3 GSTREAMER 1.16 RELEASE NOTES
6 GStreamer 1.16.0 was originally released on 19 April 2019.
8 The latest bug-fix release in the 1.16 series is 1.16.1 and was released
11 See https://gstreamer.freedesktop.org/releases/1.16/ for the latest
12 version of this document.
14 _Last updated: Sunday 22 September 2019, 21:00 UTC (log)_
19 The GStreamer team is proud to announce a new major feature release in
20 the stable 1.x API series of your favourite cross-platform multimedia
23 As always, this release is again packed with many new features, bug
24 fixes and other improvements.
29 - GStreamer WebRTC stack gained support for data channels for
30 peer-to-peer communication based on SCTP, BUNDLE support, as well as
31 support for multiple TURN servers.
33 - AV1 video codec support for Matroska and QuickTime/MP4 containers
34 and more configuration options and supported input formats for the
37 - Support for Closed Captions and other Ancillary Data in video
39 - Support for planar (non-interleaved) raw audio
41 - GstVideoAggregator, compositor and OpenGL mixer elements are now in
44 - New alternate fields interlace mode where each buffer carries a
47 - WebM and Matroska ContentEncryption support in the Matroska demuxer
49 - new WebKit WPE-based web browser source element
51 - Video4Linux: HEVC encoding and decoding, JPEG encoding, and improved
54 - Hardware-accelerated Nvidia video decoder gained support for VP8/VP9
55 decoding, whilst the encoder gained support for H.265/HEVC encoding.
57 - Many improvements to the Intel Media SDK based hardware-accelerated
58 video decoder and encoder plugin (msdk): dmabuf import/export for
59 zero-copy integration with other components; VP9 decoding; 10-bit
60 HEVC encoding; video post-processing (vpp) support including
61 deinterlacing; and the video decoder now handles dynamic resolution
64 - The ASS/SSA subtitle overlay renderer can now handle multiple
65 subtitles that overlap in time and will show them on screen
68 - The Meson build is now feature-complete (*) and it is now the
69 recommended build system on all platforms. The Autotools build is
70 scheduled to be removed in the next cycle.
72 - The GStreamer Rust bindings and Rust plugins module are now
73 officially part of upstream GStreamer.
75 - The GStreamer Editing Services gained a gesdemux element that allows
76 directly playing back serialized edit list with playbin or
79 - Many performance improvements
82 Major new features and changes
86 - GstAggregator has a new "min-upstream-latency" property that forces
87 a minimum aggregate latency for the input branches of an aggregator.
88 This is useful for dynamic pipelines where branches with a higher
89 latency might be added later after the pipeline is already up and
90 running and where a change in the latency would be disruptive. This
91 only applies to the case where at least one of the input branches is
92 live though, it won’t force the aggregator into live mode in the
93 absence of any live inputs.
95 - GstBaseSink gained a "processing-deadline" property and
96 setter/getter API to configure a processing deadline for live
97 pipelines. The processing deadline is the acceptable amount of time
98 to process the media in a live pipeline before it reaches the sink.
99 This is on top of the systemic latency that is normally reported by
100 the latency query. This defaults to 20ms and should make pipelines
101 such as v4l2src ! xvimagesink not claim that all frames are late in
102 the QoS events. Ideally, this should replace the "max-lateness"
103 property for most applications.
105 - RTCP Extended Reports (XR) parsing according to RFC 3611:
106 Loss/Duplicate RLE, Packet Receipt Times, Receiver Reference Time,
107 Delay since the last Receiver (DLRR), Statistics Summary, and VoIP
108 Metrics reports. This only provides the ability to parse such
109 packets, generation of XR packets is not supported yet and XR
110 packets are not automatically parsed by rtpbin / rtpsession but must
111 be actively handled by the application.
113 - a new mode for interlaced video was added where each buffer carries
114 a single field of interlaced video, with buffer flags indicating
115 whether the field is the top field or bottom field. Top and bottom
116 fields are expected to alternate in this mode. Caps for this
117 interlace mode must also carry a format:Interlaced caps feature to
118 ensure backwards compatibility.
120 - The video library has gained support for three new raw pixel
123 - Y410: packed 4:4:4 YUV, 10 bits per channel
124 - Y210: packed 4:2:2 YUV, 10 bits per channel
125 - NV12_10LE40: fully-packed 10-bit variant of NV12_10LE32,
126 i.e. without the padding bits
128 - GstRTPSourceMeta is a new meta that can be used to transport
129 information about the origin of depayloaded or decoded RTP buffers,
130 e.g. when mixing audio from multiple sources into a single stream. A
131 new "source-info" property on the RTP depayloader base class
132 determines whether depayloaders should put this meta on outgoing
133 buffers. Similarly, the same property on RTP payloaders determines
134 whether they should use the information from this meta to construct
135 the CSRCs list on outgoing RTP buffers.
137 - gst_sdp_message_from_text() is a convenience constructor to parse
138 SDPs from a string which is particularly useful for language
141 Support for Planar (Non-Interleaved) Raw Audio
143 Raw audio samples are usually passed around in interleaved form in
144 GStreamer, which means that if there are multiple audio channels the
145 samples for each channel are interleaved in memory, e.g.
146 |LEFT|RIGHT|LEFT|RIGHT|LEFT|RIGHT| for stereo audio. A non-interleaved
147 or planar arrangement in memory would look like
148 |LEFT|LEFT|LEFT|RIGHT|RIGHT|RIGHT| instead, possibly with
149 |LEFT|LEFT|LEFT| and |RIGHT|RIGHT|RIGHT| residing in separate memory
150 chunks or separated by some padding.
152 GStreamer has always had signalling for non-interleaved audio since
153 version 1.0, but it was never actually properly implemented in any
154 elements. audioconvert would advertise support for it, but wasn’t
155 actually able to handle it correctly.
157 With this release we now have full support for non-interleaved audio as
158 well, which means more efficient integration with external APIs that
159 handle audio this way, but also more efficient processing of certain
160 operations like interleaving multiple 1-channel streams into a
161 multi-channel stream which can be done without memory copies now.
163 New API to support this has been added to the GStreamer Audio support
164 library: There is now a new GstAudioMeta which describes how data is
165 laid out inside the buffer, and buffers with non-interleaved audio must
166 always carry this meta. To access the non-interleaved audio samples you
167 must map such buffers with gst_audio_buffer_map() which works much like
168 gst_buffer_map() or gst_video_frame_map() in that it will populate a
169 little GstAudioBuffer helper structure passed to it with the number of
170 samples, the number of planes and pointers to the start of each plane in
171 memory. This function can also be used to map interleaved audio buffers
172 in which case there will be only one plane of interleaved samples.
174 Of course support for this has also been implemented in the various
175 audio helper and conversion APIs, base classes, and in elements such as
176 audioconvert, audioresample, audiotestsrc, audiorate.
178 Support for Closed Captions and Other Ancillary Data in Video
180 The video support library has gained support for detecting and
181 extracting Ancillary Data from videos as per the SMPTE S291M
182 specification, including:
184 - a VBI (Vertical Blanking Interval) parser that can detect and
185 extract Ancillary Data from Vertical Blanking Interval lines of
186 component signals. This is currently supported for videos in v210
189 - a new GstMeta for closed captions: GstVideoCaptionMeta. This
190 supports the two types of closed captions, CEA-608 and CEA-708,
191 along with the four different ways they can be transported (other
192 systems are a superset of those).
194 - a VBI (Vertical Blanking Interval) encoder for writing ancillary
195 data to the Vertical Blanking Interval lines of component signals.
