbuffer has improved end-of-stream handling
-
-- rtpmp4vpay will be prefered over rtpmp4gpay for MPEG-4 video in
- autoplugging scenarios now
-
-- rtspsrc now allows applications to send RTSP SET_PARAMETER and
- GET_PARAMETER requests using action signals.
-
-- rtspsrc has a small (100ms) configurable teardown delay by default
- to try and make sure an RTSP TEARDOWN request gets sent out when the
- source element shuts down. This will block the downward PAUSED to
- READY state change for a short time, but can be disabled where it’s
- a problem. Some servers only allow a limited number of concurrent
- clients, so if no proper TEARDOWN is sent new clients may have
- problems connecting to the server for a while.
-
-- souphttpsrc behaves better with low bitrate streams now. Before it
- would increase the read block size too quickly which could lead to
- it not reading any data from the socket for a very long time with
- low bitrate streams that are output live downstream. This could lead
- to servers kicking off the client.
-
-- filesink: do internal buffering to avoid performance regression with
- small writes since we bypass libc buffering by using writev()
- instead of fwrite()
-
-- identity: add "eos-after" property and fix "error-after" property
- when the element is reused
-
-- input-selector: lets context queries pass through, so that
- e.g. upstream OpenGL elements can use contexts and displays
- advertised by downstream elements
-
-- queue2: avoid ping-pong between 0% and 100% buffering messages if
- upstream is pushing buffers larger than one of its limits, plus
- performance optimisations
-
-- opusdec: new "phase-inversion" property to control phase inversion.
- When enabled, this will slightly increase stereo quality, but
- produces a stream that when downmixed to mono will suffer audio
- distortions.
-
-- The x265enc HEVC encoder also exposes a "key-int-max" property to
- configure the maximum allowed GOP size now.
-
-- decklinkvideosink has seen stability improvements for long-running
- pipelines (potential crash due to overflow of leaked clock refcount)
- and clock-slaving improvements when performing flushing seeks
- (causing stalls in the output timeline), pausing and/or buffering.
-
-- srtpdec, srtpenc: add support for MKIs which allow multiple keys to
- be used with a single SRTP stream
-
-- srtpdec, srtpenc: add support for AES-GCM and also add support for
- it in gst-rtsp-server and rtspsrc.
-
-- The srt Secure Reliable Transport plugin has integrated server and
- client elements srt{client,server}{src,sink} into one (srtsrc and
- srtsink), since SRT connection mode can be changed by uri
- parameters.
+- msdkvp9enc: VP9 encoder element for the Intel MediaSDK
-- h264parse and h265parse will handle SEI recovery point messages and
- mark recovery points as keyframes as well (in addition to IDR
- frames)
+- rist: new plugin implementing support for the Video Services Forum’s
+ Reliable Internet Stream Transport (RIST) TR-06-1 Simple Profile.
+ See Nicolas’ blog post “GStreamer support for the RIST
+ Specification” for more details.
+
+- rtmp2: new RTMP client source and sink elements with fully
+ asynchronous network operations, better robustness and additional
+ features such as handling ping and stats messages, and adobe-style
+ authentication. The new rtmp2src and rtmp2sink elements should be
+ API-compatible with the old rtmpsrc / rtmpsink elements and should
+ work as drop-in replacements.
-- webrtcbin: "add-turn-server" action signal to pass multiple ICE
- relays (TURN servers).
+- new RTP source and sink elements to easily set up RTP streaming via
+ rtp:// URIs: The rtpsink and rtpsrc elements add an URI interface so
+ that streams can be decoded with decodebin using rtp:// URIs. These
+ can be used as follows: ``` gst-launch-1.0 videotestsrc ! x264enc !
+ rtph264pay config-interval=3 ! rtpsink uri=rtp://239.1.1.1:1234
-- The removesilence element has received various new features and
- properties, such as a "threshold" property, detecting silence only
- after minimum silence time/buffers, a "silent" property to control
- bus message notifications as well as a "squash" property.
+ gst-launch-1.0 videotestsrc ! x264enc ! rtph264pay config-interval=1
+ ! rtpsink uri=rtp://239.1.2.3:5000 gst-launch-1.0 rtpsrc
+ uri=rtp://239.1.2.3:5000?encoding-name=H264 ! rtph264depay !
+ avdec_h264 ! videoconvert ! xvimagesink
-- AOMedia AV1 decoder gained support for 10/12bit decoding whilst the
- AV1 encoder supports more image formats and subsamplings now and
- acquired support for rate control and profile related configuration.
+ gst-launch-1.0 videotestsrc ! avenc_mpeg4 ! rtpmp4vpay
+ config-interval=1 ! rtpsink uri=rtp://239.1.2.3:5000 gst-launch-1.0
+ rtpsrc uri=rtp://239.1.2.3:5000?encoding-name=MP4V-ES ! rtpmp4vdepay
+ ! avdec_mpeg4 ! videoconvert ! xvimagesink ```
-- The Fraunhofer fdkaac plugin can now be built against the 2.0.0
- version API and has improved multichannel support
+- svthevcenc: new SVT-HEVC-based H.265 video encoder
-- kmssink now supports unpadded 24-bit RGB and can configure mode
- setting from video info, which enables display of multi-planar
- formats such as I420 or NV12 with modesetting. It has also gained a
- number of new properties: The "restore-crtc" property does what it
- says on the tin and is enabled by default. "plane-properties" and
- "connector-properties" can be used to pass custom properties to the
- DRM.
+- switchbin: new helper element which chooses between a set of
+ processing chains (paths) based on input caps, and changes the
+ active chain if new caps arrive. Paths are child objects, which are
+ accessed by the GstChildProxy interface. See the switchbin
+ documentation for a usage example.
-- waylandsink has a "fullscreen" property now and supports the
- XDG-Shell protocol.
+- vah264dec: new experimental va plugin with an element for H.264
+ decoding with VA-API using GStreamer’s new stateless decoder
+ infrastructure (see Linux section below).
+
+- v4l2codecs: introduce an V4L2 CODECs Accelerator supporting the new
+ CODECs uAPI in the Linux kernel (see Linux section below)
-- decklinkvideosink, decklinkvideosrc support selecting between
- half/full duplex
+- zxing new plugin to detect QR codes and barcodes, based on libzxing
+
+- also see the Rust plugins section below which contains plenty of new
+ exciting plugins written in Rust!
-- The vulkan plugin gained support for macOS and iOS via MoltenVK in
- addition to the existing support for X11 and Wayland
+New element features and additions
-- imagefreeze has a new num-buffers property to limit the number of
- buffers that are produced and to send an EOS event afterwards
+GStreamer core
-- webrtcbin has a new, introspectable get-transceiver signal in
- addition to the old get-transceivers signal that couldn’t be used
- from bindings
+- filesink: Add a new “full” buffer mode. Previously the default and
+ full modes were the same. Now the default mode is like before: it
+ accumulates all buffers in a buffer list until the threshold is
+ reached and then writes them all out, potentially in multiple
+ writes. The new full mode works by always copying memory to a single
+ memory area and writing everything out with a single write once the
+ threshold is reached.
-- Support for per-element latency information was added to the latency
- tracer
+- multiqueue: Add stats property and
+ current-level-{buffers, bytes, time} pad properties to query the
+ current levels of the corresponding internal queue.
-Plugin and library moves
+Plugins Base
-- The stereo element was moved from -bad into the existing audiofx
- plugin in -good. If you get duplicate type registration warnings
- when upgrading, check that you don’t have a stale stereoplugin lying
- about somewhere.
-
-GstVideoAggregator, compositor, and OpenGL mixer elements moved from -bad to -base
-
-GstVideoAggregator is a new base class for raw video mixers and muxers
-and is based on GstAggregator. It provides defined-latency mixing of raw
-video inputs and ensures that the pipeline won’t stall even if one of
-the input streams stops producing data.
-
-As part of the move to stabilise the API there were some last-minute API
-changes and clean-ups, but those should mostly affect internal elements.
-Most notably, the "ignore-eos" pad property was renamed to
-"repeat-after-eos" and the conversion code was moved to a
-GstVideoAggregatorConvertPad subclass to avoid code duplication, make
-things less awkward for subclasses like the OpenGL-based video mixer,
-and make the API more consistent with the audio aggregator API.
-
-It is used by the compositor element, which is a replacement for
-‘videomixer’ which did not handle live inputs very well. compositor
-should behave much better in that respect and generally behave as one
-would expected in most scenarios.
-
-The compositor element has gained support for per-pad blending mode
-operators (SOURCE, OVER, ADD) which determines what operator to use for
-blending this pad over the previous ones. This can be used to implement
-crossfading and the available operators can be extended in the future as
-needed.
-
-A number of OpenGL-based video mixer elements (glvideomixer, glmixerbin,
-glvideomixerelement, glstereomix, glmosaic) which are built on top of
-GstVideoAggregator have also been moved from -bad to -base now. These
-elements have been merged into the existing OpenGL plugin, so if you get
-duplicate type registration warnings when upgrading, check that you
-don’t have a stale openglmixers plugin lying about somewhere.
+- alsa: implement a device provider
-Plugin removals
+- alsasrc: added use-driver-timestamp property to force use of
+ pipeline timestamps (and disable driver timestamps) if so desired
-The following plugins have been removed from gst-plugins-bad:
+- audioconvert: fix changing the mix-matrix property at runtime
-- The experimental daala plugin has been removed, since it’s not so
- useful now that all effort is focused on AV1 instead, and it had to
- be enabled explicitly with --enable-experimental anyway.
+- appsrc: added support for segment forwarding or custom GstSegments
+ via GstSample, enabled via the handle-segment-change property. This
+ only works for segments in TIME format for now.
-- The spc plugin has been removed. It has been replaced by the gme
- plugin.
+- compositor: various performance optimisations, checkerboard drawing
+ fixes, and support for VUYA format
-- The acmmp3dec and acmenc plugins for Windows have been removed. ACM
- is an ancient legacy API and there was no point in keeping the
- plugins around for a licensed MP3 decoder now that the MP3 patents
- have expired and we have a decoder in -good. We also didn’t ship
- these in our cerbero-built Windows packages, so it’s unlikely that
- they’ll be missed.
+- encodebin: Fix and refactor smart encoding; ensure that a single
+ segment is pushed into encoders; improve force-key-unit event
+ handling.
+- opusenc: Add low delay option (audio-type=restricted-lowdelay) to
+ disable the SILK layer and achieve only 5ms delay.
-Miscellaneous API additions
+- opusdec: add stats property to retrieve various decoder statistics.
-- GstBitwriter: new generic bit writer API to complement the existing
- bit reader
-
-- gst_buffer_new_wrapped_bytes() creates a wrap buffer from a GBytes
-
-- gst_caps_set_features_simple() sets a caps feature on all the
- structures of a GstCaps
-
-- New GST_QUERY_BITRATE query: This allows determining from downstream
- what the expected bitrate of a stream may be which is useful in
- queue2 for setting time based limits when upstream does not provide
- timing information. tsdemux, qtdemux and matroskademux have basic
- support for this query on their sink pads.
-
-- elements: there is a new “Hardware” class specifier. Elements
- interacting with hardware devices should specify this classifier in
- their element factory class metadata. This is useful to advertise as
- one might need to put such elements into READY state to test if the
- hardware is present in the system for example.
-
-- protection: Add a new definition for unspecified system protection,
- GST_PROTECTION_UNSPECIFIED_SYSTEM_ID
-
-- take functions for various mini objects that didn’t have them yet:
- gst_query_take(), gst_message_take(), gst_tag_list_take(),
- gst_buffer_list_take(). Unlike the various _replace() functions
- _take() does not increase the reference count but takes ownership of
- the mini object passed.
-
-- clear functions for various mini object types and GstObject which
- unrefs the object or mini object (if non-NULL) and sets the variable
- pointed to to NULL: gst_clear_structure(), gst_clear_tag_list(),
- gst_clear_query(), gst_clear_message(), gst_clear_event(),
- gst_clear_caps(), gst_clear_buffer_list(), gst_clear_buffer(),
- gst_clear_mini_object(), gst_clear_object()
-
-- miniobject: new API gst_mini_object_add_parent() and
- gst_mini_object_remove_parent() to set parent pointers on mini
- objects to ensure correct writability: Every container of
- miniobjects now needs to store itself as parent in the child object,
- and remove itself again later. A mini object is then only writable
- if there is at most one parent, that parent is writable itself, and
- the reference count of the mini object is 1. GstBuffer (for
- memories), GstBufferList (for buffers), GstSample (for caps, buffer,
- bufferlist), and GstVideoOverlayComposition were updated
- accordingly. Without this it was possible to have e.g. a buffer list
- with a refcount of 2 used in two places at once that both modify the
- same buffer with refcount 1 at the same time wrongly thinking it is
- writable even though it’s really not.
-
-- poll: add API to watch for POLLPRI and stop treating POLLPRI as a
- read. This is useful to wait for video4linux events which are
- signalled via POLLPRI.
-
-- sample: new API to update the contents of a GstSample and make it
- writable: gst_sample_set_buffer(), gst_sample_set_caps(),
- gst_sample_set_segment(), gst_sample_set_info(), plus
- gst_sample_is_writable() and gst_sample_make_writable(). This makes
- it possible to reuse a sample object and avoid unnecessary memory
- allocations, for example in appsink.
-
-- ClockIDs now keep a weak reference to underlying clock to avoid
- crashes in basesink in corner cases where a clock goes away while
- the ClockID is still in use, plus some new API
- (gst_clock_id_get_clock(), gst_clock_id_uses_clock()) to check the
- clock a ClockID is linked to.
-
-- The GstCheck unit test library gained a
- fail_unless_equals_clocktime() convenience macro as well as some new
- GstHarness API for for proposing meta APIs from the allocation
- query: gst_harness_add_propose_allocation_meta(). ASSERT_CRITICAL()
- checks in unit tests are now skipped if GStreamer was compiled with
- GST_DISABLE_GLIB_CHECKS.
-
-- gst_audio_buffer_truncate() convenience function to truncate a raw
- audio buffer
-
-- GstDiscoverer has support for caching the results of discovery in
- the default cache directory. This can be enabled with the use-cache
- property and is disabled by default.
-
-- GstMeta that are attached to GstBuffers are now always stored in the
- order in which they were added.
-
-- Additional support for signalling ONVIF specific features were
- added: the SEEK event can store a trickmode-interval now and support
- for the Rate-Control and Frames RTSP headers was added to the RTSP
- library.
-
-
-Miscellaneous performance and memory optimisations
+- uridecodebin3: Let decodebin3 do its stream selection if no one
+ answers
-As always there have been many performance and memory usage improvements
-across all components and modules. Some of them (such as dmabuf
-import/export) have already been mentioned elsewhere so won’t be
-repeated here.
+- decodebin3: Avoid overriding explicit user selection of streams
-The following list is only a small snapshot of some of the more
-interesting optimisations that haven’t been mentioned in other contexts
-yet:
+- playbin: add flag to force use of software decoders over any
+ hardware decoders that might also be available
-- The GstVideoEncoder and GstVideoDecoder base classes now release the
- STREAM_LOCK when pushing out buffers, which means (multi-threaded)
- encoders and decoders can now receive and continue to process input
- buffers whilst waiting for downstream elements in the pipeline to
- process the buffer that was pushed out. This increases throughput
- and reduces processing latency, also and especially for
- hardware-accelerated encoder/decoder elements.