197 The new closedcaption plugin in gst-plugins-bad then makes use of all
198 this new infrastructure and provides the following elements:
200 - cccombiner: a closed caption combiner that takes a closed captions
201 stream and another stream and adds the closed captions as
202 GstVideoCaptionMeta to the buffers of the other stream.
204 - ccextractor: a closed caption extractor which will take
205 GstVideoCaptionMeta from input buffers and output them as a separate
206 closed captions stream.
208 - ccconverter: a closed caption converter that can convert between
211 - line21encoder, line21decoder: inject/extract line21 closed captions
212 to/from SD video streams
214 - cc708overlay: decodes CEA 608/708 captions and overlays them on
217 Additionally, the following elements have also gained Closed Caption
220 - qtdemux and qtmux support CEA 608/708 Closed Caption tracks
222 - mpegvideoparse, h264parse extracts Closed Captions from MPEG-2/H.264
225 - avviddec, avvidenc, x264enc got support for extracting/injecting
228 - decklinkvideosink can output closed captions and decklinkvideosrc
229 can extract closed captions
231 - playbin and playbin3 learned how to autoplug CEA 608/708 CC overlay
234 - the externally maintained ajavideosrc element for AJA capture cards
235 has support for extracting closed captions
237 The rsclosedcaption plugin in the Rust plugins collection includes a
238 MacCaption (MCC) file parser and encoder.
242 - overlaycomposition: New element that allows applications to draw
243 GstVideoOverlayCompositions on a stream. The element will emit the
244 "draw" signal for each video buffer, and the application then
245 generates an overlay for that frame (or not). This is much more
246 performant than e.g. cairooverlay for many use cases, e.g. because
247 pixel format conversions can be avoided or the blitting of the
248 overlay can be delegated to downstream elements (such as
249 gloverlaycompositor). It’s particularly useful for cases where only
250 a small section of the video frame should be drawn on.
252 - gloverlaycompositor: New OpenGL-based compositor element that
253 flattens any overlays from GstVideoOverlayCompositionMetas into the
254 video stream. This element is also always part of glimagesink.
256 - glalpha: New element that adds an alpha channel to a video stream.
257 The values of the alpha channel can either be set to a constant or
258 can be dynamically calculated via chroma keying. It is similar to
259 the existing alpha element but based on OpenGL. Calculations are
260 done in floating point so results may not be identical to the output
261 of the existing alpha element.
263 - rtpfunnel funnels together RTP streams into a single session. Use
264 cases include multiplexing and bundle. webrtcbin uses it to
265 implement BUNDLE support.
267 - testsrcbin is a source element that provides an audio and/or video
268 stream and also announces them using the recently-introduced
269 GstStream API. This is useful for testing elements such as playbin3
270 or uridecodebin3 etc.
272 - New closed caption elements: cccombiner, ccextractor, ccconverter,
273 line21encoder, line21decoder and cc708overlay (see above)
275 - wpesrc: new source element acting as a Web Browser based on WebKit
278 - Two new OpenCV-based elements: cameracalibrate and cameraundistort
279 that can communicate to figure out distortion correction parameters
280 for a camera and correct for the distortion.
282 - New sctp plugin based on usrsctp with sctpenc and sctpdec elements.
283 These elements are used inside webrtcbin for implementing data
286 New element features and additions
288 - playbin3, playbin and playsink have gained a new "text-offset"
289 property to adjust the positioning of the selected subtitle stream
290 vis-a-vis the audio and video streams. This uses subtitleoverlay’s
291 new "subtitle-ts-offset" property. GstPlayer has gained matching API
292 for this, namely gst_player_get_text_video_offset().
294 - playbin3 buffering improvements: in network playback scenarios there
295 may be multiple inputs to decodebin3, and buffering will be done
296 before decodebin3 using queue2 or downloadbuffer elements inside
297 urisourcebin. Since this is before any parsers or demuxers there may
298 not be any bitrate information available for the various streams, so
299 it was difficult to configure the buffering there smartly within
300 global constraints. This was improved now: The queue2 elements
301 inside urisourcebin will now use the new bitrate query to figure out
302 a bitrate estimate for the stream if no bitrate was provided by
303 upstream, and urisourcebin will use the bitrates of the individual
304 queues to distribute the globally-set "buffer-size" budget in bytes
305 to the various queues. urisourcebin also gained "low-watermark" and
306 "high-watermark" properties which will be proxied to the internal
307 queues, as well as a read-only "statistics" property which allows
308 querying of the minimum/maximum/average byte and time levels of the
309 queues inside the urisourcebin in question.
311 - splitmuxsink has gained a couple of new features:
313 - new "async-finalize" mode: This mode is useful for muxers or
314 outputs that can take a long time to finalize a file. Instead of
315 blocking the whole upstream pipeline while the muxer is doing
316 its stuff, we can unlink it and spawn a new muxer + sink
317 combination to continue running normally. This requires us to
318 receive the muxer and sink (if needed) as factories via the new
319 "muxer-factory" and "sink-factory" properties, optionally
320 accompanied by their respective properties structures (set via
321 the new "muxer-properties" and "sink-properties" properties).
322 There are also new "muxer-added" and "sink-added" signals in
323 case custom code has to be called for them to configure them.
325 - "split-at-running-time" action signal: When called by the user,
326 this action signal ends the current file (and starts a new one)
327 as soon as the given running time is reached. If called multiple
328 times, running times are queued up and processed in the order
331 - "split-after" action signal to finish outputting the current GOP
332 to the current file and then start a new file as soon as the GOP
333 is finished and a new GOP is opened (unlike the existing
334 "split-now" which immediately finishes the current file and
335 writes the current GOP into the next newly-started file).
337 - "reset-muxer" property: when unset, the muxer is reset using
338 flush events instead of setting its state to NULL and back. This
339 means the muxer can keep state across resets, e.g. mpegtsmux
340 will keep the continuity counter continuous across segments as
341 required by hlssink2.
343 - qtdemux gained PIFF track encryption box support in addition to the
344 already-existing PIFF sample encryption support, and also allows
345 applications to select which encryption system to use via a
346 "drm-preferred-decryption-system-id" context in case there are
349 - qtmux: the "start-gap-threshold" property determines now whether an
350 edit list will be created to account for small gaps or offsets at
351 the beginning of a stream in case the start timestamps of tracks
352 don’t line up perfectly. Previously the threshold was hard-coded to
353 1% of the (video) frame duration, now it is 0 by default (so edit
354 list will be created even for small differences), but fully
357 - rtpjitterbuffer has improved end-of-stream handling
359 - rtpmp4vpay will be preferred over rtpmp4gpay for MPEG-4 video in
360 autoplugging scenarios now
362 - rtspsrc now allows applications to send RTSP SET_PARAMETER and
363 GET_PARAMETER requests using action signals.
365 - rtspsrc has a small (100ms) configurable teardown delay by default
366 to try and make sure an RTSP TEARDOWN request gets sent out when the
367 source element shuts down. This will block the downward PAUSED to
368 READY state change for a short time, but can be disabled where it’s
369 a problem. Some servers only allow a limited number of concurrent
370 clients, so if no proper TEARDOWN is sent new clients may have
371 problems connecting to the server for a while.
373 - souphttpsrc behaves better with low bitrate streams now. Before it
374 would increase the read block size too quickly which could lead to
375 it not reading any data from the socket for a very long time with
376 low bitrate streams that are output live downstream. This could lead
377 to servers kicking off the client.
379 - filesink: do internal buffering to avoid performance regression with
380 small writes since we bypass libc buffering by using writev()
383 - identity: add "eos-after" property and fix "error-after" property
384 when the element is reused
386 - input-selector: lets context queries pass through, so that
387 e.g. upstream OpenGL elements can use contexts and displays
388 advertised by downstream elements
390 - queue2: avoid ping-pong between 0% and 100% buffering messages if
391 upstream is pushing buffers larger than one of its limits, plus
392 performance optimisations
394 - opusdec: new "phase-inversion" property to control phase inversion.
395 When enabled, this will slightly increase stereo quality, but
396 produces a stream that when downmixed to mono will suffer audio
399 - The x265enc HEVC encoder also exposes a "key-int-max" property to
400 configure the maximum allowed GOP size now.
402 - decklinkvideosink has seen stability improvements for long-running
403 pipelines (potential crash due to overflow of leaked clock refcount)
404 and clock-slaving improvements when performing flushing seeks
405 (causing stalls in the output timeline), pausing and/or buffering.
407 - srtpdec, srtpenc: add support for MKIs which allow multiple keys to
408 be used with a single SRTP stream
410 - srtpdec, srtpenc: add support for AES-GCM and also add support for
411 it in gst-rtsp-server and rtspsrc.
413 - The srt Secure Reliable Transport plugin has integrated server and
414 client elements srt{client,server}{src,sink} into one (srtsrc and
415 srtsink), since SRT connection mode can be changed by uri
418 - h264parse and h265parse will handle SEI recovery point messages and
419 mark recovery points as keyframes as well (in addition to IDR
422 - webrtcbin: "add-turn-server" action signal to pass multiple ICE
423 relays (TURN servers).
425 - The removesilence element has received various new features and
426 properties, such as a "threshold" property, detecting silence only
427 after minimum silence time/buffers, a "silent" property to control
428 bus message notifications as well as a "squash" property.
430 - AOMedia AV1 decoder gained support for 10/12bit decoding whilst the
431 AV1 encoder supports more image formats and subsamplings now and
432 acquired support for rate control and profile related configuration.
434 - The Fraunhofer fdkaac plugin can now be built against the 2.0.0
435 version API and has improved multichannel support
437 - kmssink now supports unpadded 24-bit RGB and can configure mode
438 setting from video info, which enables display of multi-planar
439 formats such as I420 or NV12 with modesetting. It has also gained a
440 number of new properties: The "restore-crtc" property does what it
441 says on the tin and is enabled by default. "plane-properties" and
442 "connector-properties" can be used to pass custom properties to the
445 - waylandsink has a "fullscreen" property now and supports the
448 - decklinkvideosink, decklinkvideosrc support selecting between
451 - The vulkan plugin gained support for macOS and iOS via MoltenVK in
452 addition to the existing support for X11 and Wayland
454 - imagefreeze has a new num-buffers property to limit the number of
455 buffers that are produced and to send an EOS event afterwards
457 - webrtcbin has a new, introspectable get-transceiver signal in
458 addition to the old get-transceivers signal that couldn’t be used
461 - Support for per-element latency information was added to the latency
464 Plugin and library moves
466 - The stereo element was moved from -bad into the existing audiofx
467 plugin in -good. If you get duplicate type registration warnings
468 when upgrading, check that you don’t have a stale stereoplugin lying
471 GstVideoAggregator, compositor, and OpenGL mixer elements moved from -bad to -base
473 GstVideoAggregator is a new base class for raw video mixers and muxers
474 and is based on GstAggregator. It provides defined-latency mixing of raw
475 video inputs and ensures that the pipeline won’t stall even if one of
476 the input streams stops producing data.
478 As part of the move to stabilise the API there were some last-minute API
479 changes and clean-ups, but those should mostly affect internal elements.
480 Most notably, the "ignore-eos" pad property was renamed to
481 "repeat-after-eos" and the conversion code was moved to a
482 GstVideoAggregatorConvertPad subclass to avoid code duplication, make
483 things less awkward for subclasses like the OpenGL-based video mixer,
484 and make the API more consistent with the audio aggregator API.
486 It is used by the compositor element, which is a replacement for
487 ‘videomixer’ which did not handle live inputs very well. compositor
488 should behave much better in that respect and generally behave as one
489 would expected in most scenarios.
491 The compositor element has gained support for per-pad blending mode
492 operators (SOURCE, OVER, ADD) which determines what operator to use for
493 blending this pad over the previous ones. This can be used to implement
494 crossfading and the available operators can be extended in the future as
497 A number of OpenGL-based video mixer elements (glvideomixer, glmixerbin,
498 glvideomixerelement, glstereomix, glmosaic) which are built on top of
499 GstVideoAggregator have also been moved from -bad to -base now. These
500 elements have been merged into the existing OpenGL plugin, so if you get
501 duplicate type registration warnings when upgrading, check that you
502 don’t have a stale openglmixers plugin lying about somewhere.
506 The following plugins have been removed from gst-plugins-bad:
508 - The experimental daala plugin has been removed, since it’s not so
509 useful now that all effort is focused on AV1 instead, and it had to
510 be enabled explicitly with --enable-experimental anyway.
512 - The spc plugin has been removed. It has been replaced by the gme
515 - The acmmp3dec and acmenc plugins for Windows have been removed. ACM
516 is an ancient legacy API and there was no point in keeping the
517 plugins around for a licensed MP3 decoder now that the MP3 patents
518 have expired and we have a decoder in -good. We also didn’t ship
519 these in our cerbero-built Windows packages, so it’s unlikely that
523 Miscellaneous API additions
525 - GstBitwriter: new generic bit writer API to complement the existing
528 - gst_buffer_new_wrapped_bytes() creates a wrap buffer from a GBytes
530 - gst_caps_set_features_simple() sets a caps feature on all the
531 structures of a GstCaps
533 - New GST_QUERY_BITRATE query: This allows determining from downstream
534 what the expected bitrate of a stream may be which is useful in
535 queue2 for setting time based limits when upstream does not provide
536 timing information. tsdemux, qtdemux and matroskademux have basic
537 support for this query on their sink pads.
539 - elements: there is a new “Hardware” class specifier. Elements
540 interacting with hardware devices should specify this classifier in
541 their element factory class metadata. This is useful to advertise as
542 one might need to put such elements into READY state to test if the
543 hardware is present in the system for example.
545 - protection: Add a new definition for unspecified system protection,
546 GST_PROTECTION_UNSPECIFIED_SYSTEM_ID
548 - take functions for various mini objects that didn’t have them yet:
549 gst_query_take(), gst_message_take(), gst_tag_list_take(),
550 gst_buffer_list_take(). Unlike the various _replace() functions
551 _take() does not increase the reference count but takes ownership of
552 the mini object passed.
554 - clear functions for various mini object types and GstObject which
555 unrefs the object or mini object (if non-NULL) and sets the variable
556 pointed to to NULL: gst_clear_structure(), gst_clear_tag_list(),
557 gst_clear_query(), gst_clear_message(), gst_clear_event(),
558 gst_clear_caps(), gst_clear_buffer_list(), gst_clear_buffer(),
559 gst_clear_mini_object(), gst_clear_object()
561 - miniobject: new API gst_mini_object_add_parent() and
562 gst_mini_object_remove_parent() to set parent pointers on mini
563 objects to ensure correct writability: Every container of
564 miniobjects now needs to store itself as parent in the child object,
565 and remove itself again later. A mini object is then only writable
566 if there is at most one parent, that parent is writable itself, and
567 the reference count of the mini object is 1. GstBuffer (for
568 memories), GstBufferList (for buffers), GstSample (for caps, buffer,
569 bufferlist), and GstVideoOverlayComposition were updated
570 accordingly. Without this it was possible to have e.g. a buffer list
571 with a refcount of 2 used in two places at once that both modify the
572 same buffer with refcount 1 at the same time wrongly thinking it is
573 writable even though it’s really not.