+- playbin3, playbin: propagate sink context
-- GstQueueArray has seen a few API additions
- (gst_queue_array_peek_nth(), gst_queue_array_set_clear_func(),
- gst_queue_array_clear()) so that it can be used in other places like
- GstAdapter instead of a GList, which reduces allocations and
- improves performance.
+- rawvideoparse: Fix tiling support, allow setting colorimetry
-- appsink now reuses the sample object in pull_sample() if possible
+- subparse: output plain utf8 text instead of pango-markup formatted
+ text if downstream requires it, useful for interop with elements
+ that only accept utf8-formatted subtitles such as muxers or closed
+ caption converters.
-- rtpsession only starts the RTCP thread when it’s actually needed now
+- tcpserversrc, tcpclientsrc: add stats property with TCP connection
+ stats (some are only available on Linux though)
-- udpsrc uses a buffer pool now and the GstUdpSrc object structure was
- optimised for better cache performance
+- timeoverlay: add show-times-as-dates, datetime-format and
+ datetime-epoch properties to display times with dates
-GstPlayer
+- videorate: Fix changing rate property during playback; reverse
+ playback fixes; update QoS events taking into account our rate
-- API was added to fine-tune the synchronisation offset between
- subtitles and video
+- videoscale: pass through and transform size sensitive metas instead
+ of just dropping them
+
+Plugins Good
+
+- avidemux can handle H.265 video now. Our advice remains to
+ immediately cease all contact and communication with anyone who
+ hands you H.265 video in an AVI container, however.
+
+- avimux: Add support for S24LE and S32LE raw audio and v210 raw video
+ formats; support more than 2 channels of raw audio.
+
+- souphttpsrc: disable session sharing and cookie jar when the cookies
+ property is set; correctly handle seeks past the end of the content
+
+- deinterlace: new YADIF deinterlace method which should provide
+ better quality than the existing methods and is LGPL licensed;
+ alternate fields are supported as input to the deinterlacer as well
+ now, and there were also fixes for switching the deinterlace mode on
+ the fly.
+
+- flvmux: in streamable mode allow adding new pads even if the initial
+ header has already been written. Old clients will only process the
+ initial stream, new clients will get a header with the new streams.
+ The skip-backwards-streams property can be used to force flvmux to
+ skip and drop a few buffers rather than produce timestamps that go
+ backward and confuse librtmp-based clients. There’s also better
+ handling for timestamp rollover when streaming for a long time.
+
+- imagefreeze: Add live mode, which can be enabled via the new is-live
+ property. In this mode frames will only be output in PLAYING state
+ according to the negotiated framerate, skipping frames if the output
+ can’t keep up (e.g. because it’s blocked downstream). This makes it
+ possible to actually use imagefreeze in live pipelines without
+ having to manually ensure somehow that it starts outputting at the
+ current running time and without still risking to fall behind
+ without recovery.
+
+- matroskademux, qtdemux: Provide audio lead-in for some lossy formats
+ when doing accurate seeks, to make sure we can actually decode
+ samples at the desired position. This is especially important for
+ non-linear audio/video editing use-cases.
+
+- matroskademux, matroskamux: Handle interlaced field order (tff, bff)
+
+- matroskamux:
+
+ - new offset-to-zero property to offset all streams to start at
+ zero. This takes the timestamp of the earliest stream and
+ offsets it so that it starts at 0. Some software (VLC,
+ ffmpeg-based) does not properly handle Matroska files that start
+ at timestamps much bigger than zero, which could happen with
+ live streams.
+ - added a creation-time property to explicitly set the creation
+ time to write into the file headers. Useful when remuxing, for
+ example, but also for live feeds where the DateUTC header can be
+ set a UTC timestamp corresponding to the beginning of the file.
+ - the muxer now also always waits for caps on sparse streams, and
+ warns if caps arrive after the header has already been sent,
+ otherwise the subtitle track might be silently absent in the
+ final file. This might affect applications that send sparse data
+ into matroskamux via an appsrc element, which will usually not
+ send out the initial caps before it sends out the first buffer.
+
+- pulseaudio: device provider improvements: fix discovery of
+ newly-added devices and hide the alsa device provider if we provide
+ alsa devices
+
+- qtdemux: raw audio handling improvements, support for AC4 audio, and
+ key-units trickmode interval support
+
+- qtmux:
+
+ - was ported to the GstAggregator base class which allows for
+ better handling of live inputs, but might entail minor
+ behavioural changes for sparse inputs if inputs are not live.
+ - has also gained a force-create-timecode-trak property to create
+ a timecode trak in non-mov flavors, which may not be supported
+ by Apple but is supported by other software such as Final Cut
+ Pro X
+ - also a force-chunks property to force the creation of chunks
+ even in single-stream files, which is required for Apple ProRes
+ certification.
+ - also supports 8k resolutions in prefill mode with ProRes.
+
+- rtpbin gained a request-jitterbuffer signal which allows
+ applications to plug in their own jitterbuffer implementation such
+ as the threadsharing jitterbuffer from the Rust plugins, for
+ example.
+
+- rtprtxsend: add clock-rate-map property to allow generic RTP input
+ caps without a clock-rate whilst still supporting the max-size-time
+ property for bundled streams.
+
+- rtpssrcdemux: introduce max-streams property to guard against
+ attacks where the sender changes SSRC for every RTP packet.
+
+- rtph264pay, rtph264pay: implement STAP-A and various aggregation
+ modes controled by the new aggegrate-mode property: none to not
+ aggregate NAL units (as before), zero-latency to aggregate NAL units
+ until a VCL or suffix unit is included, or max to aggregate all NAL
+ units with the same timestamp (which adds one frame of latency). The
+ default has been kept at none for backwards compatibility reasons
+ and because various RTP/RTSP implementions don’t handle aggregation
+ well. For WebRTC use cases this should be set to zero-latency,
+ however.
+- rtpmp4vpay: add support for config-interval=-1 to resend headers
+ with each IDR keyframe, like other video payloaders.
-Miscellaneous changes
+- rtpvp8depay: Add wait-for-keyframe property for waiting until the
+ next keyframe after packet loss. Useful if the video stream was not
+ encoded with error resilience enabled, in which case packet loss
+ tends to cause very bad artefacts when decoding, and waiting for the
+ next keyframe instead improves user experience considerably.
-- As a result of moving to newer FFmpeg APIs, encoder and decoder
- elements exposed by the GStreamer FFmpeg wrapper plugin (gst-libav)
- may have seen possibly incompatible changes to property names and/or
- types, and not all properties exposed might be functional. We are
- still reviewing the new properties and aim to minimise breaking
- changes at least for the most commonly-used properties, so please
- report any issues you run into!
+- splitmuxsink and splitmuxsrc can now handle auxiliary video streams
+ in addition to the primary video stream. The primary video stream is
+ still used to select fragment cut points at keyframe boundaries.
+ Auxilliary video streams may be broken up at any packet - so
+ fragments may not start with a keyframe for those streams.
-OpenGL integration
+- splitmuxsink:
-- The OpenGL mixer elements have been moved from -bad to
- gst-plugins-base (see above)
+ - new muxer-preset and sink-preset properties for setting
+ muxer/sink presets
+ - a new start-index property to set the initial fragment id
+ - and a new muxer-pad-map property which explicitly maps
+ splitmuxsink pads to the muxer pads they should connect to,
+ overriding the implicit logic that tries to match pads but
+ yields arbitrary names.
+ - Also includes the actual sink element in the fragment-opened and
+ fragment-closed element messages now, which is especially useful
+ for sinks without a location property or when finalisation of
+ the fragments is done asynchronously.
+
+- videocrop: add support for Y444, Y41B and Y42B pixel formats
+
+- vp8enc, vp9enc: change default value of VP8E_SET_STATIC_THRESHOLD
+ from 0 to 1 which matches what Google WebRTC does and results in
+ lower CPU usage; also added a new bit-per-pixel property to select a
+ better default bitrate
+
+- v4l2: add support for ABGR, xBGR, RGBA, and RGBx formats and for
+ handling interlaced video in alternate fields interlace mode (one
+ field per buffer instead of one frame per picture with both fields
+ interleaved)
+
+- v4l2: Profile and level probing support for H264, H265, MPEG-4,
+ MPEG-2, VP8, and VP9 video encoders and decoders
+
+Plugins Ugly
+
+- asfdemux: extract more metadata: disc number and disc count
+
+- x264enc:
+
+ - respect YouTube bitrate recommendation when user sets the
+ YouTube profile preset
+ - separate high-10 video formats from 8-bit formats to improve
+ depth negotiation and only advertise suitable input raw formats
+ for the desired output depth
+ - forward downstream colorimetry and chroma-site restrictions to
+ upstream elements
+ - support more color primaries/mappings
+
+Plugins Bad
+
+- av1enc: add threads, row-mt and tile-{columns,rows} properties for
+ this AOMedia AV1 encoder
+
+- ccconverter: implement support for CDP framerate conversions
+
+- ccextractor: Add remove-caption-meta property to remove caption
+ metas from the outgoing video buffers
+
+- decklink: add support for 2K DCI video modes, widescreen NTSC/PAL,
+ and for parsing/outputting AFD/Bar data. Also implement a simple
+ device provider for Decklink devices.
+
+- dtlsrtpenc: add rtp-sync property which synchronises RTP streams to
+ the pipeline clock before passing them to funnel for merging with
+ RTCP.
+
+- fdkaac: also decode MPEG-2 AAC; encoder now supports more
+ multichannel/surround sound layouts
+
+- hlssink2: add action signals for custom playlist/fragment handling:
+ Instead of always going through the file system API we allow the
+ application to modify the behaviour. For the playlist itself and
+ fragments, the application can provide a GOutputStream. In addition
+ the sink notifies the application whenever a fragment can be
+ deleted.
+
+- interlace: can now output data in alternate fields mode; added field
+ switching mode for 2:2 field pattern
+
+- iqa: Add a mode property to enable strict mode that checks that all
+ the input streams have the exact same number of frames; also
+ implement the child proxy interface
+
+- mpeg2enc: add disable-encode-retries property for lower CPU usage
+
+- mpeg4videoparse: allow re-sending codec config at IDR via
+ config-interval=-1
+
+- mpegtsparse: new alignment property to determine number of TS
+ packets per output buffer, useful for feeding an MPEG-TS stream for
+ sending via udpsink. This can be used in combination with the
+ split-on-rai property that makes sure to start a new output buffer
+ for any TS packet with the Random Access Indicator set. Also set
+ delta unit buffer flag on non-random-access buffers.
+
+- mpegdemux: add an ignore-scr property to ignore the SCR in
+ non-compliant MPEG-PS streams with a broken SCR, which will work as
+ long as PTS/DTS in the PES header is consistently increasing.
+
+- tsdemux:
+
+ - add an ignore-pcr property to ignore MPEG-TS streams with broken
+ PCR streams on which we can’t reliably recover correct
+ timestamps.
+ - new latency property to allow applications to lower the
+ advertised worst-case latency of 700ms if they know their
+ streams support this (must have timestamps in higher frequency
+ than required by the spec)
+ - support for AC4 audio
+
+- msdk - Intel Media SDK plugin for hardware-accelerated video
+ decoding and encoding on Windows and Linux:
+
+ - mappings for more video formats: Y210, Y410, P012_LE, Y212_LE
+ - encoders now support bitrate changes and input format changes in
+ playing state
+ - msdkh264enc, msdkh265enc: add support for CEA708 closed caption
+ insertion
+ - msdkh264enc, msdkh265enc: set Region of Interest (ROI) region
+ from ROI metas
+ - msdkh264enc, msdkh265enc: new tune property to enable low-power
+ mode
+ - msdkh265enc: add support 12-bit 4:2:0 encoding and 8-bit 4:2:2
+ encoding and VUYA, Y210, and Y410 as input formats
+ - msdkh265enc: add support for screen content coding extension
+ - msdkh265dec: add support for main-12/main-12-intra,
+ main-422-10/main-422-10-intra 10bit,
+ main-422-10/main-422-10-intra 8bit,
+ main-422-12/main-422-12-intra, main-444-10/main-444-10-intra,
+ main-444-12/main-444-12-intra, and main-444 profiles
+ - msdkvp9dec: add support for 12-bit 4:4:4
+ - msdkvpp: add support for Y410 and Y210 formats, cropping via
+ properties, and a new video-direction property.
+
+- mxf: Add support for CEA-708 CDP from S436 essence tracks. mxfdemux
+ can now handle Apple ProRes
+
+- nvdec: add H264 + H265 stateless codec implementation nvh264sldec
+ and nvh265sldec with fewer features but improved latency. You can
+ set the environment variable GST_USE_NV_STATELESS_CODEC=h264 to use
+ the stateless decoder variant as nvh264dec instead of the “normal”
+ NVDEC decoder implementation.
+
+- nvdec: add support for 12-bit 4:4:4/4:2:0 and 10-bit 4:2:0 decoding
+
+- nvenc:
+
+ - add more rate-control options, support for B-frame encoding (if
+ device supports it), an aud property to toggle Access Unit
+ Delimiter insertion, and qp-{min,max,const}-{i,p,b} properties.
+ - the weighted-pred property enables weighted prediction.
+ - support for more input formats, namely 8-bit and 10-bit RGB
+ formats (BGRA, RGBA, RGB10A2, BGR10A2) and YV12 and VUYA.
+ - on-the-fly resolution changes are now supported as well.
+ - in case there are multiple GPUs on the system, there are also
+ per-GPU elements registered now, since different devices will
+ have different capabilities.
+ - nvh265enc can now support 10-bit YUV 4:4:4 encoding and 8-bit
+ 4:4:4 / 10-bit 4:2:0 formats up to 8K resolution (with some
+ devices). In case of HDR content HDR related SEI nals will be
+ inserted automatically.
+
+- openjpeg: enable multi-threaded decoding and add support for
+ sub-frame encoding (for lower latency)
+
+- rtponviftimestamp: add opt-out “drop-out-of-segment” property
+
+- spanplc: new stats property
+
+- srt: add support for IPv6 and for using hostnames instead of IP
+ addresses; add streamid property, but also allow passing the id via
+ the stream URI; add wait-for-connection property to srtsink
+
+- timecodestamper: this element was rewritten with an updated API
+ (properties); it has gained many new properties, seeking support and
+ support for linear timecode (LTC) from an audio stream.
+
+- uvch264src now comes with a device provider to advertise available
+ camera sources that support this interface (mostly Logitech C920s)
+
+- wpe: Add software rendering support and support for mouse scroll
+ events
+
+- x265enc: support more 8/10/12 bits 4:2:0, 4:2:2 and 4:4:4 profiles;
+ add support for mastering display info and content light level
+ encoding SEIs
+
+gst-libav
+
+- Add mapping for SpeedHQ video codec used by NDI
+
+- Add mapping for aptX and aptX-HD
+
+- avivf_mux: support VP9 and AV1
+
+- avvidenc: shift output buffer timestamps and output segment by 1h
+ just like x264enc does, to allow for negative DTS.
+
+- avviddec: Limit default number of decoder threads on systems with
+ more than 16 cores, as the number of threads used in avdec has a
+ direct impact on the latency of the decoder, which is of as many
+ frames as threads, so a large numbers of threads can make for
+ latency levels that can be problematic in some applications.
+
+- avviddec: Add thread-type property that allows applications to
+ specify the preferred multithreading method (auto, frame, slice).
+ Note that thread-type=frame may introduce additional latency
+ especially in live pipelines, since it introduces a decoding delay
+ of number of thread frames.
-- The Mesa GBM backend now supports headless mode
+Plugin and library moves
-- gloverlaycompositor: New OpenGL-based compositor element that
- flattens any overlays from GstVideoOverlayCompositionMetas into the
- video stream.
+- There were no plugin moves or library moves in this cycle.
-- glalpha: New element that adds an alpha channel to a video stream.
- The values of the alpha channel can either be set to a constant or
- can be dynamically calculated via chroma keying. It is similar to
- the existing alpha element but based on OpenGL. Calculations are
- done in floating point so results may not be identical to the output
- of the existing alpha element.
+- The rpicamsrc element was moved into -good from an external
+ repository on github.
-- glupload: Implement direct dmabuf uploader, the idea being that some
- GPUs (like the Vivante series) can actually perform the YUV->RGB
- conversion internally, so no custom conversion shaders are needed.
- To make use of this feature, we need an additional uploader that can
- import DMABUF FDs and also directly pass the pixel format, relying
- on the GPU to do the conversion.
+Plugin removals
-- The OpenGL library no longer restores the OpenGL viewport. This is a
- performance optimization to not require performing multiple
- expensive glGet*() function calls per frame. This affects any
- application or plugin use of the following functions and objects:
- - glcolorconvert library object (not the element)
- - glviewconvert library object (not the element)
- - gst_gl_framebuffer_draw_to_texture()
- - custom GstGLWindow implementations
+The following elements or plugins have been removed:
+- The yadif video deinterlacing plugin from gst-plugins-bad, which was
+ one of the few GPL licensed plugins, has been removed in favour of
+ deinterlace method=yadif.
-Tracing framework and debugging improvements
+- The avdec_cdgraphics CD Graphics video decoder element from
+ gst-libav was never usable in GStreamer and we now have a cdgdec
+ element written in Rust in gst-plugins-rs to replace it.
-- There is now a GDB PRETTY PRINTER FOR VARIOUS GSTREAMER TYPES: For
- GstObject pointers the type and name is added, e.g.
- 0x5555557e4110 [GstDecodeBin|decodebin0]. For GstMiniObject pointers
- the object type is added, e.g. 0x7fffe001fc50 [GstBuffer]. For
- GstClockTime and GstClockTimeDiff the time is also printed in human
- readable form, e.g. 150116219955 [+0:02:30.116219955].
+- The VDPAU plugin has been unmaintained and unsupported for a very
+ long time and does not have the feature set we expect from
+ hardware-accelerated video decoders. It’s been superseded by the
+ nvcodec plugin leveraging NVIDIA’s NVDEC API.
-- GDB EXTENSION WITH TWO CUSTOM GDB COMMANDS gst-dot AND gst-print:
+Miscellaneous API additions
- - gst-dot creates dot files that a very close to what
- GST_DEBUG_BIN_TO_DOT_FILE() produces, but object properties and
- buffer contents such as codec-data in caps are not available.
+GStreamer core
+
+- gst_task_resume(): This new API allows resuming a task if it was
+ paused, while leaving it in stopped state if it was stopped or not
+ started yet. This can be useful for callback-based driver workflows,
+ where you basically want to pause and resume the task when buffers
+ are notified while avoiding the race with a gst_task_stop() coming
+ from another thread.
+
+- info: add printf extensions GST_TIMEP_FORMAT and GST_STIMEP_FORMAT
+ for printing GstClockTime/GstClockTimeDiff pointers, which is much
+ more convenient to use in debug log statements than the usual
+ GST_TIME_FORMAT-followed-by-GST_TIME_ARGS dance. Also add an
+ explicit GST_STACK_TRACE_SHOW_NONE enum value.
+
+- gst_element_get_current_clock_time() and
+ gst_element_get_current_running_time(): new helper functions for
+ getting an element clock’s time, and the clock time minus base time,
+ respectively. Useful when adding additional input branches to
+ elements such as compositor, audiomixer, flvmux, interleave or
+ input-selector to determine initial pad offsets and such.
+
+- seeking: Add GST_SEEK_FLAG_TRICKMODE_FORWARD_PREDICTED to just skip
+ B-frames during trick mode, showing both keyframes + P-frame, and
+ add support for it in h264parse and h265parse.
+
+- elementfactory: add GST_ELEMENT_FACTORY_TYPE_HARDWARE to allow
+ elements to advertise that they are hardware-based or interact with
+ hardware. This has multiple applications:
+
+ - it makes it possible to easily differentiate hardware and
+ software based element implementations such as audio or video
+ encoders and decoders. This is useful in order to force the use
+ of software decoders for specific use cases, or to check if a
+ selected decoder is actually hardware-accelerated or not.
+ - elements interacting with hardware and their respective drivers
+ typically don’t know the actually supported capabilities until
+ the element is set into at least READY state and can open a
+ device handle and probe the hardware.
+
+- gst_uri_from_string_escaped(): identical to gst_uri_from_string()
+ except that the userinfo and fragment components of the URI will not
+ be unescaped while parsing. This is needed for correctly parsing
+ usernames or passwords with : in them .
+
+- paramspecs: new GstParamSpec flag GST_PARAM_CONDITIONALLY_AVAILABLE
+ to indicate that a property might not always exist.
+
+- gst_bin_iterate_all_by_element_factory_name() finds elements in a
+ bin by factory name
+
+- pad: gst_pad_get_single_internal_link() is a new convenience
+ function to return the single internal link of a pad, which is
+ useful e.g. to retrieve the output pad of a new multiqueue request
+ pad.
+
+- datetime: Add constructors to create datetimes with timestamps in
+ microseconds, gst_date_time_new_from_unix_epoch_local_time_usecs()
+ and gst_date_time_new_from_unix_epoch_utc_usecs().
+
+- gst_debug_log_get_lines() gets debug log lines formatted in the same
+ way the default log handler would print them
+
+- GstSystemClock: Add GST_CLOCK_TYPE_TAI as GStreamer abstraction for
+ CLOCK_TAI, to support transmission offloading features where network
+ packets are timestamped with the time they are deemed to be actually
+ transmitted. Useful in combination with the new AVTP plugin.
+
+- miscellaneous utility functions: gst_clear_uri(),
+ gst_structure_take().
+
+- harness: Added gst_harness_pull_until_eos()
+
+- GstBaseSrc:
+
+ - gst_base_src_new_segment() allows subclasses to update the
+ segment to be used at runtime from the ::create() function. This
+ deprecates gst_base_src_new_seamless_segment()
+ - gst_base_src_negotiate() allows subclasses to trigger format
+ renegotiation at runtime from inside the ::create() or ::alloc()
+ function
+
+- GstBaseSink: new stats property and gst_base_sink_get_stats() method
+ to retrieve various statistics such as average frame rate and
+ dropped/rendered buffers.
+
+- GstBaseTransform: gst_base_transform_reconfigure() is now public
+ API, useful for subclasses that need to completely re-implement the
+ ::submit_input_buffer() virtual method
+
+- GstAggregator:
+
+ - gst_aggregator_update_segment() allows subclasses to update the
+ output segment at runtime. Subclasses should use this function
+ rather than push a segment event onto the source pad directly.
+ - new sample selection API:
+ - subclasses should now call gst_aggregator_selected_samples()
+ from their ::aggregate() implementation to signal that they
+ have selected the next samples they will aggregate
+ - GstAggregator will then emit the samples-selected signal
+ where handlers can then look up samples per pad via
+ gst_aggregator_peek_next_sample().
+ - This is useful for example to atomically update input pad
+ properties in mixer subclasses such as compositor.
+ Applications can now update properties with precise control
+ of when these changes will take effect, and for which input
+ buffer(s).
+ - gst_aggregator_finish_buffer_list() allows subclasses to push
+ out a buffer list, improving efficiency in some cases.
+ - a ::negotiate() virtual method was added, for consistency with
+ other base classes and to allow subclasses to completely
+ override the negotiation behaviour.
+ - the new ::sink_event_pre_queue() and ::sink_query_pre_queue()
+ virtual methods allow subclasses to intercept or handle
+ serialized events and queries before they’re queued up
+ internally.
+
+GStreamer Plugins Base Libraries
+
+Audio library
+
+- audioaggregator, audiomixer: new output-buffer-duration-fraction
+ property which allows use cases such as keeping the buffers output
+ by compositor on one branch and audiomixer on another perfectly
+ aligned, by requiring the compositor to output a n/d frame rate, and
+ setting output-buffer-duration-fraction to d/n on the audiomixer.
+
+- GstAudioDecoder: new max-errors property so applications can
+ configure at what point the decoder should error out, or tell it to
+ just keep going
+
+- gst_audio_make_raw_caps() and gst_audio_formats_raw() are
+ bindings-friendly versions of the GST_AUDIO_CAPS_MAKE() C macro.
+
+- gst_audio_info_from_caps() now handles encoded audio formats as well
+
+PbUtils library
+
+- GstEncodingProfile:
+ - Do not restrict number of similar profiles in a container
+ - add GstValue serialization function
+- codec utils now support more H.264/H.265 profiles/levels and have
+ improved extension handling
+
+RTP library
+
+- rtpbasepayloader: Add scale-rtptime property for scaling RTP
+ timestamp according to the segment rate (equivalent to RTSP speed
+ parameter). This is useful for ONVIF trickmodes via RTSP.
+
+- rtpbasepayload: add experimental property for embedding twcc
+ sequencenumbers for Transport-Wide Congestion Control (gated behind
+ the GST_RTP_ENABLE_EXPERIMENTAL_TWCC_PROPERTY environment
+ variable) - more generic API for enabling this is expected to land
+ in the next development cycle.
+
+- rtcpbuffer: add RTPFB_TYPE_TWCC for Transport-Wide Congestion
+ Control
+
+- rtpbuffer: add
+ gst_rtp_buffer_get_extension_onebyte_header_from_bytes()``, so that one can parse theGBytes`
+ returned by gst_rtp_buffer_get_extension_bytes()
+
+- rtpbasedepayload: Add max-reorder property to make the
+ previously-hardcoded value when to consider a sender to have
+ restarted configurable. In some scenarios it’s particularly useful
+ to set max-reorder=0 to disable the behaviour that the depayloader
+ will drop packets: when max-reorder is set to 0 all
+ reordered/duplicate packets are considered coming from a restarted
+ sender.
+
+RTSP library
+
+- add gst_rtsp_url_get_request_uri_with_control() to create request
+ uri combined with control url
+
+- GstRTSPConnection: add the possibility to limit the Content-Length
+ for RTSP messages via
+ gst_rtsp_connection_set_content_length_limit(). The same
+ functionality is also exposed in gst-rtsp-server.
+
+SDP library
+
+- add support for parsing the extmap attribute from caps and storing
+ inside caps The extmap attribute allows mapping RTP extension header
+ IDs to well-known RTP extension header specifications. See RFC8285
+ for details.
+
+Tags library
+
+- update to latest iso-code and support more languages
+
+- add tags for acoustid id & acoustid fingerprint, plus MusicBrainz ID
+ handling fixes
+
+Video library
+
+- High Dynamic Range (HDR) video information representation and
+ signalling enhancements:
+
+ - New APIs for HDR video information representation and
+ signalling:
+ - GstVideoMasteringDisplayInfo: display color volume info as
+ per SMPTE ST 2086
+ - GstVideoContentLightLevel: content light level specified in
+ CEA-861.3, Appendix A.
+ - plus functions to serialise/deserialise and add them to or
+ parse them from caps
+ - gst_video_color_{matrix,primaries,transfer}_{to,from}_iso():
+ new utilility functions for conversion from/to ISO/IEC
+ 23001-8
+ - add ARIB STD-B67 transfer chracteristic function
+ - add SMPTE ST 2084 support and BT 2100 colorimetry
+ - define bt2020-10 transfer characteristics for clarity:
+ bt707, bt2020-10, and bt2020-12 transfer characteristics are
+ functionally identical but have their own unique values in
+ the specification.
+ - h264parse, h265parse: Parse mastering display info and content
+ light level from SEIs.
+ - matroskademux: parse HDR metadata
+ - matroskamux: Write MasteringMetadata and Max{CLL,FALL}. Enable
+ muxing with HDR meta data if upstream provided it
+ - avviddec: Extract HDR information if any and map bt2020-10, PQ
+ and HLG transfer functions
+
+- added bt601 transfer function (for completeness)
+
+- support for more pixel formats:
+
+ - Y412 (packed 12 bits 4:4:4:4)
+ - Y212 (packed 12 bits 4:2:2)
+ - P012 (semi-planar 4:2:0)
+ - P016_{LE,BE} (semi-planar 16 bits 4:2:0)
+ - Y444_16{LE,BE} (planar 16 bits 4:4:4)
+ - RGB10A2_LE (packed 10-bit RGB with 2-bit alpha channel)
+ - NV12_32L32 (NV12 with 32x32 tiles in linear order)
+ - NV12_4L4 (NV12 with 4x4 tiles in linear order)
+
+- GstVideoDecoder:
+
+ - new max-errors property so applications can configure at what
+ point the decoder should error out, or tell it to just keep
+ going
+
+ - new qos property to disable dropping frames because of QoS, and
+ post QoS messages on the bus when dropping frames. This is
+ useful for example in a scenario where the decoded video is
+ tee-ed off to go into a live sink that syncs to the clock in one
+ branch, and an encoding and save to file pipeline in the other
+ branch. In that case one wouldn’t want QoS events from the video
+ sink make the decoder drop frames because that would also leave
+ gaps in the encoding branch then.
+
+- GstVideoEncoder:
+
+ - gst_video_encoder_finish_subframe() is new API to push out
+ subframes (e.g. slices), so encoders can split the encoding into
+ subframes, which can be useful to reduce the overall end-to-end
+ latency as we no longer need to wait for the full frame to be
+ encoded to start decoding or sending out the data.
+ - new min-force-key-unit-interval property allows configuring the
+ minimum interval between force-key-unit requests and prevents a
+ big bitrate increase if a lot of key-units are requested in a
+ short period of time (as might happen in live streaming RTP
+ pipelines when packet loss is detected).
+ - various force-key-unit event handling fixes
+
+- GstVideoAggregator, compositor, glvideomixer: expose
+ max-last-buffer-repeat property on pads. This can be used to have a
+ compositor display either the background or a stream on a lower
+ zorder after a live input stream freezes for a certain amount of
+ time, for example because of network issues.
+
+- gst_video_format_info_component() is new API to find out which
+ components are packed into a given plane, which is useful to prevent
+ us from assuming a 1-1 mapping between planes and components.
+
+- gst_video_make_raw_caps() and gst_video_formats_raw() are
+ bindings-friendly versions of the GST_VIDEO_CAPS_MAKE() C macro.