575 - poll: add API to watch for POLLPRI and stop treating POLLPRI as a
576 read. This is useful to wait for video4linux events which are
577 signalled via POLLPRI.
579 - sample: new API to update the contents of a GstSample and make it
580 writable: gst_sample_set_buffer(), gst_sample_set_caps(),
581 gst_sample_set_segment(), gst_sample_set_info(), plus
582 gst_sample_is_writable() and gst_sample_make_writable(). This makes
583 it possible to reuse a sample object and avoid unnecessary memory
584 allocations, for example in appsink.
586 - ClockIDs now keep a weak reference to underlying clock to avoid
587 crashes in basesink in corner cases where a clock goes away while
588 the ClockID is still in use, plus some new API
589 (gst_clock_id_get_clock(), gst_clock_id_uses_clock()) to check the
590 clock a ClockID is linked to.
592 - The GstCheck unit test library gained a
593 fail_unless_equals_clocktime() convenience macro as well as some new
594 GstHarness API for for proposing meta APIs from the allocation
595 query: gst_harness_add_propose_allocation_meta(). ASSERT_CRITICAL()
596 checks in unit tests are now skipped if GStreamer was compiled with
597 GST_DISABLE_GLIB_CHECKS.
599 - gst_audio_buffer_truncate() convenience function to truncate a raw
602 - GstDiscoverer has support for caching the results of discovery in
603 the default cache directory. This can be enabled with the use-cache
604 property and is disabled by default.
606 - GstMeta that are attached to GstBuffers are now always stored in the
607 order in which they were added.
609 - Additional support for signalling ONVIF specific features were
610 added: the SEEK event can store a trickmode-interval now and support
611 for the Rate-Control and Frames RTSP headers was added to the RTSP
615 Miscellaneous performance and memory optimisations
617 As always there have been many performance and memory usage improvements
618 across all components and modules. Some of them (such as dmabuf
619 import/export) have already been mentioned elsewhere so won’t be
622 The following list is only a small snapshot of some of the more
623 interesting optimisations that haven’t been mentioned in other contexts
626 - The GstVideoEncoder and GstVideoDecoder base classes now release the
627 STREAM_LOCK when pushing out buffers, which means (multi-threaded)
628 encoders and decoders can now receive and continue to process input
629 buffers whilst waiting for downstream elements in the pipeline to
630 process the buffer that was pushed out. This increases throughput
631 and reduces processing latency, also and especially for
632 hardware-accelerated encoder/decoder elements.
634 - GstQueueArray has seen a few API additions
635 (gst_queue_array_peek_nth(), gst_queue_array_set_clear_func(),
636 gst_queue_array_clear()) so that it can be used in other places like
637 GstAdapter instead of a GList, which reduces allocations and
638 improves performance.
640 - appsink now reuses the sample object in pull_sample() if possible
642 - rtpsession only starts the RTCP thread when it’s actually needed now
644 - udpsrc uses a buffer pool now and the GstUdpSrc object structure was
645 optimised for better cache performance
649 - API was added to fine-tune the synchronisation offset between
653 Miscellaneous changes
655 - As a result of moving to newer FFmpeg APIs, encoder and decoder
656 elements exposed by the GStreamer FFmpeg wrapper plugin (gst-libav)
657 may have seen possibly incompatible changes to property names and/or
658 types, and not all properties exposed might be functional. We are
659 still reviewing the new properties and aim to minimise breaking
660 changes at least for the most commonly-used properties, so please
661 report any issues you run into!
665 - The OpenGL mixer elements have been moved from -bad to
666 gst-plugins-base (see above)
668 - The Mesa GBM backend now supports headless mode
670 - gloverlaycompositor: New OpenGL-based compositor element that
671 flattens any overlays from GstVideoOverlayCompositionMetas into the
674 - glalpha: New element that adds an alpha channel to a video stream.
675 The values of the alpha channel can either be set to a constant or
676 can be dynamically calculated via chroma keying. It is similar to
677 the existing alpha element but based on OpenGL. Calculations are
678 done in floating point so results may not be identical to the output
679 of the existing alpha element.
681 - glupload: Implement direct dmabuf uploader, the idea being that some
682 GPUs (like the Vivante series) can actually perform the YUV->RGB
683 conversion internally, so no custom conversion shaders are needed.
684 To make use of this feature, we need an additional uploader that can
685 import DMABUF FDs and also directly pass the pixel format, relying
686 on the GPU to do the conversion.
688 - The OpenGL library no longer restores the OpenGL viewport. This is a
689 performance optimization to not require performing multiple
690 expensive glGet*() function calls per frame. This affects any
691 application or plugin use of the following functions and objects:
692 - glcolorconvert library object (not the element)
693 - glviewconvert library object (not the element)
694 - gst_gl_framebuffer_draw_to_texture()
695 - custom GstGLWindow implementations
698 Tracing framework and debugging improvements
700 - There is now a GDB PRETTY PRINTER FOR VARIOUS GSTREAMER TYPES: For
701 GstObject pointers the type and name is added, e.g.
702 0x5555557e4110 [GstDecodeBin|decodebin0]. For GstMiniObject pointers
703 the object type is added, e.g. 0x7fffe001fc50 [GstBuffer]. For
704 GstClockTime and GstClockTimeDiff the time is also printed in human
705 readable form, e.g. 150116219955 [+0:02:30.116219955].
707 - GDB EXTENSION WITH TWO CUSTOM GDB COMMANDS gst-dot AND gst-print:
709 - gst-dot creates dot files that a very close to what
710 GST_DEBUG_BIN_TO_DOT_FILE() produces, but object properties and
711 buffer contents such as codec-data in caps are not available.
713 - gst-print produces high-level information about a GStreamer
714 object. This is currently limited to pads for GstElements and
715 events for the pads. The output may look like this:
717 - gst_structure_to_string() now serialises the actual value of
718 pointers when serialising GstStructures instead of claiming they’re
719 NULL. This makes debug logging in various places less confusing,
720 because it’s clear now that structure fields actually hold valid
721 objects. Such object pointer values will never be deserialised
727 - gst-inspect-1.0 has coloured output now and will automatically use a
728 pager if the output does not fit on a page. This only works in a
729 UNIX environment and if the output is not piped, and on Windows 10
730 build 16257 or newer. If you don’t like the colours you can disable
731 them by setting the GST_INSPECT_NO_COLORS=1 environment variable or
732 passing the --no-color command line option.
735 GStreamer RTSP server
737 - Improved backlog handling when using TCP interleaved for data
738 transport. Before there was a fixed maximum size for backlog
739 messages, which was prone to deadlocks and made it difficult to
740 control memory usage with the watch backlog. The RTSP server now
741 limits queued TCP data messages to one per stream, moving queuing of
742 the data into the pipeline and leaving the RTSP connection
743 responsive to RTSP messages in both directions, preventing all those
746 - Initial ULP Forward Error Correction support in rtspclientsink and
747 for RECORD mode in the server.