+
+- video-blend: Add support for blending on top of 16 bit per component
+ formats, which makes sure we can support every currently supported
+ raw video format for blending subtitles or logos on top of video.
+
+- GST_VIDEO_BUFFER_IS_TOP_FIELD() and
+ GST_VIDEO_BUFFER_IS_BOTTOM_FIELD() convenience macros to check
+ whether the video buffer contains only the top field or bottom field
+ of an interlaced picture.
+
+- GstVideoMeta now includes an alignment field with the
+ GstVideoAlignment so buffer producers can explicitly specify the
+ exact geometry of the planes, allowing users to easily know the
+ padded size and height of each plane. Default values will be used if
+ this is not set.
+
+ Use gst_video_meta_set_alignment() to set the alignment and
+ gst_video_meta_get_plane_size() or gst_video_meta_get_plane_height()
+ to compute the plane sizes or plane heights based on the information
+ in the video meta.
+
+- gst_video_info_align_full() works like gst_video_info_align() but
+ also retrieves the plane sizes.
+
+MPEG-TS library
+
+- support for SCTE-35 sections
+
+- extend support for ATSC tables:
+
+ - System Time Table (STT)
+ - Master Guide Table (MGT)
+ - Rating Region Table (RRT)
+
+Miscellaneous performance, latency and memory optimisations
- - gst-print produces high-level information about a GStreamer
- object. This is currently limited to pads for GstElements and
- events for the pads. The output may look like this:
+As always there have been many performance and memory usage improvements
+across all components and modules. Some of them have already been
+mentioned elsewhere so won’t be repeated here.
-- gst_structure_to_string() now serialises the actual value of
- pointers when serialising GstStructures instead of claiming they’re
- NULL. This makes debug logging in various places less confusing,
- because it’s clear now that structure fields actually hold valid
- objects. Such object pointer values will never be deserialised
- however.
+The following list is only a small snapshot of some of the more
+interesting optimisations that haven’t been mentioned in other contexts
+yet:
+- caps negotiation, structure and GValue performance optimizations
-Tools
+- systemclock: clock waiting performance improvements (moved from
+ GstPoll to GCond for waiting), especially on Windows.
-- gst-inspect-1.0 has coloured output now and will automatically use a
- pager if the output does not fit on a page. This only works in a
- UNIX environment and if the output is not piped, and on Windows 10
- build 16257 or newer. If you don’t like the colours you can disable
- them by setting the GST_INSPECT_NO_COLORS=1 environment variable or
- passing the --no-color command line option.
+- rtpsession: add support for buffer lists on the recv path for better
+ performance with higher packet rate streams.
+- rtpjitterbuffer: internal timer handling has been rewritten for
+ better performance, see Nicolas’ talk “Revisiting RTP Jitter Buffer
+ Timers” for more details.
-GStreamer RTSP server
+- H.264/H.265 parsers and RTP payloaders/depayloaders have been
+ optimised for latency to make sure data is processed and pushed out
+ as quickly as possible
-- Improved backlog handling when using TCP interleaved for data
- transport. Before there was a fixed maximum size for backlog
- messages, which was prone to deadlocks and made it difficult to
- control memory usage with the watch backlog. The RTSP server now
- limits queued TCP data messages to one per stream, moving queuing of
- the data into the pipeline and leaving the RTSP connection
- responsive to RTSP messages in both directions, preventing all those
- problems.
+- video-scaler: correctness and performance improvements, esp. for
+ interlaced formats and GBRA
-- Initial ULP Forward Error Correction support in rtspclientsink and
- for RECORD mode in the server.
+- GstVideoEncoder has gained new API to push out subframes
+ (e.g. slices), so encoders can split the encoding into subframes,
+ which can be useful to reduce the overall end-to-end latency as we
+ no longer need to wait for the full frame to be encoded to start
+ decoding or sending out the data.
-- API to explicitly enable retransmission requests (RTX)
+ This is complemented by the new GST_VIDEO_BUFFER_FLAG_MARKER which
+ is a video-specific buffer flag to mark the end of a video frame, so
+ elements can know that they have received all data for a frame
+ without waiting for the beginning of the next frame. This is similar
+ to how the RTP marker flag is used in many RTP video mappings.
-- Lots of multicast-related fixes
+ The video encoder base class now also releases the internal stream
+ lock before pushing out data, so as to not block the input side of
+ things from processing more data in the meantime.
-- rtsp-auth: Add support for parsing .htdigest files
+Miscellaneous other changes and enhancements
+- it is now possible to modify the initial rank of plugin features
+ without modifying the source code or writing code to do so
+ programmatically via the GST_PLUGIN_FEATURE_RANK environment
+ variable. Users can adjust the rank of plugin(s) by passing a
+ comma-separated list of feature:rank pairs where rank can be a
+ numerical value or one of NONE, MARGINAL, SECONDARY, PRIMARY, and
+ MAX. Example: GST_PLUGIN_FEATURE_RANK=myh264dec:MAX,avdec_h264:NONE
+ sets the rank of the myh264dec element feature to the maximum and
+ that of avdec_h264 to 0 (none), thus ensuring that myh264dec is
+ prefered as H264 decoder in an autoplugging context.
-GStreamer VAAPI
+- GstDeviceProvider now does a static probe on start as fallback for
+ providers that don’t support dynamic probing to make things easier
+ for users
-- Support Wayland’s display for context sharing, so the application
- can pass its own wl_display in order to be used for the VAAPI
- display creation.
+WebRTC
-- A lot of work to support new Intel hardware using media-driver as VA
- backend.
+- webrtcbin now contains initial support for renegotiation involving
+ stream addition and removal. There are a number of caveats to this
+ initial renegotiation support and many complex scenarios are known
+ to require some work.
-- For non-x86 devices, VAAPI display can instantiate, through DRM,
- with no PCI bus. This enables the usage of libva-v4l2-request
- driver.
+- webrtcbin now exposes the internal ICE object for advanced
+ configuration options. Using the internal ICE object, it is possible
+ to toggle UDP or TCP connection usage as well as provide local
+ network addresses.
-- Added support for XDG-shell protocol as wl_shell replacement which
- is currently deprecated. This change add as dependency
- wayland-protocol.
+- Fix a number of call flows within webrtcbin’s GstPromise handling
+ where a promise was never replied to. This has been fixed and now a
+ promise will always receive a reply.
-- GstVaapiFilter, GstVaapiWindow, and GstVaapiDecoder classes now
- inherit from GstObject, gaining all the GStreamer’s instrumentation
- support.
+- webrtcbin now exposes a latency property for configuring the
+ internal rtpjitterbuffer latency and buffering when receiving
+ streams.
-- The metadata now specifies the plugin as Hardware class.
+- webrtcbin now only synchronises the RTP part of a stream, allowing
+ RTCP messages to skip synchronisation entirely.
-- H264 decoder is more stable with problematic streams.
+- Fixed most of the webrtcbin state properties (connection-state,
+ ice-connection-state, signaling-state, but not ice-gathering-state
+ as that requires newer API in libnice and will be fixed in the next
+ release series) to advance through the state values correctly. Also
+ implemented DTLS connection states in the DTLS elements so that
+ peer-connection-state is not always new.
-- In H265 decoder added support for profiles main-422-10 (P010_10LE),
- main-444 (AYUV) and main-444-10 (Y410)
+- webrtcbin now accounts for the a=ice-lite attribute in a remote SDP
+ offer and will configure the internal ICE implementation
+ accordingly.
-- JPEG decoder handles dynamic resolution changes.
+- webrtcbin will now resolve .local candidate addresses using the
+ system DNS resolver. .local candidate addresses are now produced by
+ web browsers to help protect the privacy of users.
-- More specification adherence in H264 and H265 encoders.
+- webrtcbin will now add candidates found in the SDP to the internal
+ ICE agent. This was previously unsupported and required using the
+ add-ice-candidate signal manually from the application.
+- webrtcbin will now correctly parse a TURN URI that contains a
+ username or password with a : in it.
-GStreamer OMX
+- The GStreamer WebRTC library gained a GstWebRTCDataChannel object
+ roughly matching the interface exposed by the WebRTC specification
+ to allow for easier binding generation and use of data channels.
-- Add support of NV16 format to video encoders input.
+OpenGL integration
-- Video decoders now handle the ALLOCATION query to tell upstream
- about the number of buffers they require. Video encoders will also
- use this query to adjust their number of allocated buffers
- preventing starvation when using dynamic buffer mode.
+GStreamer OpenGL bindings/build related changes
-- The OMX_PERFORMANCE debug category has been renamed to OMX_API_TRACE
- and can now be used to track a widder variety of interactions
- between OMX and GStreamer.
+- The GStreamer OpenGL library (libgstgl) now ships pkg-config files
+ for platform-specific API where libgstgl provides a public
+ integration interface and a pkg-config file for a dependency on the
+ detected OpenGL headers. The new list of pkg-config files in
+ addition to the original gstreamer-gl-1.0 are gstreamer-gl-x11-1.0,
+ gstreamer-gl-wayland-1.0, gstreamer-gl-egl-1.0, and
+ gstreamer-gl-prototypes-1.0 (for OpenGL headers when including
+ gst/gl/gstglfuncs.h).
-- Video encoders will now detect frame rate only changes and will
- inform OMX about it rather than doing a full format reset.
+- GStreamer OpenGL now ships some platform-specific introspection data
+ for platforms that have a public interface. This should allow for
+ easier integration with bindings involving platform specific
+ functionality. The new introspection data files are named
+ GstGLX11-1.0, GstGLWayland-1.0, and GstGLEGL-1.0.
-- Various Zynq UltraScale+ specific improvements:
- - Video encoders are now able to import dmabuf from upstream.
- - Support for HEVC range extension profiles and more AVC profiles.
- - We can now request video encoders to generate an IDR using the
- force key unit event.
+GStreamer OpenGL Features
+- The iOS implementation no longer accesses UIKit objects off the main
+ thread fixing a loud warning message when used in iOS applications.
-GStreamer Editing Services and NLE
+- Support for mouse and keyboard handling using the GstNavigation
+ interface was added for the wayland implementation complementing the
+ already existing support for the X11 and Windows implementations.
-- Added a gesdemux element, it is an auto pluggable element that
- allows decoding edit list like files supported by GES
+- A new helper base class for source elements, GstGLBaseSrc is
+ provided to ease writing source elements producing OpenGL video
+ frames.
-- Added gessrc which wraps a GESTimeline as a standard source element
- (implementing the ges protocol handler)
+- Support for some more 12-bit and 16-bit video formats (Y412_LE,
+ Y412_BE, Y212_LE, Y212_BE, P012_LE, P012_BE, P016, NV16, NV61) was
+ added to glcolorconvert.
-- Added basic support for videorate::rate property potentially
- allowing changing playback speed
+- glupload can now import dma-buf’s into external-oes textures.
-- Layer priority is now fully automatic and they should be moved with
- the new ges_timeline_move_layer method, ges_layer_set_priority is
- now deprecated.
+- A new display type for EGLDevice-based systems was added. It is
+ currently opt-in by using either the GST_GL_PLATFORM=egl-device
+ environment variable or manual construction
+ (gst_gl_display_egl_device_new*()) due to compatibility issues with
+ some platforms.
-- Added a ges_timeline_element_get_layer_priority so we can simply get
- all information about GESTimelineElement position in the timeline
+- Support was added for WinRT/UWP using the ANGLE project for running
+ OpenGL-based pipelines within a UWP application.
-- GESVideoSource now auto orientates the images if it is defined in a
- meta (overridable).
+- Various elements now support changing the GstGLDisplay to be used at
+ runtime in simple cases. This is primarily helpful for changing or
+ adding an OpenGL-based video sink that must share an OpenGL context
+ with an external source to an already running pipeline.
-- Added some PyGObject overrides to make the API more pythonic
+GStreamer Vulkan integration
-- The threading model has been made more explicit with safe guard to
- make sure not thread safe APIs are not used from the wrong threads.
- It is also now possible to properly handle in what thread the API
- should be used.
+- There is now a GStreamer Vulkan library to provide integration
+ points and helpers with applications and external GStreamer Vulkan
+ based elements. The structure of the library is modelled similarly
+ to the already existing GStreamer OpenGL library. Please note that
+ the API is still unstable and may change in future releases,
+ particularly around memory handling. The GStreamer Vulkan library
+ contains objects for sharing the vkInstance, vkDevice, vkQueue,
+ vkImage, VkMemory, etc with other elements and/or the application as
+ well as some helper objects for using Vulkan in an application or
+ element.
-- Optimized GESClip and GESTrackElement creation
+- Added support for building and running on/for the Android and
+ Windows systems to complement the existing XCB, Wayland, MacOS, and
+ iOS implementations.
-- Added a way to compile out the old, unused and deprecated
- GESPitiviFormatter
+- XCB gained support for mouse/keyboard events using the GstNavigation
+ API.
-- Re implemented the timeline editing API making it faster and making
- the code much more maintainable
+- New vulkancolorconvert element for converting between color formats.
+ vulkancolorconvert can currently convert to/from all 8-bit RGBA
+ formats as well as 8-bit RGBA formats to/from the YUV formats AYUV,
+ NV12, and YUY2.
-- Simplified usage of nlecomposition outside GES by removing quirks in
- it API usage and removing the need to treat it specially from an
- application perspective.
+- New vulkanviewconvert element for converting between stereo view
+ layouts. vulkanviewconvert can currently convert between all of the
+ single memory formats (side-by-side, top-bottom, column-interleaved,
+ row-interleaved, checkerboard, left, right, mono).
-- ges-launch-1.0:
+- New vulkanimageidentity element for a blit from the input vulkan
+ image/s to a new vulkan image/s.
- - Added support to add titles to the timeline
- - Enhance the help auto generating it from the code
+- The vulkansink element can now scale the input image to the output
+ window/surface size where that information is available.
-- Deprecate ges_timeline_load_from_uri as loading the timeline should
- be done through a project now
+- The vulkanupload element can now configure a transfer from system
+ memory to VulkanImage-based memory. Previously, this required two
+ vulkanupload elements.
-- MANY leaks have been plugged and the unit testsuite is now “leak
- free”
+Tracing framework and debugging improvements
+- gst_tracing_get_active_tracers() returns a list of active tracer
+ objects. This can be used to interact with tracers at runtime using
+ GObject API such as action signals. This has been implemented in the
+ leaks tracer for snapshotting and retrieving leaked/active objects
+ at runtime.
+
+- The leaks tracer can now be interacted with programmatically at
+ runtime via GObject action signals:
+
+ - get-live-object returns a list of live (allocated) traced
+ objects
+ - log-live-objects logs a list of live objects into the debug log.
+ This is the same as sending the SIGUSR1 signal on unix systems,
+ but works on all operating systems including Windows.
+ - activity-start-tracking, activity-get-checkpoint,
+ activity-log-checkpoint, activity-stop-tracking: add support for
+ tracking and checkpointing objects, similar to what was
+ previously available via SIGUSR2 on unix systems, but works on
+ all operating systems including Windows.
+
+- various GStreamer gdb debug helper improvements:
+
+ - new ‘gst-pipeline-tree’ command
+ - more gdb helper functions: gst_element_pad(), gst_pipeline() and
+ gst_bin_get()
+ - support for queries and buffers
+ - print more info for segment events, print event seqnums, object
+ pointers and structures
+ - improve gst-print command to show more pad and element
+ information
-GStreamer validate
+Tools
-- Added an action type to verify the checksum of the sink last-sample
+gst-launch-1.0
-- Added an include keyword to validate scenarios
+- now prints the pipeline position and duration if available when the
+ pipeline is advancing. This is hopefully more user-friendly and
+ gives visual feedback on the terminal that the pipeline is actually
+ up and running. This can be disabled with the --no-position command
+ line option.
-- Added the notion of variable in scenarios, with the set-vars keyword
+- the parse-launch pipeline syntax now has support for presets:
+ use@preset=<preset-name>" after an element to load a preset.
-- Started adding support for “performance” like tests by allowing to
- define the number of dropped buffers or the minimum buffer frequency
- on a specific pad
+gst-inspect-1.0
-- Added a validateflow plugin which allows defining the data flow to
- be seen on a particular pad and verifying that following runs match
- the expectations
+- new --color command line option to force coloured output even if not
+ connected to a tty
-- Added support for appsrc based test definition so we can instrument
- the data pushed into the pipeline from scenarios
+gst-tester-1.0 (new)
-- Added a mockdecryptor allowing adding tests with on encrypted files,
- the element will potentially be instrumented with a validate
- scenario
+- gst-tester-1.0 is a new tool for plugin developers to launch
+ .validatetest files with TAP compatible output, meaning it can
+ easily and cleanly be integrated with the meson test harness. It
+ allows you to use gst-validate (from the gst-devtools module) to
+ write integration tests in any GStreamer repository whilst keeping
+ the tests as close as possible to the code. The tool transparently
+ handles gst-validate being installed or not: if it is not installed
+ those integration tests will simply be skipped.
-- gst-validate-launcher:
+gst-play-1.0
- - Cleaned up output
+- interactive keyboard controls now also work on Windows
- - Changed the default for “muting” tests as user doesn’t expect
- hundreds of windows to show up when running the testsuite
+gst-transcoder-1.0 (new)
- - Fixed the outputted xunit files to be compatible with GitLab
+- gst-transcoder-1.0 is a new command line tool to transcode one URI
+ into another URI based on the specified encoding profile using the
+ new GstTranscoder API (see above).
- - Added support to run tests on media files in push mode (using
- pushfile://)
+GStreamer RTSP server
- - Added support for running inside gst-build
+- Fix issue where the first few packets (i.e. keyframes) could
+ sometimes be dropped if the rtsp media pipeline had a live input.
+ This was a regression from GStreamer 1.14. There are more fixes
+ pending for that which will hopefully land in 1.18.1.
+
+- Fix backpressure handling when sending data in TCP interleave mode
+ where RTSP requests and responses and RTP/RTCP packets flow over the
+ same RTSP TCP connection: The previous implementation would at some
+ point stop sending data to other clients when a single client
+ stopped consuming data or did not consume data fast enough. This
+ obviously created problems for shared media, where the same stream
+ from a single producer pipeline is sent to multiple clients. Instead
+ we now manage a backlog in the server’s stream-transport component
+ and remove slow clients once this backlog exceeds a maximum duration
+ (which is currently hardcoded).
+
+- Onvif Streaming Specification trick modes support (see section at
+ the beginning)
+
+- Scale/Speed header support: Speed will deliver the data at the
+ requested speed, which means increasing the data bandwidth for
+ speeds > 1.0. Scale will attempt to do the same without affecting
+ the overall bandwidth requirement vis-a-vis normal playback speed
+ (e.g. it might drop data for fast-forward playback).
+
+- rtspclientsink: send buffer lists in one go for better performance
- - Added support for running ssim tests on rendered files
+GStreamer VAAPI
- - Added a way to simply define tests on pipelines through a simple
- .json file
+- A lot of work was done adding support for media-driver (iHD), the
+ new VAAPI driver for Intel, mostly for Gen9 onwards.
- - Added a python app to easily run python testsuite reusing all
- the launcher features
+- Available color formats and frame sizes are now detected at run-time
+ according to the context configuration.
- - Added flatpak knowledge so we can print backtrace even when
- running from within flatpak
+- Gallium drivers have been re-enabled in the allowed drivers list
- - Added a way to automatically generated “known issues”
- suppressions lines
+- Improved the mapping between VA formats and GStreamer formats by
+ generating a mapping table at run-time since even among different
+ drivers the mapping might be different, particularly for RGB with
+ little endianness.
- - Added a way to rerun tests to check if they are flaky and added
- a way to tolerate tests known to be flaky
+- The experimental Flexible Encoding Infrastructure (FEI) elements
+ have been removed since they were not really actively maintained or
+ tested.
- - Add a way to output html log files
+- Enhanced the juggling of DMABuf buffers and VASurface metas
+- New vaapioverlay element: a compositor element using VA VPP blend
+ capabilities to accelerate overlaying and compositing. Example
+ pipeline:
-GStreamer Python Bindings
+ gst-launch-1.0 -vf videotestsrc ! vaapipostproc ! tee name=testsrc ! queue \
+ ! vaapioverlay sink_1::xpos=300 sink_1::alpha=0.75 name=overlay ! vaapisink \
+ testsrc. ! queue ! overlay.
-- add binding for gst_pad_set_caps()
+vaapipostproc
-- pygobject dependency requirement was bumped to >= 3.8
+- added video-orientation support, supporting frame mirroring and
+ rotation
-- new audiotestsrc, audioplot, and mixer plugin examples, and a
- dynamic pipeline example
+- added cropping support, either via properties (crop-left,
+ crop-right, crop-bottom and crop-top) or buffer meta.
+- new skin-tone-enhancenment-level property which is the iHD
+ replacement of the i965 driver’s sink-tone-level. Both are
+ incompatible with each other, so both were kept.
-GStreamer C# Bindings
+- handle video colorimetry
-- bindings for the GstWebRTC library
+- support HDR10 tone mapping
+vaapisink
-GStreamer Rust Bindings
+- resurrected wayland backend for non-weston compositors by extracting
+ the DMABuf from the VASurface and rendering it.
-The GStreamer Rust bindings are now officially part of the GStreamer
-project and are also maintained in the GStreamer GitLab.
+- merged the video overlay API for wayland. Now applications can
+ define the “window” to render on.
-The releases will generally not be synchronized with the releases of
-other GStreamer parts due to dependencies on other projects.
+- demoted the vaapisink element to secondary rank since libva
+ considers rendering as a second-class feature.
-Also unlike the other GStreamer libraries, the bindings will not commit
-to full API stability but instead will follow the approach that is
-generally taken by Rust projects, e.g.:
+VAAPI Encoders
-1) 0.12.X will be completely API compatible with all other 0.12.Y
- versions.
-2) 0.12.X+1 will contain bugfixes and compatible new feature additions.
-3) 0.13.0 will _not_ be backwards compatible with 0.12.X but projects
- will be able to stay at 0.12.X without any problems as long as they
- don’t need newer features.
+- new common target-percentage property which is the desired target
+ percentage of bitrate for variable rate control.
-The current stable release is 0.12.2 and the next release series will be
-0.13, probably around March 2019.
+- encoders now extract their caps from the driver at registration
+ time.
-At this point the bindings cover most of GStreamer core (except for most
-notably GstAllocator and GstMemory), and most parts of the app, audio,
-base, check, editing-services, gl, net. pbutils, player, rtsp,
-rtsp-server, sdp, video and webrtc libraries.
+- vaapivp9enc: added support for low power mode and support for
+ profile 2 (profile 0 by default)
-Also included is support for creating subclasses of the following types
-and writing GStreamer plugins:
+- vaapih264enc: new max-qp property that sets the maximum quantization
+ value. Support for ICQ and QBVR bitrate control mode, adding a
+ quality-factor property for these modes. Support baseline profile as
+ constrained-baseline
-- gst::Element
-- gst::Bin and gst::Pipeline
-- gst::URIHandler and gst::ChildProxy
-- gst::Pad, gst::GhostPad
-- gst_base::Aggregator and gst_base::AggregatorPad
-- gst_base::BaseSrc and gst_base::BaseSink
-- gst_base::BaseTransform
+- vaapih265enc:
-Changes to 0.12.X since 0.12.0
+ - support for main-444 and main-12 encoding profiles.
+ - new max-qp property that sets the maximum quantization value.
+ - support for ICQ and QBVR bitrate control mode, adding a
+ quality-factor property for these modes.
+ - handle SCC profiles.
+ - num-tile-cols and num-tile-row properties to specify the number
+ of tiles to use.
+ - the low-delay-b property was deprecated and is now determined
+ automatically.
+ - improved profile selection through caps.
-Fixed
+VAAPI Decoders
-- PTP clock constructor actually creates a PTP instead of NTP clock
+- Decoder surfaces are not bound to their context any longer and can
+ thus be created and used dynamically, removing the deadlock
+ headache.
-Added
+- Reverse playback is now fluid
-- Bindings for GStreamer Editing Services
-- Bindings for GStreamer Check testing library
-- Bindings for the encoding profile API (encodebin)
+- Forward Region-of-Interest (ROI) metas downstream
-- VideoFrame, VideoInfo, AudioInfo, StructureRef implements Send and
- Sync now
-- VideoFrame has a function to get the raw FFI pointer
-- From impls from the Error/Success enums to the combined enums like
- FlowReturn
-- Bin-to-dot file functions were added to the Bin trait
-- gst_base::Adapter implements SendUnique now
-- More complete bindings for the gst_video::VideoOverlay interface,
- especially
- gst_video::is_video_overlay_prepare_window_handle_message()
+- GLTextureUploadMeta uses DMABuf when GEM is not available. Now
+ Gallium drivers can use this meta for rendering with EGL.
-Changed
+- vaapivp9dec: support for 4:2:2 and 4:4:4 chroma type streams
-- All references were updated from GitHub to freedesktop.org GitLab
-- Fix various links in the README.md
-- Link to the correct location for the documentation
-- Remove GitLab badge as that only works with gitlab.com currently
+- vaapih265dec: skip all pictures prior to the first I-frame. Enable
+ passing range extension flags to the driver. Handle SCC profiles.
-Changes in git master for 0.13
+- vaapijpegdec: support for 4:0:0, 4:1:1, 4:2:2 and 4:4:4 chroma types
+ pictures
-Fixed
+- vaapih264dec: handle baseline streams as constrained-baseline if
+ possible and make it more tolerant when encountering unknown NALs
-- gst::tag::Album is the album tag now instead of artist sortname
+GStreamer OMX
-Added
+- omxvideoenc: use new video encoder subframe API to push out slices
+ as soon as they’re ready
-- Subclassing infrastructure was moved directly into the bindings,
- making the gst-plugin crate deprecated. This involves many API
- changes but generally cleans up code and makes it more flexible.
- Take a look at the gst-plugins-rs crate for various examples.
+- omxh264enc, omxh265enc: negotiate subframe mode via caps. To enable
+ it, force downstream caps to video/x-h264,alignment=nal or
+ video/x-h265,alignment=nal.
-- Bindings for CapsFeatures and Meta
-- Bindings for
- ParentBufferMeta,VideoMetaandVideoOverlayCompositionMeta`
-- Bindings for VideoOverlayComposition and VideoOverlayRectangle
-- Bindings for VideoTimeCode
+- omxh264enc: Add ref-frames property
-- UniqueFlowCombiner and UniqueAdapter wrappers that make use of the
- Rust compile-time mutability checks and expose more API in a safe
- way, and as a side-effect implement Sync and Send now
+- Zynq ultrascale+ specific video encoder/decoder improvements:
-- More complete bindings for Allocation Query
-- pbutils functions for codec descriptions
-- TagList::iter() for iterating over all tags while getting a single
- value per tag. The old ::iter_tag_list() function was renamed to
- ::iter_generic() and still provides access to each value for a tag
-- Bus::iter() and Bus::iter_timed() iterators around the corresponding
- ::pop\*() functions
+ - GRAY8 format support
+ - support for alternate fields interlacing mode
+ - video encoder: look-ahead, long-term-ref, and long-term-freq
+ properties
-- serde serialization of Value can also handle Buffer now
+GStreamer Editing Services and NLE
-- Extensive comments to all examples with explanations
-- Transmuxing example showing how to use typefind, multiqueue and
- dynamic pads
-- basic-tutorial-12 was ported and added
+- Added nested timelines and subproject support so that GES projects
+ can be used as clips, potentially serializing nested projects in the
+ main file or referencing external project files.
-Changed
+- Implemented an OpenTimelineIO GES formatter. This means GES and
+ GStreamer can now load and save projects in all the formats
+ supported by otio.
-- Rust 1.31 is the minimum supported Rust version now
-- Update to latest gir code generator and glib bindings
+- Implemented a GESMarkerList object which allow setting timed
+ metadata on any GES object.
-- Functions returning e.g. gst::FlowReturn or other “combined” enums
- were changed to return split enums like
- Result<gst::FlowSuccess, gst::FlowError> to allow usage of the
- standard Rust error handling.
+- Fixed audio rendering issues during clip transition by ensuring that
+ a single segment is pushed into encoders.
-- MiniObject subclasses are now newtype wrappers around the underlying
- GstRc<FooRef> wrapper. This does not change the API in any breaking
- way for the current usages, but allows MiniObjects to also be
- implemented in other crates and makes sure rustdoc places the
- documentation in the right places.
+- The GESUriClipAsset API is now MT safe.
-- BinExt extension trait was renamed to GstBinExt to prevent conflicts
- with gtk::Bin if both are imported
+- Added ges_meta_container_register_static_meta() to allow fixing a
+ type for a specific metadata without actually setting a value.
-- Buffer::from_slice() can’t possible return None
+- The framepositioner element now handles resizing the project and
+ keeps the same positioning when the aspect ratio is not changed .
-- Various clippy warnings
+- Reworked the documentation, making it more comprehensive and much
+ more detailed.
+- Added APIs to retrieve natural size and framerate of a clip (for
+ example in the case of URIClip it is the framerate/size of the
+ underlying file).
-GStreamer Rust Plugins
+- ges_container_edit() is now deprecated and GESTimelineElement gained
+ the ges_timeline_element_edit() method so the editing API is now
+ usable from any element in the timeline.
-Like the GStreamer Rust bindings, the Rust plugins are now officially
-part of the GStreamer project and are also maintained in the GStreamer
-GitLab.
+- GESProject::loading was added so applications can be notified about
+ when a new timeline starts loading.
-In the 0.3.x versions this contained infrastructure for writing
-GStreamer plugins in Rust, and a set of plugins.
+- Implemented the GstStream API in GESTimeline.
-In git master that infrastructure was moved to the GLib and GStreamer
-bindings directly, together with many other improvements that were made
-possible by this, so the gst-plugins-rs repository only contains
-GStreamer elements now.
+- Added a way to add a timeoverlay inside the test source (potentially
+ with timecodes).