749 - API to explicitly enable retransmission requests (RTX)
751 - Lots of multicast-related fixes
753 - rtsp-auth: Add support for parsing .htdigest files
758 - Support Wayland’s display for context sharing, so the application
759 can pass its own wl_display in order to be used for the VAAPI
762 - A lot of work to support new Intel hardware using media-driver as VA
765 - For non-x86 devices, VAAPI display can instantiate, through DRM,
766 with no PCI bus. This enables the usage of libva-v4l2-request
769 - Added support for XDG-shell protocol as wl_shell replacement which
770 is currently deprecated. This change add as dependency
773 - GstVaapiFilter, GstVaapiWindow, and GstVaapiDecoder classes now
774 inherit from GstObject, gaining all the GStreamer’s instrumentation
777 - The metadata now specifies the plugin as Hardware class.
779 - H264 decoder is more stable with problematic streams.
781 - In H265 decoder added support for profiles main-422-10 (P010_10LE),
782 main-444 (AYUV) and main-444-10 (Y410)
784 - JPEG decoder handles dynamic resolution changes.
786 - More specification adherence in H264 and H265 encoders.
791 - Add support of NV16 format to video encoders input.
793 - Video decoders now handle the ALLOCATION query to tell upstream
794 about the number of buffers they require. Video encoders will also
795 use this query to adjust their number of allocated buffers
796 preventing starvation when using dynamic buffer mode.
798 - The OMX_PERFORMANCE debug category has been renamed to OMX_API_TRACE
799 and can now be used to track a widder variety of interactions
800 between OMX and GStreamer.
802 - Video encoders will now detect frame rate only changes and will
803 inform OMX about it rather than doing a full format reset.
805 - Various Zynq UltraScale+ specific improvements:
806 - Video encoders are now able to import dmabuf from upstream.
807 - Support for HEVC range extension profiles and more AVC profiles.
808 - We can now request video encoders to generate an IDR using the
809 force key unit event.
812 GStreamer Editing Services and NLE
814 - Added a gesdemux element, it is an auto pluggable element that
815 allows decoding edit list like files supported by GES
817 - Added gessrc which wraps a GESTimeline as a standard source element
818 (implementing the ges protocol handler)
820 - Added basic support for videorate::rate property potentially
821 allowing changing playback speed
823 - Layer priority is now fully automatic and they should be moved with
824 the new ges_timeline_move_layer method, ges_layer_set_priority is
827 - Added a ges_timeline_element_get_layer_priority so we can simply get
828 all information about GESTimelineElement position in the timeline
830 - GESVideoSource now auto orientates the images if it is defined in a
833 - Added some PyGObject overrides to make the API more pythonic
835 - The threading model has been made more explicit with safe guard to
836 make sure not thread safe APIs are not used from the wrong threads.
837 It is also now possible to properly handle in what thread the API
840 - Optimized GESClip and GESTrackElement creation
842 - Added a way to compile out the old, unused and deprecated
845 - Re implemented the timeline editing API making it faster and making
846 the code much more maintainable
848 - Simplified usage of nlecomposition outside GES by removing quirks in
849 it API usage and removing the need to treat it specially from an
850 application perspective.
854 - Added support to add titles to the timeline
855 - Enhance the help auto generating it from the code
857 - Deprecate ges_timeline_load_from_uri as loading the timeline should
858 be done through a project now
860 - MANY leaks have been plugged and the unit testsuite is now “leak
866 - Added an action type to verify the checksum of the sink last-sample
868 - Added an include keyword to validate scenarios
870 - Added the notion of variable in scenarios, with the set-vars keyword
872 - Started adding support for “performance” like tests by allowing to
873 define the number of dropped buffers or the minimum buffer frequency
876 - Added a validateflow plugin which allows defining the data flow to
877 be seen on a particular pad and verifying that following runs match
880 - Added support for appsrc based test definition so we can instrument
881 the data pushed into the pipeline from scenarios
883 - Added a mockdecryptor allowing adding tests with on encrypted files,
884 the element will potentially be instrumented with a validate
887 - gst-validate-launcher:
891 - Changed the default for “muting” tests as user doesn’t expect
892 hundreds of windows to show up when running the testsuite
894 - Fixed the outputted xunit files to be compatible with GitLab
896 - Added support to run tests on media files in push mode (using
899 - Added support for running inside gst-build
901 - Added support for running ssim tests on rendered files
903 - Added a way to simply define tests on pipelines through a simple
906 - Added a python app to easily run python testsuite reusing all
907 the launcher features
909 - Added flatpak knowledge so we can print backtrace even when
910 running from within flatpak
912 - Added a way to automatically generated “known issues”
915 - Added a way to rerun tests to check if they are flaky and added
916 a way to tolerate tests known to be flaky
918 - Add a way to output html log files
921 GStreamer Python Bindings
923 - add binding for gst_pad_set_caps()
925 - pygobject dependency requirement was bumped to >= 3.8
927 - new audiotestsrc, audioplot, and mixer plugin examples, and a
928 dynamic pipeline example
931 GStreamer C# Bindings
933 - bindings for the GstWebRTC library
936 GStreamer Rust Bindings
938 The GStreamer Rust bindings are now officially part of the GStreamer
939 project and are also maintained in the GStreamer GitLab.
941 The releases will generally not be synchronized with the releases of
942 other GStreamer parts due to dependencies on other projects.
944 Also unlike the other GStreamer libraries, the bindings will not commit
945 to full API stability but instead will follow the approach that is
946 generally taken by Rust projects, e.g.:
948 1) 0.12.X will be completely API compatible with all other 0.12.Y
950 2) 0.12.X+1 will contain bugfixes and compatible new feature additions.
951 3) 0.13.0 will _not_ be backwards compatible with 0.12.X but projects
952 will be able to stay at 0.12.X without any problems as long as they
953 don’t need newer features.
955 The current stable release is 0.12.2 and the next release series will be
956 0.13, probably around March 2019.
958 At this point the bindings cover most of GStreamer core (except for most
959 notably GstAllocator and GstMemory), and most parts of the app, audio,
960 base, check, editing-services, gl, net. pbutils, player, rtsp,
961 rtsp-server, sdp, video and webrtc libraries.
963 Also included is support for creating subclasses of the following types
964 and writing GStreamer plugins:
967 - gst::Bin and gst::Pipeline
968 - gst::URIHandler and gst::ChildProxy
969 - gst::Pad, gst::GhostPad
970 - gst_base::Aggregator and gst_base::AggregatorPad
971 - gst_base::BaseSrc and gst_base::BaseSink
972 - gst_base::BaseTransform
974 Changes to 0.12.X since 0.12.0
978 - PTP clock constructor actually creates a PTP instead of NTP clock
982 - Bindings for GStreamer Editing Services
983 - Bindings for GStreamer Check testing library
984 - Bindings for the encoding profile API (encodebin)
986 - VideoFrame, VideoInfo, AudioInfo, StructureRef implements Send and
988 - VideoFrame has a function to get the raw FFI pointer
989 - From impls from the Error/Success enums to the combined enums like
991 - Bin-to-dot file functions were added to the Bin trait
992 - gst_base::Adapter implements SendUnique now
993 - More complete bindings for the gst_video::VideoOverlay interface,
995 gst_video::is_video_overlay_prepare_window_handle_message()
999 - All references were updated from GitHub to freedesktop.org GitLab
1000 - Fix various links in the README.md
1001 - Link to the correct location for the documentation
1002 - Remove GitLab badge as that only works with gitlab.com currently
1004 Changes in git master for 0.13
1008 - gst::tag::Album is the album tag now instead of artist sortname
1012 - Subclassing infrastructure was moved directly into the bindings,
1013 making the gst-plugin crate deprecated. This involves many API
1014 changes but generally cleans up code and makes it more flexible.