-Elements included are:
+- Added APIs to convert times to frame numbers and vice versa:
-- Tutorials plugin: identity, rgb2gray and sinesrc with extensive
- comments
+ - ges_timeline_get_frame_time()
-- rsaudioecho, a port of the audiofx element
+ - ges_timeline_get_frame_at()
-- rsfilesrc, rsfilesink
+ - ges_clip_asset_get_frame_time()
-- rsflvdemux, a FLV demuxer. Not feature-equivalent with flvdemux yet
+ - ges_clip_get_timeline_time_from_source_frame()
-- threadshare plugin: ts-appsrc, ts-proxysrc/sink, ts-queue, ts-udpsrc
- and ts-tcpclientsrc elements that use a fixed number of threads and
- share them between instances. For more background about these
- elements see Sebastian’s talk “When adding more threads adds more
- problems - Thread-sharing between elements in GStreamer” at the
- GStreamer Conference 2017.
+ Quite a few validate tests have been implemented to check the
+ behavior for various demuxer/codec formats
-- rshttpsrc, a HTTP source around the hyper/reqwest Rust libraries.
- Not feature-equivalent with souphttpsrc yet.
+- Added ges_layer_set_active_for_tracks() which allows muting layers
+ for the specified tracks
-- togglerecord, an element that allows to start/stop recording at any
- time and keeps all audio/video streams in sync.
+- Deprecated GESImageSource and GESMultiFileSource now that we have
+ imagesequencesrc which handles the imagesequence “protocol”
-- mccparse and mccenc, parsers and encoders for the MCC closed caption
- file format.
+- Stopped exposing ‘deinterlacing’ children properties for clip types
+ where they do not make sense.
-Changes to 0.3.X since 0.3.0
+- Added support for simple time remapping effects
-- All references were updated from GitHub to freedesktop.org GitLab
-- Fix various links in the README.md
-- Link to the correct location for the documentation
+GStreamer validate
-Changes in git master for 0.4
+- Introduced the concept of “Test files” allowing to implement “all
+ included” test cases, meaning that inside the file the following can
+ be defined:
-- togglerecord: Switch to parking_lot crate for mutexes/condition
- variables for lower overhead
-- Merge threadshare plugin here
-- New closedcaption plugin with mccparse and mccenc elements
-- New identity element for the tutorials plugin
+ - The application arguments
+ - The validate configurations
+ - The validate scenario
-- Register plugins statically in tests instead of relying on the
- plugin loader to find the shared library in a specific place
+ This replaces the previous big dictionary file in
+ gst-validate-launcher to implement specific test cases.
-- Update to the latest API changes in the GLib and GStreamer bindings
-- Update to the latest versions of all crates
+ We set several variables inside the files (as well as inside
+ scenarios and config files) to make them relocatable.
+ The file format has been enhanced so it is easier to read and write,
+ for example line ending with a coma or (curly) brackets can now be
+ used as continuation marker so you do not need to add \ at the end
+ of lines to write a structure on several lines.
-Build and Dependencies
+- Support the imagesequence “protocol” and added integration tests for
+ it.
-- The MESON BUILD SYSTEM BUILD IS NOW FEATURE-COMPLETE (*) and it is
- now the recommended build system on all platforms and also used by
- Cerbero to build GStreamer on all platforms. The Autotools build is
- scheduled to be removed in the next cycle. Developers who currently
- use gst-uninstalled should move to gst-build. The build option
- naming has been cleaned up and made consistent and there are now
- feature options to enable/disable plugins and various other features
- on a case-by-case basis. (*) with the exception of plugin docs which
- will be handled differently in future
-
-- Symbol export in libraries is now controlled via explicit exports
- using symbol visibility or export defines where supported, to ensure
- consistency across all platforms. This also allows libraries to have
- exports that vary based on detected platform features and configure
- options as is the case with the GStreamer OpenGL integration library
- for example. A few symbols that had been exported by accident in
- earlier versions may no longer be exported. These symbols will not
- have had declarations in any public header files then though and
- would not have been usable.
-
-- The GStreamer FFmpeg wrapper plugin (gst-libav) now depends on
- FFmpeg 4.x and uses the new FFmpeg 4.x API and stopped relying on
- ancient API that was removed with the FFmpeg 4.x release. This means
- that it is no longer possible to build this module against an older
- system-provided FFmpeg 3.x version. Use the internal FFmpeg 4.x copy
- instead if you build using autotools, or use gst-libav 1.14.x
- instead which targets the FFmpeg 3.x API and _should_ work fine in
- combination with a newer GStreamer. It’s difficult for us to support
- both old and new FFmpeg APIs at the same time, apologies for any
- inconvenience caused.
-
-- Hardware-accelerated Nvidia video encoder/decoder plugins nvdec and
- nvenc can be built against CUDA Toolkit versions 9 and 10.0 now. The
- dynlink interface has been dropped since it’s deprecated in 10.0.
-
-- The (optional) OpenCV requirement has been bumped to >= 3.0.0 and
- the plugin can also be built against OpenCV 4.x now.
-
-- New sctp plugin based on usrsctp (for WebRTC data channels)
+- Added action types to allow the scenario to run the Test Clock for
+ better reproducibility of tests.
-Cerbero
+- Support generating tests to check that seeking is frame accurate
+ (base on ssim).
-Cerbero is a meta build system used to build GStreamer plus dependencies
-on platforms where dependencies are not readily available, such as
-Windows, Android, iOS and macOS.
+- Added ways to record buffers checksum (in different ways) in the
+ validateflow module.
-Cerbero has seen a number of improvements:
+- Added vp9 encoding tests.
-- Cerbero has been ported to Python 3 and requires Python 3.5 or newer
- now
+- Enhanced seeking action types implementation to allow support for
+ segment seeks.
-- Source tarballs are now protected by checksums in the recipes to
- guard against download errors and malicious takeover of projects or
- websites. In addition, downloads are only allowed via secure
- transports now and plain HTTP, FTP and git:// transports are not
- allowed anymore.
+- Output improvements:
-- There is now a new fetch-bootstrap command which downloads sources
- required for bootstrapping, with an optional --build-tools-only
- argument to match the bootstrap --build-tools-only command.
+ - Logs are now in markdown formats (and bat is used to dump them
+ if available).
+ - File format issues in scenarios/configs/tests files are nicely
+ reported with the line numbers now.
-- The bootstrap, build, package and bundle-source commands gained a
- new --offline switch that ensures that only sources from the cache
- are used and never downloaded via the network. This is useful in
- combination with the fetch and fetch-bootstrap commands that acquire
- sources ahead of time before any build steps are executed. This
- allows more control over the sources used and when sources are
- updated, and is particularly useful for build environments that
- don’t have network access.
+GStreamer Python Bindings
-- bootstrap --assume-yes will automatically say ‘yes’ to any
- interactive prompts during the bootstrap stage, such as those from
- apt-get or yum.
+- Python 2.x is no longer supported
-- bootstrap --system-only will only bootstrap the system without build
- tools.
+- Support mapping buffers without any memcpy:
-- Manifest support: The build manifest can be used in continuous
- integration (CI) systems to fixate the Git revision of certain
- projects so that all builds of a pipeline are on the same reference.
- This is used in GStreamer’s gitlab CI for example. It can also be
- used in order to re-produce a specific build. To set a manifest, you
- can set manifest = 'my_manifest.xml' in your configuration file, or
- use the --manifest command line option. The command line option will
- take precendence over anything specific in the configuration file.
+ - Added a ContextManager to make the API more pythonic
-- The new build-deps command can be used to build only the
- dependencies of a recipe, without the recipe itself.
+ with buf.map(Gst.MapFlags.READ | Gst.MapFlags.WRITE) as info:
+ info.data[42] = 0
-- new --list-variants command to list available variants
+- Added high-level helper API for constructing pipelines:
-- variants can now be set on the command line via the -v option as a
- comma-separated list. This overrides any variants set in any
- configuration files.
+ - Gst.Bin.make_and_add(factory_name, instance_name=None)
+ - Gst.Element.link_many(element, ...)
-- new qt5, intelmsdk and nvidia variants for enabling Qt5 and hardware
- codec support. See the Enabling Optional Features with Variants
- section in the Cerbero documentation for more details how to enable
- and use these variants.
+GStreamer C# Bindings
-- A new -t / --timestamp command line switch makes commands print
- timestamps
+- Bind gst_buffer_new_wrapped() manually to fix memory handling.
+
+- Fix gst_promise_new_with_change_func() where bindgen didn’t properly
+ detect the func as a closure.
+
+- Declare GstVideoOverlayComposition and GstVideoOverlayRectangle as
+ opaque type and subclasses of Gst.MiniObject. This changes the API
+ but without this all usage will cause memory corruption or simply
+ not work.
+
+- on Windows, look for gstreamer, glib and gobject DLLs using the MSVC
+ naming convention (i.e. gstvideo-1.0-0.dll instead of
+ libgstvideo-1.0-0.dll).
+
+ The names of these DLLs have to be hardcoded in the bindings, and
+ most C# users will probably be using the Microsoft toolchain anyway.
+
+ This means that the MSVC compiler is now required to build the
+ bindings, MingW will no longer work out of the box.
+
+GStreamer Rust Bindings and Rust Plugins
+
+The GStreamer Rust bindings are released separately with a different
+release cadence that’s tied to gtk-rs, but the latest release has
+already been updated for the new GStreamer 1.18 API, so there’s
+absolutely no excuse why your next GStreamer application can’t be
+written in Rust anymore.
+
+gst-plugins-rs, the module containing GStreamer plugins written in Rust,
+has also seen lots of activity with many new elements and plugins.
+
+What follows is a list of elements and plugins available in
+gst-plugins-rs, so people don’t miss out on all those potentially useful
+elements that have no C equivalent.
+
+Rust audio plugins
+
+- audiornnoise: New element for audio denoising which implements the
+ noise removal algorithm of the Xiph RNNoise library, in Rust
+- rsaudioecho: Port of the audioecho element from gst-plugins-good
+ rsaudioloudnorm: Live audio loudness normalization element based on
+ the FFmpeg af_loudnorm filter
+- claxondec: FLAC lossless audio codec decoder element based on the
+ pure-Rust claxon implementation
+- csoundfilter: Audio filter that can use any filter defined via the
+ Csound audio programming language
+- lewtondec: Vorbis audio decoder element based on the pure-Rust
+ lewton implementation
+
+Rust video plugins
+
+- cdgdec/cdgparse: Decoder and parser for the CD+G video codec based
+ on a pure-Rust CD+G implementation, used for example by karaoke CDs
+- cea608overlay: CEA-608 Closed Captions overlay element
+- cea608tott: CEA-608 Closed Captions to timed-text (e.g. VTT or SRT
+ subtitles) converter
+- tttocea608: CEA-608 Closed Captions from timed-text converter
+- mccenc/mccparse: MacCaption Closed Caption format encoder and parser
+- sccenc/sccparse: Scenarist Closed Caption format encoder and parser
+- dav1dec: AV1 video decoder based on the dav1d decoder implementation
+ by the VLC project
+- rav1enc: AV1 video encoder based on the fast and pure-Rust rav1e
+ encoder implementation
+- rsflvdemux: Alternative to the flvdemux FLV demuxer element from
+ gst-plugins-good, not feature-equivalent yet
+- rsgifenc/rspngenc: GIF/PNG encoder elements based on the pure-Rust
+ implementations by the image-rs project
+
+Rust text plugins
+
+- textwrap: Element for line-wrapping timed text (e.g. subtitles) for
+ better screen-fitting, including hyphenation support for some
+ languages
+
+Rust network plugins
+
+- reqwesthttpsrc: HTTP(S) source element based on the Rust
+ reqwest/hyper HTTP implementations and almost feature-equivalent
+ with the main GStreamer HTTP source souphttpsrc
+- s3src/s3sink: Source/sink element for the Amazon S3 cloud storage
+- awstranscriber: Live audio to timed text transcription element using
+ the Amazon AWS Transcribe API
+
+Generic Rust plugins
+
+- sodiumencrypter/sodiumdecrypter: Encryption/decryption element based
+ on libsodium/NaCl
+- togglerecord: Recording element that allows to pause/resume
+ recordings easily and considers keyframe boundaries
+- fallbackswitch/fallbacksrc: Elements for handling potentially
+ failing (network) sources, restarting them on errors/timeout and
+ showing a fallback stream instead
+- threadshare: Set of elements that provide alternatives for various
+ existing GStreamer elements but allow to share the streaming threads
+ between each other to reduce the number of threads
+- rsfilesrc/rsfilesink: File source/sink elements as replacements for
+ the existing filesrc/filesink elements
+Build and Dependencies
-Platform-specific changes and improvements
+- The Autotools build system has finally been removed in favour of the
+ Meson build system. Developers who currently use gst-uninstalled
+ should move to gst-build.
+
+- API and plugin documentation are no longer built with gtk_doc. The
+ gtk_doc documentation has been removed in favour of a new unified
+ documentation module built with hotdoc (also see “Documentation
+ improvements” section below). Distributors should use the
+ documentation release tarball instead of trying to package hotdoc
+ and building the documentation from scratch.
+
+- gst-plugins-bad now includes an internal copy of libusrsctp, as
+ there are problems in usrsctp with global shared state, lack of API
+ stability guarantees, and the absence of any kind of release
+ process. We also can’t rely on distros shipping a version with the
+ fixes we need. Both firefox and Chrome bundle their own copies too.
+ It is still possible to build against an external copy of usrsctp if
+ so desired.
+
+- nvcodec no longer needs the NVIDIA NVDEC/NVENC SDKs available at
+ build time, only at runtime. This allows distributions to ship this
+ plugin by default and it will just start to work when the required
+ run-time SDK libraries are installed by the user, without users
+ needing to build and install the plugin from source.
+
+- the gst-editing-services tarball is now named gst-editing-services
+ for consistency (used to be gstreamer-editing-services).
+
+- the gst-validate tarball has been superseded by the gst-devtools
+ tarball for consistency with the git module name.
+
+gst-build
+
+gst-build is a meta-module and serves primarily as our uninstalled
+development environment. It makes it easy to build most of GStreamer,
+but unlike Cerbero it only comes with a limited number of external
+dependencies that can be built as subprojects if they are not found on
+the system.
+
+gst-build is based on Meson and replaces the old autotools
+gst-uninstalled script.
+
+- The ‘uninstalled’ target has been renamed to ‘devenv’
+
+- Experimental gstreamer-full library containing all built plugins and
+ their deps when building with -Ddefault_library=static. A monolithic
+ library is easier to distribute, and may be required in some
+ environments. GStreamer core, GLib and GObject are always included,
+ but external dependencies are still dynamically linked. The
+ gst-full-libraries meson option allows adding other GStreamer
+ libraries to the gstreamer-full build. This is an experiment for now
+ and its behaviour or API may still change in future releases.
+
+- Add glib-networking as a subproject when glib is a subproject and
+ load gio modules in the devenv, tls option control whether to use
+ openssl or gnutls.
+
+- git-worktree: Allow multiple worktrees for subproject branches
+
+- Guard against meson being run from inside the uninstalled devenv, as
+ this might have unexpected consequences.
+
+- our ffmpeg and x264 meson ports have been updated to the latest
+ stable version (you might need to update the subprojects checkout
+ manually though, or just remove the checkouts so meson checks out
+ the latest version again; improvements for this are pending in
+ meson, but not merged yet).
-Android
+Cerbero
-- toolchain: update compiler to clang and NDKr18. NDK r18 removed the
- armv5 target and only has Android platforms that target at least
- armv7 so the armv5 target is not useful anymore.
+Cerbero is a meta build system used to build GStreamer plus dependencies
+on platforms where dependencies are not readily available, such as
+Windows, Android, iOS and macOS.
-- The way that GIO modules are named has changed due to upstream GLib
- natively adding support for loading static GIO modules. This means
- that any GStreamer application using gnutls for SSL/TLS on the
- Android or iOS platforms (or any other setup using static libraries)
- will fail to link looking for the g_io_module_gnutls_load_static()
- function. The new function name is now
- g_io_gnutls_load(gpointer data). data can be NULL for a static
- library. Look at this commit for the necessary change in the
- examples.
+General improvements
+
+- Recipe build steps are done in parallel wherever possible. This
+ leads to massive improvements in overall build time.
+- Several recipes were ported to Meson, which improved build times
+- Moved from using both GnuTLS and OpenSSL to only OpenSSL
+- Moved from yasm to nasm for all assembly compilation
+- Support zsh when running the cerbero shell command
+- Numerous version upgrades for dependencies
+- Default to xz for tarball binary packages. bz2 can be selected with
+ the --compress-method option to package.
+- Added boolean variant for controlling the optimization level:
+ -v optimization
+- Ship .pc pkgconfig files for all plugins in the binary packages
+- CMake and nasm will only be built by Cerbero if the system versions
+ are unusable
+- The nvcodec variant was removed and the nvcodec plugin is built by
+ default now (as it no longer requires the SDK to be installed at
+ build time, only at runtime)
+
+macOS / iOS
+
+- Minimum iOS SDK version bumped to 11.0
+- Minimum macOS SDK version bumped to 10.11
+- No longer need to manually add support for newer iOS SDK versions
+- Added Vulkan elements via MoltenVK
+- Build times were improved by code-signing all build tools
+- macOS framework ships all gstreamer libraries instead of an outdated
+ subset
+- Ship pkg-config in the macOS framework package
+- fontconfig: Fix EXC_BAD_ACCESS crash on iOS ARM64
+- Improved App Store compatibility by setting LC_VERSION_MIN_MACOSX,
+ fixing relocations, and improved bitcode support
-- various build issues on Android have been fixed.
+Windows
-macOS and iOS
+- MinGW-GCC toolchain was updated to 8.2. It uses the Universal CRT
+ instead of MSVCRT which eliminates cross-CRT issues in the Visual
+ Studio build.
+- Require Windows 7 or newer for running binaries produced by Cerbero
+- Require Windows x86_64 for running Cerbero to build binary packages
+- Cerbero no longer uses C:/gstreamer/1.0 as a prefix when building.
+ That prefix is reserved for use by the MSI installers.
+- Several recipes can now be buit with Visual Studio instead of MinGW.
+ Ported to meson: opus, libsrtp, harfbuzz, cairo, openh264, libsoup,
+ libusrsctp. Existing build system: libvpx, openssl.
+- Support building using Visual Studio for 32-bit x86. Previously we
+ only supported building for 32-bit x86 using the MinGW toolchain.
+- Fixed annoying msgmerge popups in the middle of cerbero builds
+- Added configuration options vs_install_path and vs_install_version
+ for specifying custom search locations for older Visual Studio
+ versions that do not support vswhere. You can set these in
+ ~/.cerbero/cerbero.cbc where ~ is the MSYS homedir, not your Windows
+ homedir.
+- New Windows-specific plugins: d3d11, mediafoundation, wasapi2
+- Numerous compatibility and reliability fixes when running Cerbero on
+ Windows, especially non-English locales
+- proxy-libintl now exports the same symbols as gettext, which makes
+ it a drop-in replacement
+- New mapping variant for selecting the Visual Studio CRT to use:
+ -v vscrt=<value>. Valid values are md, mdd, and auto (default). A
+ separate prefix is used when building with either md (release) or
+ mdd (debug), and the outputted package will have +debug in the
+ filename. This variant is also used for selecting the correct Qt
+ libraries (debug vs release) to use when building with -v qt5 on
+ Windows.
+- Support cross-compile on Windows to Windows ARM64 and ARMv7
+- Support cross-compile on Windows to the Universal Windows Platform
+ (UWP). Only the subset of plugins that can be built entirely with
+ Visual Studio will be selected in this case. To do so, use the
+ config/cross-uwp-universal.cbc configuration, which will build
+ ARM64, x86, and x86_64 binaries linked to the release CRT, with
+ optimizations enabled, and debugging turned on. You can combine this
+ with -v vscrt=mdd to produce binaries linked to the debug CRT. You
+ can turn off optimizations with the -v nooptimization variant.
+
+Windows MSI installer
+
+- Require Windows 7 or newer for running GStreamer
+- Fixed some issues with shipping of pkg-config in the Windows
+ installers
+- Plugin PDB debug files are now shipped in the development package,
+ not the runtime package
+- Ship installers for 32-bit binaries built with Visual Studio
+- Ship debug and release “universal” (ARM64, X86, and X86_64) tarballs
+ built for the Universal Windows Platform
+- Windows MSI installers now install into separate prefixes when
+ building with MSVC and MinGW. Previously both would be installed
+ into C:/gstreamer/1.0/x86 or C:/gstreamer/1.0/x86_64. Now, the
+ installation prefixes are:
+
+ ----------------------------------------------------------------------------------------------------------------
+ Target Path Build options
+ --------------------------- ------------------------------------ -----------------------------------------------
+ MinGW 32-bit C:/gstreamer/1.0/mingw_x86 -c config/win32.cbc
+
+ MinGW 64-bit C:/gstreamer/1.0/mingw_x86_64 -c config/win64.cbc
+
+ MSVC 32-bit C:/gstreamer/1.0/msvc_x86 -c config/win32.cbc -v visualstudio
+
+ MSVC 64-bit C:/gstreamer/1.0/msvc_x86_64 -c config/win64.cbc -v visualstudio
+
+ MSVC 32-bit (debug) C:/gstreamer/1.0/msvc-debug_x86 -c config/win32.cbc -v visualstudio,vscrt=mdd
+
+ MSVC 64-bit (debug) C:/gstreamer/1.0/msvc-debug_x86_64 -c config/win64.cbc -v visualstudio,vscrt=mdd
+ ----------------------------------------------------------------------------------------------------------------
+
+Note: UWP binary packages are tarballs, not MSI installers.
+
+Linux
+
+- Support creating MSI installers using WiX when cross-compiling to
+ Windows
+- Support running cross-windows binaries with Wine when using the
+ shell and runit cerbero commands
+- Added bash-completion support inside the cerbero shell on Linux
+- Require a system-wide installation of openssl on Linux
+- Added variant -v vaapi to build gstreamer-vaapi and the new gstva
+ plugin
+- Debian packaging was disabled because it does not work. Help in
+ fixing this is appreciated.
+- Trimmed the list of packages needed for bootstrap on Linux
-- various build issues on iOS have been fixed.
+Android
-- the minimum required iOS version is now 9.0. The difference in
- adoption between 8.0 and 9.0 is 0.1% and the bump to 9.0 fixes some
- build issues.
+- Updated to NDK r21
+- Support Vulkan
+- Support Qt 5.14+ binary package layout
-- The way that GIO modules are named has changed due to upstream GLib
- natively adding support for loading static GIO modules. This means
- that any GStreamer application using gnutls for SSL/TLS on the
- Android or iOS platforms (or any other setup using static libraries)
- will fail to link looking for the g_io_module_gnutls_load_static()
- function. The new function name is now
- g_io_gnutls_load(gpointer data). data can be NULL for a static
- library. Look at this commit for the necessary change in the
- examples.
+Platform-specific changes and improvements
-Windows
+Android
-- The webrtcdsp element is shipped again as part of the Windows binary
- packages, the build system issue has been resolved.
+- opensles: Remove hard-coded buffer-/latency-time values and allow
+ openslessink to handle 48kHz streams.
-- ‘Inconsistent DLL linkage’ warnings when building with MSVC have
- been fixed
+- photography interface and camera source: Add additional settings
+ relevant to Android such as: Exposure mode property, extra colour
+ tone values (aqua, emboss, sketch, neon), extra scene modes
+ (backlight, flowers, AR, HDR), and missing virtual methods for
+ exposure mode, analog gain, lens focus, colour temperature, min &
+ max exposure time. Add new effects and scene modes to Camera
+ parameters.
-- Hardware-accelerated Nvidia video encoder/decoder plugins nvdec and
- nvenc build on Windows now, also with MSVC and using Meson.
+macOS and iOS
-- The ksvideosrc camera capture plugin supports 16-bit grayscale video
- now
+- vtdec can now output to Vulkan-backed memory for zerocopy support
+ with the Vulkan elements.
-- The wasapisrc audio capture element implements loopback recording
- from another output device or sink
+Windows
-- wasapisink recover from low buffer levels in shared mode and some
- exclusive mode fixes
+- d3d11videosink: new Direct3D11-based video sink with support for
+ HDR10 rendering if supported.
+
+- Hardware-accelerated video decoding on Windows via DXVA2 /
+ Direct3D11 using native Windows APIs rather than per-vendor SDKs
+ (like MSDK for Intel or NVCODEC for NVidia). Plus modern Direct3D11
+ integration rather than the almost 20-year old Direct3D9 from
+ Windows XP times used in d3dvideosink. Formats supported for
+ decoding are H.264, H.265, VP8, and VP9, and zero-copy operation
+ should be supported in combination with the new d3d11videosink. See
+ Seungha’s blog post “Windows DXVA2 (via Direct3D 11) Support in
+ GStreamer 1.17” for more details.
+
+- Microsoft Media Foundation plugin for hardware-accelerated video
+ encoding on Windows using native Windows APIs rather than per-vendor
+ SDKs. Formats supported for encoding are H.264, H.265 and VP9. Also
+ includes audio encoders for AAC and MP3. See Seungha’s blog post
+ “Bringing Microsoft Media Foundation to GStreamer” for some more
+ details about this.
+
+- new mfvideosrc video capture source element using the latest Windows
+ APIs rather than ancient APIs used by ksvideosrc/winks. ksvideosrc
+ should be considered deprecated going forward.
+
+- d3d11: add d3d11convert, a color space conversion and rescaling
+ element using shaders, and introduce d3d11upload and d3d11download
+ elements that work just like glupload and gldownload but for D3D11.
+
+- Universal Windows Platform (UWP) support, including official
+ GStreamer binary packages for it. Check out Nirbheek’s latest blog
+ post “GStreamer 1.18 supports the Universal Windows Platform” for
+ more details.
+
+- systemclock correctness and reliability fixes, and also don’t start
+ the system clock at 0 any longer (which shouldn’t make any
+ difference to anyone, as absolute clock time values are supposed to
+ be meaningless in themselves, only the rate of increase matters).
+
+- toolchain specific plugin registry: the registry cache is now named
+ differently for MSVC and MinGW toolchains/packages, which should
+ avoid problems when switching between binaries built with a
+ different toolchain.
+
+- new wasapi2 plugin mainly to support UWP applications. The core
+ logic of this plugin is almost identical to existing wasapi plugin,
+ but the main target is Windows 10 and UWP. This plugin uses WinRT
+ APIs, so will likely not work on Windows 8 or older. Unlike the
+ existing wasapi plugin, this plugin supports automatic stream
+ routing (auto fallback when device was removed) and device level
+ mute/volume control. Exclusive streaming mode is not supported,
+ however, and loopback features are not implemented yet. It is also
+ only possible to build this plugin with MSVC and the Windows 10 SDK,
+ it can’t be cross-compiled with the MingW toolchain.
+
+- new dxgiscreencapsrc element which uses the Desktop Duplication API
+ to capture the desktop screen at high speed. This is only supported
+ on Windows 8 or later. Compared to the existing elements
+ dxgiscreencapsrc offers much better performance, works in High DPI
+ environments and draws an accurate mouse cursor.
+
+- d3dvideosink was downgraded to secondary rank, d3d11videosink is
+ preferred now. Support OverlayComposition for GPU overlay
+ compositing of subtitles and logos.
+
+- debug log output fixes, esp. with a non-UTF8 locale/codepage
+
+- speex, jack: fixed crashes on Windows caused by cross-CRT issues
+
+- gst-play-1.0 interactive keyboard controls now also work on Windows
+
+Linux
+
+- kmssink: Add support for P010 and P016 formats
+
+- vah264dec: new experimental va plugin with an element for H.264
+ decoding with VA-API. This novel approach, different from
+ gstreamer-vaapi, uses the gstcodecs library for decoder state
+ handling, which it is hoped will make for cleaner code because it
+ uses VA-API without further layers or wrappers. Check out Víctor’s
+ blog post “New VA-API H.264 decoder in gst-plugins-bad” for the full
+ lowdown and the limitations of this new plugin, and how to give it a
+ spin.
+
+- v4l2codecs: introduce a V4L2 CODECs Accelerator. This plugin will
+ support the new CODECs uAPI in the Linux kernel, which consists of
+ an accelerator interface similar to DXVA, NVDEC, VDPAU and VAAPI. So
+ far H.264 and VP8 are supported. This is used on certain embedded
+ systems such as i.mx8m, rk3288, rk3399, Allwinner H-series SoCs.
+
+Documentation improvements
+
+- unified documentation containing tutorials, API docs, plugin docs,
+ etc. all under one roof, shipped in form of a documentation release
+ tarball containing both devhelp and html documentation.
+
+- all documentation is now generated using hotdoc, gtk-doc is no
+ longer used. Distributors should use the above-mentioned
+ documentation release tarball instead of trying to package hotdoc
+ and building the documentation from scratch.
+
+- there is now documentation for wrapper plugins like gst-libav and
+ frei0r, as well as tracer plugins.
+
+- for more info, check out Thibault’s “GStreamer Documentation”
+ lightning talk from the 2019 GStreamer Conference.
+
+- new API for plugins to support the documentation system:
+
+ - new GParamSpecFlag GST_PARAM_DOC_SHOW_DEFAULT to make
+ gst-inspect-1.0 (and the documentation) show the paramspec’s
+ default value rather than the actually set value as default
+ - GstPadTemplate getter and setter for “documentation caps”,
+ gst_pad_template_set_documentation_caps() and
+ gst_pad_template_get_documentation_caps(): This can be used in
+ elements where the caps of pad templates are dynamically
+ generated and/or dependent on the environment, to override the
+ caps shown in the documentation (usually to advertise the full
+ set of possible caps).
+ - gst_type_mark_as_plugin_api() for marking types as plugin API,
+ used for plugin-internal types like enums, flags, pad
+ subclasses, boxed types, and such.
+
+Possibly Breaking Changes
+
+- GstVideo: the canonical list of raw video formats (for use in caps)
+ has been reordered, so video elements such as videotestsrc or
+ videoconvert might negotiate to a different format now than before.
+ The new format might be a higher-quality format or require more
+ processing overhead, which might affect pipeline performance.
+
+- mpegtsdemux used to wrongly advertise H.264 and H.265 video
+ elementary streams as alignment=nal. This has now been fixed and
+ changed to alignment=none, which means an h264parse or h265parse
+ element is now required after tsdemux for some pipelines where there
+ wasn’t one before, e.g. in transmuxing scenarios (tsdemux ! tsmux).