1015 Take a look at the gst-plugins-rs crate for various examples.
1017 - Bindings for CapsFeatures and Meta
1019 ParentBufferMeta,VideoMetaandVideoOverlayCompositionMeta`
1020 - Bindings for VideoOverlayComposition and VideoOverlayRectangle
1021 - Bindings for VideoTimeCode
1023 - UniqueFlowCombiner and UniqueAdapter wrappers that make use of the
1024 Rust compile-time mutability checks and expose more API in a safe
1025 way, and as a side-effect implement Sync and Send now
1027 - More complete bindings for Allocation Query
1028 - pbutils functions for codec descriptions
1029 - TagList::iter() for iterating over all tags while getting a single
1030 value per tag. The old ::iter_tag_list() function was renamed to
1031 ::iter_generic() and still provides access to each value for a tag
1032 - Bus::iter() and Bus::iter_timed() iterators around the corresponding
1035 - serde serialization of Value can also handle Buffer now
1037 - Extensive comments to all examples with explanations
1038 - Transmuxing example showing how to use typefind, multiqueue and
1040 - basic-tutorial-12 was ported and added
1044 - Rust 1.31 is the minimum supported Rust version now
1045 - Update to latest gir code generator and glib bindings
1047 - Functions returning e.g. gst::FlowReturn or other “combined” enums
1048 were changed to return split enums like
1049 Result<gst::FlowSuccess, gst::FlowError> to allow usage of the
1050 standard Rust error handling.
1052 - MiniObject subclasses are now newtype wrappers around the underlying
1053 GstRc<FooRef> wrapper. This does not change the API in any breaking
1054 way for the current usages, but allows MiniObjects to also be
1055 implemented in other crates and makes sure rustdoc places the
1056 documentation in the right places.
1058 - BinExt extension trait was renamed to GstBinExt to prevent conflicts
1059 with gtk::Bin if both are imported
1061 - Buffer::from_slice() can’t possible return None
1063 - Various clippy warnings
1066 GStreamer Rust Plugins
1068 Like the GStreamer Rust bindings, the Rust plugins are now officially
1069 part of the GStreamer project and are also maintained in the GStreamer
1072 In the 0.3.x versions this contained infrastructure for writing
1073 GStreamer plugins in Rust, and a set of plugins.
1075 In git master that infrastructure was moved to the GLib and GStreamer
1076 bindings directly, together with many other improvements that were made
1077 possible by this, so the gst-plugins-rs repository only contains
1078 GStreamer elements now.
1080 Elements included are:
1082 - Tutorials plugin: identity, rgb2gray and sinesrc with extensive
1085 - rsaudioecho, a port of the audiofx element
1087 - rsfilesrc, rsfilesink
1089 - rsflvdemux, a FLV demuxer. Not feature-equivalent with flvdemux yet
1091 - threadshare plugin: ts-appsrc, ts-proxysrc/sink, ts-queue, ts-udpsrc
1092 and ts-tcpclientsrc elements that use a fixed number of threads and
1093 share them between instances. For more background about these
1094 elements see Sebastian’s talk “When adding more threads adds more
1095 problems - Thread-sharing between elements in GStreamer” at the
1096 GStreamer Conference 2017.
1098 - rshttpsrc, a HTTP source around the hyper/reqwest Rust libraries.
1099 Not feature-equivalent with souphttpsrc yet.
1101 - togglerecord, an element that allows to start/stop recording at any
1102 time and keeps all audio/video streams in sync.
1104 - mccparse and mccenc, parsers and encoders for the MCC closed caption
1107 Changes to 0.3.X since 0.3.0
1109 - All references were updated from GitHub to freedesktop.org GitLab
1110 - Fix various links in the README.md
1111 - Link to the correct location for the documentation
1113 Changes in git master for 0.4
1115 - togglerecord: Switch to parking_lot crate for mutexes/condition
1116 variables for lower overhead
1117 - Merge threadshare plugin here
1118 - New closedcaption plugin with mccparse and mccenc elements
1119 - New identity element for the tutorials plugin
1121 - Register plugins statically in tests instead of relying on the
1122 plugin loader to find the shared library in a specific place
1124 - Update to the latest API changes in the GLib and GStreamer bindings
1125 - Update to the latest versions of all crates
1128 Build and Dependencies
1130 - The MESON BUILD SYSTEM BUILD IS NOW FEATURE-COMPLETE (*) and it is
1131 now the recommended build system on all platforms and also used by
1132 Cerbero to build GStreamer on all platforms. The Autotools build is
1133 scheduled to be removed in the next cycle. Developers who currently
1134 use gst-uninstalled should move to gst-build. The build option
1135 naming has been cleaned up and made consistent and there are now
1136 feature options to enable/disable plugins and various other features
1137 on a case-by-case basis. (*) with the exception of plugin docs which
1138 will be handled differently in future
1140 - Symbol export in libraries is now controlled via explicit exports
1141 using symbol visibility or export defines where supported, to ensure
1142 consistency across all platforms. This also allows libraries to have
1143 exports that vary based on detected platform features and configure
1144 options as is the case with the GStreamer OpenGL integration library
1145 for example. A few symbols that had been exported by accident in
1146 earlier versions may no longer be exported. These symbols will not
1147 have had declarations in any public header files then though and
1148 would not have been usable.
1150 - The GStreamer FFmpeg wrapper plugin (gst-libav) now depends on
1151 FFmpeg 4.x and uses the new FFmpeg 4.x API and stopped relying on
1152 ancient API that was removed with the FFmpeg 4.x release. This means
1153 that it is no longer possible to build this module against an older
1154 system-provided FFmpeg 3.x version. Use the internal FFmpeg 4.x copy
1155 instead if you build using autotools, or use gst-libav 1.14.x
1156 instead which targets the FFmpeg 3.x API and _should_ work fine in
1157 combination with a newer GStreamer. It’s difficult for us to support
1158 both old and new FFmpeg APIs at the same time, apologies for any
1159 inconvenience caused.
1161 - Hardware-accelerated Nvidia video encoder/decoder plugins nvdec and
1162 nvenc can be built against CUDA Toolkit versions 9 and 10.0 now. The
1163 dynlink interface has been dropped since it’s deprecated in 10.0.
1165 - The (optional) OpenCV requirement has been bumped to >= 3.0.0 and
1166 the plugin can also be built against OpenCV 4.x now.
1168 - New sctp plugin based on usrsctp (for WebRTC data channels)
1172 Cerbero is a meta build system used to build GStreamer plus dependencies
1173 on platforms where dependencies are not readily available, such as
1174 Windows, Android, iOS and macOS.
1176 Cerbero has seen a number of improvements:
1178 - Cerbero has been ported to Python 3 and requires Python 3.5 or newer
1181 - Source tarballs are now protected by checksums in the recipes to
1182 guard against download errors and malicious takeover of projects or
1183 websites. In addition, downloads are only allowed via secure
1184 transports now and plain HTTP, FTP and git:// transports are not
1187 - There is now a new fetch-bootstrap command which downloads sources
1188 required for bootstrapping, with an optional --build-tools-only
1189 argument to match the bootstrap --build-tools-only command.
1191 - The bootstrap, build, package and bundle-source commands gained a
1192 new --offline switch that ensures that only sources from the cache
1193 are used and never downloaded via the network. This is useful in
1194 combination with the fetch and fetch-bootstrap commands that acquire
1195 sources ahead of time before any build steps are executed. This
1196 allows more control over the sources used and when sources are
1197 updated, and is particularly useful for build environments that
1198 don’t have network access.