+ Pipelines without such a parser may now fail to link or error out at
+ runtime. As parsers after demuxers and before muxers have been
+ generally required for a long time now it is hoped that this will
+ only affect a small number of applications or pipelines.
+
+- The Android opensles audio source and sink used to have hard-coded
+ buffer-/latency-time values of 20ms. This is no longer needed with
+ newer Android versions and has now been removed. This means a higher
+ or lower value might now be negotiated by default, which can affect
+ pipeline performance and latency.
-- dshowsrc now implements the GstDeviceMonitor interface
+Known Issues
+- None in particular
Contributors
-Aaron Boxer, Aleix Conchillo Flaqué, Alessandro Decina, Alexandru Băluț,
-Alex Ashley, Alexey Chernov, Alicia Boya García, Amit Pandya, Andoni
-Morales Alastruey, Andreas Frisch, Andre McCurdy, Andy Green, Anthony
-Violo, Antoine Jacoutot, Antonio Ospite, Arun Raghavan, Aurelien Jarno,
-Aurélien Zanelli, ayaka, Bananahemic, Bastian Köcher, Branko Subasic,
-Brendan Shanks, Carlos Rafael Giani, Charlie Turner, Christoph Reiter,
-Corentin Noël, Daeseok Youn, Damian Vicino, Dan Kegel, Daniel Drake,
-Daniel Klamt, Danilo Spinella, Dardo D Kleiner, David Ing, David
-Svensson Fors, Devarsh Thakkar, Dimitrios Katsaros, Edward Hervey,
-Emilio Pozuelo Monfort, Enrique Ocaña González, Erlend Eriksen, Ezequiel
-Garcia, Fabien Dessenne, Fabrizio Gennari, Florent Thiéry, Francisco
-Velazquez, Freyr666, Garima Gaur, Gary Bisson, George Kiagiadakis, Georg
-Lippitsch, Georg Ottinger, Geunsik Lim, Göran Jönsson, Guillaume
-Desmottes, H1Gdev, Haihao Xiang, Haihua Hu, Harshad Khedkar, Havard
-Graff, He Junyan, Hoonhee Lee, Hosang Lee, Hyunjun Ko, Ilya Smelykh,
-Ingo Randolf, Iñigo Huguet, Jakub Adam, James Stevenson, Jan Alexander
-Steffens, Jan Schmidt, Jerome Laheurte, Jimmy Ohn, Joakim Johansson,
-Jochen Henneberg, Johan Bjäreholt, John-Mark Bell, John Bassett, John
-Nikolaides, Jonathan Karlsson, Jonny Lamb, Jordan Petridis, Josep Torra,
-Joshua M. Doe, Jos van Egmond, Juan Navarro, Julian Bouzas, Jun Xie,
-Junyan He, Justin Kim, Kai Kang, Kim Tae Soo, Kirill Marinushkin, Kyrylo
-Polezhaiev, Lars Petter Endresen, Linus Svensson, Louis-Francis
-Ratté-Boulianne, Lucas Stach, Luis de Bethencourt, Luz Paz, Lyon Wang,
-Maciej Wolny, Marc-André Lureau, Marc Leeman, Marco Trevisan (Treviño),
-Marcos Kintschner, Marian Mihailescu, Marinus Schraal, Mark Nauwelaerts,
-Marouen Ghodhbane, Martin Kelly, Matej Knopp, Mathieu Duponchelle,
-Matteo Valdina, Matthew Waters, Matthias Fend, memeka, Michael Drake,
-Michael Gruner, Michael Olbrich, Michael Tretter, Miguel Paris, Mike
-Wey, Mikhail Fludkov, Naveen Cherukuri, Nicola Murino, Nicolas Dufresne,
-Niels De Graef, Nirbheek Chauhan, Norbert Wesp, Ognyan Tonchev, Olivier
-Crête, Omar Akkila, Pat DeSantis, Patricia Muscalu, Patrick Radizi,
-Patrik Nilsson, Paul Kocialkowski, Per Forlin, Peter Körner, Peter
-Seiderer, Petr Kulhavy, Philippe Normand, Philippe Renon, Philipp Zabel,
-Pierre Labastie, Piotr Drąg, Roland Jon, Roman Sivriver, Roman Shpuntov,
-Rosen Penev, Russel Winder, Sam Gigliotti, Santiago Carot-Nemesio,
-Sean-Der, Sebastian Dröge, Seungha Yang, Shi Yan, Sjoerd Simons, Snir
-Sheriber, Song Bing, Soon, Thean Siew, Sreerenj Balachandran, Stefan
-Ringel, Stephane Cerveau, Stian Selnes, Suhas Nayak, Takeshi Sato,
-Thiago Santos, Thibault Saunier, Thomas Bluemel, Tianhao Liu,
-Tim-Philipp Müller, Tobias Ronge, Tomasz Andrzejak, Tomislav Tustonić,
-U. Artie Eoff, Ulf Olsson, Varunkumar Allagadapa, Víctor Guzmán, Víctor
-Manuel Jáquez Leal, Vincenzo Bono, Vineeth T M, Vivia Nikolaidou, Wang
-Fei, wangzq, Whoopie, Wim Taymans, Wind Yuan, Wonchul Lee, Xabier
-Rodriguez Calvar, Xavier Claessens, Haihao Xiang, Yacine Bandou,
-Yeongjin Jeong, Yuji Kuwabara, Zeeshan Ali,
+Aaron Boxer, Adam Duskett, Adam x Nilsson, Adrian Negreanu, Akinobu
+Mita, Alban Browaeys, Alcaro, Alexander Lapajne, Alexandru Băluț, Alex
+Ashley, Alex Hoenig, Alicia Boya García, Alistair Buxton, Ali Yousuf,
+Ambareesh “Amby” Balaji, Amr Mahdi, Andoni Morales Alastruey, Andreas
+Frisch, Andre Guedes, Andrew Branson, Andrey Sazonov, Antonio Ospite,
+aogun, Arun Raghavan, Askar Safin, AsociTon, A. Wilcox, Axel Mårtensson,
+Ayush Mittal, Bastian Bouchardon, Benjamin Otte, Bilal Elmoussaoui,
+Brady J. Garvin, Branko Subasic, Camilo Celis Guzman, Carlos Rafael
+Giani, Charlie Turner, Cheng-Chang Wu, Chris Ayoup, Chris Lord,
+Christoph Reiter, cketti, Damian Hobson-Garcia, Daniel Klamt, Daniel
+Molkentin, Danny Smith, David Bender, David Gunzinger, David Ing, David
+Svensson Fors, David Trussel, Debarshi Ray, Derek Lesho, Devarsh
+Thakkar, dhilshad, Dimitrios Katsaros, Dmitriy Purgin, Dmitry Shusharin,
+Dominique Leuenberger, Dong Il Park, Doug Nazar, dudengke, Dylan McCall,
+Dylan Yip, Ederson de Souza, Edward Hervey, Eero Nurkkala, Eike Hein,
+ekwange, Eric Marks, Fabian Greffrath, Fabian Orccon, Fabio D’Urso,
+Fabrice Bellet, Fabrice Fontaine, Fanchao L, Felix Yan, Fernando
+Herrrera, Francisco Javier Velázquez-García, Freyr, Fuwei Tang, Gaurav
+Kalra, George Kiagiadakis, Georgii Staroselskii, Georg Lippitsch, Georg
+Ottinger, gla, Göran Jönsson, Gordon Hart, Gregor Boirie, Guillaume
+Desmottes, Guillermo Rodríguez, Haakon Sporsheim, Haihao Xiang, Haihua
+Hu, Havard Graff, Håvard Graff, Heinrich Kruger, He Junyan, Henry
+Wilkes, Hosang Lee, Hou Qi, Hu Qian, Hyunjun Ko, ibauer, Ignacio Casal
+Quinteiro, Ilya Smelykh, Jake Barnes, Jakub Adam, James Cowgill, James
+Westman, Jan Alexander Steffens, Jan Schmidt, Jan Tojnar, Javier Celaya,
+Jeffy Chen, Jennifer Berringer, Jens Göpfert, Jérôme Laheurte, Jim
+Mason, Jimmy Ohn, J. Kim, Joakim Johansson, Jochen Henneberg, Johan
+Bjäreholt, Johan Sternerup, John Bassett, Jonas Holmberg, Jonas Larsson,
+Jonathan Matthew, Jordan Petridis, Jose Antonio Santos Cadenas, Josep
+Torra, Jose Quaresma, Josh Matthews, Joshua M. Doe, Juan Navarro,
+Juergen Werner, Julian Bouzas, Julien Isorce, Jun-ichi OKADA, Justin
+Chadwell, Justin Kim, Keri Henare, Kevin JOLY, Kevin King, Kevin Song,
+Knut Andre Tidemann, Kristofer Björkström, krivoguzovVlad, Kyrylo
+Polezhaiev, Lenny Jorissen, Linus Svensson, Loïc Le Page, Loïc Minier,
+Lucas Stach, Ludvig Rappe, Luka Blaskovic, luke.lin, Luke Yelavich,
+Marcin Kolny, Marc Leeman, Marco Felsch, Marcos Kintschner, Marek
+Olejnik, Mark Nauwelaerts, Markus Ebner, Martin Liska, Martin Theriault,
+Mart Raudsepp, Matej Knopp, Mathieu Duponchelle, Mats Lindestam, Matthew
+Read, Matthew Waters, Matus Gajdos, Maxim Paymushkin, Maxim P.
+Dementiev, Michael Bunk, Michael Gruner, Michael Olbrich, Miguel París
+Díaz, Mikhail Fludkov, Milian Wolff, Millan Castro, Muhammet Ilendemli,
+Nacho García, Nayana Topolsky, Nian Yan, Nicola Murino, Nicolas
+Dufresne, Nicolas Pernas Maradei, Niels De Graef, Nikita Bobkov, Niklas
+Hambüchen, Nirbheek Chauhan, Ognyan Tonchev, okuoku, Oleksandr
+Kvl,Olivier Crête, Ondřej Hruška, Pablo Marcos Oltra, Patricia Muscalu,
+Peter Seiderer, Peter Workman, Philippe Normand, Philippe Renon, Philipp
+Zabel, Pieter Willem Jordaan, Piotr Drąg, Ralf Sippl, Randy Li, Rasmus
+Thomsen, Ratchanan Srirattanamet, Raul Tambre, Ray Tiley, Richard
+Kreckel, Rico Tzschichholz, R Kh, Robert Rosengren, Robert Tiemann,
+Roman Shpuntov, Roman Sivriver, Ruben Gonzalez, Rubén Gonzalez,
+rubenrua, Ryan Huang, Sam Gigliotti, Santiago Carot-Nemesio, Saunier
+Thibault, Scott Kanowitz, Sebastian Dröge, Sebastiano Barrera, Seppo
+Yli-Olli, Sergey Nazaryev, Seungha Yang, Shinya Saito, Silvio
+Lazzeretti, Simon Arnling Bååth, Siwon Kang, sohwan.park, Song Bing,
+Soohyun Lee, Srimanta Panda, Stefano Buora, Stefan Sauer, Stéphane
+Cerveau, Stian Selnes, Sumaid Syed, Swayamjeet, Thiago Santos, Thibault
+Saunier, Thomas Bluemel, Thomas Coldrick, Thor Andreassen, Tim-Philipp
+Müller, Ting-Wei Lan, Tobias Ronge, trilene, Tulio Beloqui, U. Artie
+Eoff, VaL Doroshchuk, Varunkumar Allagadapa, Vedang Patel, Veerabadhran
+G, Víctor Manuel Jáquez Leal, Vivek R, Vivia Nikolaidou, Wangfei, Wang
+Zhanjun, Wim Taymans, Wonchul Lee, Xabier Rodriguez Calvar, Xavier
+Claessens, Xidorn Quan, Xu Guangxin, Yan Wang, Yatin Maan, Yeongjin
+Jeong, yychao, Zebediah Figura, Zeeshan Ali, Zeid Bekli, Zhiyuan Sraf,
+Zoltán Imets,
… and many others who have contributed bug reports, translations, sent
suggestions or helped testing.
+Stable 1.18 branch
-Stable 1.16 branch
-
-After the 1.16.0 release there will be several 1.16.x bug-fix releases
+After the 1.18.0 release there will be several 1.18.x bug-fix releases
which will contain bug fixes which have been deemed suitable for a
stable branch, but no new features or intrusive changes will be added to
-a bug-fix release usually. The 1.16.x bug-fix releases will be made from
-the git 1.16 branch, which is a stable branch.
-
-1.16.0
-
-1.16.0 was released on 19 April 2019.
-
-
-Known Issues
-
-- possibly breaking/incompatible changes to properties of wrapped
- FFmpeg decoders and encoders (see above).
+a bug-fix release usually. The 1.18.x bug-fix releases will be made from
+the git 1.18 branch, which will be a stable branch.
-- The way that GIO modules are named has changed due to upstream GLib
- natively adding support for loading static GIO modules. This means
- that any GStreamer application using gnutls for SSL/TLS on the
- Android or iOS platforms (or any other setup using static libraries)
- will fail to link looking for the g_io_module_gnutls_load_static()
- function. The new function name is now
- g_io_gnutls_load(gpointer data). See Android/iOS sections above for
- further details.
+1.18.0
+1.18.0 was released on 7 September 2020.
-Schedule for 1.18
+Schedule for 1.20
-Our next major feature release will be 1.18, and 1.17 will be the
-unstable development version leading up to the stable 1.18 release. The
-development of 1.17/1.18 will happen in the git master branch.
+Our next major feature release will be 1.20, and 1.19 will be the
+unstable development version leading up to the stable 1.20 release. The
+development of 1.19/1.20 will happen in the git master branch.
-The plan for the 1.18 development cycle is yet to be confirmed, but it
-is possible that the next cycle will be a short one in which case
-feature freeze would be perhaps around August 2019 with a new 1.18
-stable release in September.
+The plan for the 1.20 development cycle is yet to be confirmed, but it
+is now expected that feature freeze will take place some time in January
+2021, with the first 1.20 stable release around February/March 2021.
-1.18 will be backwards-compatible to the stable 1.16, 1.14, 1.12, 1.10,
-1.8, 1.6, 1.4, 1.2 and 1.0 release series.
+1.20 will be backwards-compatible to the stable 1.18, 1.16, 1.14, 1.12,
+1.10, 1.8, 1.6, 1.4, 1.2 and 1.0 release series.
------------------------------------------------------------------------
-_These release notes have been prepared by Tim-Philipp Müller with_
-_contributions from Sebastian Dröge, Guillaume Desmottes, Matthew
-Waters, _ _Thibault Saunier, and Víctor Manuel Jáquez Leal._
+These release notes have been prepared by Tim-Philipp Müller with
+contributions from Mathieu Duponchelle, Matthew Waters, Nirbheek
+Chauhan, Sebastian Dröge, Thibault Saunier, and Víctor Manuel Jáquez
+Leal.
-_License: CC BY-SA 4.0_
+License: CC BY-SA 4.0
projects / platform / upstream / gstreamer.git / blobdiff