1200 - bootstrap --assume-yes will automatically say ‘yes’ to any
1201 interactive prompts during the bootstrap stage, such as those from
1204 - bootstrap --system-only will only bootstrap the system without build
1207 - Manifest support: The build manifest can be used in continuous
1208 integration (CI) systems to fixate the Git revision of certain
1209 projects so that all builds of a pipeline are on the same reference.
1210 This is used in GStreamer’s gitlab CI for example. It can also be
1211 used in order to re-produce a specific build. To set a manifest, you
1212 can set manifest = 'my_manifest.xml' in your configuration file, or
1213 use the --manifest command line option. The command line option will
1214 take precedence over anything specific in the configuration file.
1216 - The new build-deps command can be used to build only the
1217 dependencies of a recipe, without the recipe itself.
1219 - new --list-variants command to list available variants
1221 - variants can now be set on the command line via the -v option as a
1222 comma-separated list. This overrides any variants set in any
1223 configuration files.
1225 - new qt5, intelmsdk and nvidia variants for enabling Qt5 and hardware
1226 codec support. See the Enabling Optional Features with Variants
1227 section in the Cerbero documentation for more details how to enable
1228 and use these variants.
1230 - When building on Windows, Cerbero can now build GStreamer recipes
1231 and core dependencies such as glib with Visual Studio. This is
1232 controlled by the visualstudio variant. Visual Studio 2015, 2017,
1233 and 2019 are supported. Currently, only 64-bit x86 is supported due
1234 to a known bug which will be fixed for the next release.
1236 - A new -t / --timestamp command line switch makes commands print
1240 Platform-specific changes and improvements
1244 - toolchain: update compiler to clang and NDKr18. NDK r18 removed the
1245 armv5 target and only has Android platforms that target at least
1246 armv7 so the armv5 target is not useful anymore.
1248 - The way that GIO modules are named has changed due to upstream GLib
1249 natively adding support for loading static GIO modules. This means
1250 that any GStreamer application using gnutls for SSL/TLS on the
1251 Android or iOS platforms (or any other setup using static libraries)
1252 will fail to link looking for the g_io_module_gnutls_load_static()
1253 function. The new function name is now
1254 g_io_gnutls_load(gpointer data). data can be NULL for a static
1255 library. Look at this commit for the necessary change in the
1258 - various build issues on Android have been fixed.
1262 - various build issues on iOS have been fixed.
1264 - the minimum required iOS version is now 9.0. The difference in
1265 adoption between 8.0 and 9.0 is 0.1% and the bump to 9.0 fixes some
1268 - The way that GIO modules are named has changed due to upstream GLib
1269 natively adding support for loading static GIO modules. This means
1270 that any GStreamer application using gnutls for SSL/TLS on the
1271 Android or iOS platforms (or any other setup using static libraries)
1272 will fail to link looking for the g_io_module_gnutls_load_static()
1273 function. The new function name is now
1274 g_io_gnutls_load(gpointer data). data can be NULL for a static
1275 library. Look at this commit for the necessary change in the
1280 - The webrtcdsp element is shipped again as part of the Windows binary
1281 packages, the build system issue has been resolved.
1283 - ‘Inconsistent DLL linkage’ warnings when building with MSVC have
1286 - Hardware-accelerated Nvidia video encoder/decoder plugins nvdec and
1287 nvenc build on Windows now, also with MSVC and using Meson.
1289 - The ksvideosrc camera capture plugin supports 16-bit grayscale video
1292 - The wasapisrc audio capture element implements loopback recording
1293 from another output device or sink
1295 - wasapisink recover from low buffer levels in shared mode and some
1296 exclusive mode fixes
1298 - dshowsrc now implements the GstDeviceMonitor interface
1303 Aaron Boxer, Aleix Conchillo Flaqué, Alessandro Decina, Alexandru Băluț,
1304 Alex Ashley, Alexey Chernov, Alicia Boya García, Amit Pandya, Andoni
1305 Morales Alastruey, Andreas Frisch, Andre McCurdy, Andy Green, Anthony
1306 Violo, Antoine Jacoutot, Antonio Ospite, Arun Raghavan, Aurelien Jarno,
1307 Aurélien Zanelli, ayaka, Bananahemic, Bastian Köcher, Branko Subasic,
1308 Brendan Shanks, Carlos Rafael Giani, Charlie Turner, Christoph Reiter,
1309 Corentin Noël, Daeseok Youn, Damian Vicino, Dan Kegel, Daniel Drake,
1310 Daniel Klamt, Danilo Spinella, Dardo D Kleiner, David Ing, David
1311 Svensson Fors, Devarsh Thakkar, Dimitrios Katsaros, Edward Hervey,
1312 Emilio Pozuelo Monfort, Enrique Ocaña González, Erlend Eriksen, Ezequiel
1313 Garcia, Fabien Dessenne, Fabrizio Gennari, Florent Thiéry, Francisco
1314 Velazquez, Freyr666, Garima Gaur, Gary Bisson, George Kiagiadakis, Georg
1315 Lippitsch, Georg Ottinger, Geunsik Lim, Göran Jönsson, Guillaume
1316 Desmottes, H1Gdev, Haihao Xiang, Haihua Hu, Harshad Khedkar, Havard
1317 Graff, He Junyan, Hoonhee Lee, Hosang Lee, Hyunjun Ko, Ilya Smelykh,
1318 Ingo Randolf, Iñigo Huguet, Jakub Adam, James Stevenson, Jan Alexander
1319 Steffens, Jan Schmidt, Jerome Laheurte, Jimmy Ohn, Joakim Johansson,
1320 Jochen Henneberg, Johan Bjäreholt, John-Mark Bell, John Bassett, John
1321 Nikolaides, Jonathan Karlsson, Jonny Lamb, Jordan Petridis, Josep Torra,
1322 Joshua M. Doe, Jos van Egmond, Juan Navarro, Julian Bouzas, Jun Xie,
1323 Junyan He, Justin Kim, Kai Kang, Kim Tae Soo, Kirill Marinushkin, Kyrylo
1324 Polezhaiev, Lars Petter Endresen, Linus Svensson, Louis-Francis
1325 Ratté-Boulianne, Lucas Stach, Luis de Bethencourt, Luz Paz, Lyon Wang,
1326 Maciej Wolny, Marc-André Lureau, Marc Leeman, Marco Trevisan (Treviño),
1327 Marcos Kintschner, Marian Mihailescu, Marinus Schraal, Mark Nauwelaerts,
1328 Marouen Ghodhbane, Martin Kelly, Matej Knopp, Mathieu Duponchelle,
1329 Matteo Valdina, Matthew Waters, Matthias Fend, memeka, Michael Drake,
1330 Michael Gruner, Michael Olbrich, Michael Tretter, Miguel Paris, Mike
1331 Wey, Mikhail Fludkov, Naveen Cherukuri, Nicola Murino, Nicolas Dufresne,
1332 Niels De Graef, Nirbheek Chauhan, Norbert Wesp, Ognyan Tonchev, Olivier
1333 Crête, Omar Akkila, Pat DeSantis, Patricia Muscalu, Patrick Radizi,
1334 Patrik Nilsson, Paul Kocialkowski, Per Forlin, Peter Körner, Peter
1335 Seiderer, Petr Kulhavy, Philippe Normand, Philippe Renon, Philipp Zabel,
1336 Pierre Labastie, Piotr Drąg, Roland Jon, Roman Sivriver, Roman Shpuntov,
1337 Rosen Penev, Russel Winder, Sam Gigliotti, Santiago Carot-Nemesio,
1338 Sean-Der, Sebastian Dröge, Seungha Yang, Shi Yan, Sjoerd Simons, Snir
1339 Sheriber, Song Bing, Soon, Thean Siew, Sreerenj Balachandran, Stefan
1340 Ringel, Stephane Cerveau, Stian Selnes, Suhas Nayak, Takeshi Sato,
1341 Thiago Santos, Thibault Saunier, Thomas Bluemel, Tianhao Liu,
1342 Tim-Philipp Müller, Tobias Ronge, Tomasz Andrzejak, Tomislav Tustonić,
1343 U. Artie Eoff, Ulf Olsson, Varunkumar Allagadapa, Víctor Guzmán, Víctor
1344 Manuel Jáquez Leal, Vincenzo Bono, Vineeth T M, Vivia Nikolaidou, Wang
1345 Fei, wangzq, Whoopie, Wim Taymans, Wind Yuan, Wonchul Lee, Xabier
1346 Rodriguez Calvar, Xavier Claessens, Haihao Xiang, Yacine Bandou,
1347 Yeongjin Jeong, Yuji Kuwabara, Zeeshan Ali,
1349 … and many others who have contributed bug reports, translations, sent
1350 suggestions or helped testing.
1355 After the 1.16.0 release there will be several 1.16.x bug-fix releases
1356 which will contain bug fixes which have been deemed suitable for a
1357 stable branch, but no new features or intrusive changes will be added to
1358 a bug-fix release usually. The 1.16.x bug-fix releases will be made from
1359 the git 1.16 branch, which is a stable branch.
1363 1.16.0 was released on 19 April 2019.
1367 The first 1.16 bug-fix release (1.16.1) was released on 23 September
1370 This release only contains bugfixes and it _should_ be safe to update
1373 Highlighted bugfixes in 1.16.1
1375 - GStreamer-vaapi: fix green frames and decoding artefacts in some
1377 - OpenGL: fix wayland event source burning CPU in certain
1379 - Memory leak fixes and memory footprint improvements
1380 - Performance improvements
1381 - Stability and security fixes
1382 - Fix enum for GST_MESSAGE_DEVICE_CHANGED which is technically an API
1383 break, but this is only used internally in GStreamer and duplicated
1384 another message enum
1385 - hls: Make crypto dependency optional when hls-crypto is auto
1386 - player: fix switching back and forth between forward and reverse
1388 - decklinkaudiosink: Drop late buffers
1389 - openh264enc: Fix compilation with openh264 v2.0
1390 - wasapisrc: fix segtotal value being always 2
1391 - Fix issues on Android Q
1395 - device: gst_device_create_element() is transfer floating, not
1397 - filesink, fdsink: respect IOV_MAX for the writev iovec array
1399 - miniobject: free qdata array when the last qdata is removed (reduces
1401 - bin: Fix minor race when adding to a bin
1402 - aggregator: Actually handle NEED_DATA return from update_src_caps()
1403 - aggregator: Ensure that the source pad is created as a
1404 GstAggregatorPad if no type is given in the pad template
1405 - latency: fix custom event leaks
1406 - registry: Use plugin directory from the build system for
1407 relocateable Windows builds
1408 - message: fix up enum value for GST_MESSAGE_DEVICE_CHANGED
1409 - info: Fix deadlock in gst_ring_buffer_logger_log()
1410 - downloadbuffer: Check for flush after seek
1411 - identity: Non-live upstream have no max latency
1412 - identity: Fix the ts-offset property getter
1413 - aggregator: Make parsing of explicit sink pad names more robust
1414 - bufferpool: Fix the buffer size reset code
1415 - fakesink, fakesrc, identity: sync gst_buffer_get_flags_string() with
1417 - multiqueue: never unref queries we do not own
1418 - concat: Reset last_stop on FLUSH_STOP too
1419 - aggregator: fix flow-return boolean return type mismatch
1420 - gstpad: Handle probes that reset the data field
1421 - gst: Add support for g_autoptr(GstPromise)
1422 - gst-inspect: fix unused-const-variable error in windows
1423 - base: Include gstbitwriter.h in the single-include header
1424 - Add various Since: 1.16 markers
1425 - GST_MESSAGE_DEVICE_CHANGED duplicates GST_MESSAGE_REDIRECT
1426 - Targetting wrong meson version
1427 - meson: Make get_flex_version.py script executable
1428 - meson: Link to objects instead of static helper library
1429 - meson: set correct install path for gdb helper
1430 - meson: fix warning about configure_file() install kwarg
1434 - video-info: parse field-order for all interleaved formats
1435 - tests: fix up valgrind suppressions for glibc getaddrinfo leaks
1436 - meson: Reenable NEON support (in audio resampler)
1437 - audio-resampler: Update NEON to handle remainders not multiples of 4
1438 - eglimage: Fix memory leak
1439 - audiodecoder: Set output caps with negotiated caps to avoid critical
1441 - video-frame: Take TFF flag from the video info if it was set in
1443 - glcolorconvert: Fix external-oes shader
1444 - video-anc: Fix ADF detection when trying to extract data from vanc
1445 - gl/wayland: fix wayland event source burning CPU
1446 - configure: add used attribute in order to make NEON detection
1448 - audioaggregator: Return a valid rate range from caps query if
1449 downstream supports a whole range
1450 - rtspconnection: data-offset increase not set
1451 - rtpsconnection: Fix number of n_vectors
1452 - video-color: Add compile-time assert for ColorimetryInfo enum
1453 - audiodecoder: Fix leak on failed audio gaps
1454 - glupload: Keep track of cached EGLImage texture format
1455 - playsink: Set ts-offset to text sink.
1456 - meson.build: use join_paths() on prefix
1457 - compositor: copy frames as-is when possible
1458 - compositor: Skip background when a pad obscures it completely
1459 - rtspconnection: Start CSeq at 1 (some servers don’t cope well with
1461 - viv-fb: fix build break for GST_GL_API
1462 - gl/tests: fix shader creation tests part 2
1463 - gl/tests: fix shader creation tests
1464 - wayland: set the event queue also for the xdg_wm_base object
1465 - video: Added GI annotation for gstvideoaffinetransformationmeta
1467 - compositor: Remove unneeded left shift for ARGB/AYUV SOURCE operator
1469 - alsasrc: Don’t use driver timestamp if it’s zero
1470 - gloverlaycompositor: fix crash if buffer doesn’t have video meta
1471 - meson: Don’t try to find gio-unix on Windows
1472 - glshader: fix default external-oes shaders
1473 - subparse: fix pushing WebVTT cue with no newline at the end
1474 - meson: Missing “android” choice in gl_winsys
1475 - video test: Keep BE test inline with LE test
1476 - id3tag: Correctly validate the year from v1 tags before passing to
1478 - gl/wayland: Don’t prefix wl_shell struct field
1479 - eglimage: Add compatibility define for DRM_FORMAT_NV24
1480 - Add various Since: 1.16 markers
1481 - video-anc: Handle SD formats correctly
1482 - Docs: add GL_CFLAGS to GTK_DOC_CFLAGS
1483 - GL: using vaapi and showing on glimagesink on wayland loads one core
1485 - GL: external-oes shader places precision qualifier before #extension
1486 (was: androidmedia amcviddec fail after 1.15.90 1.16.0 update)
1490 - alpha: Fix one_over_kc calculation on arm/aarch64
1491 - souphttpsrc: Fix incompatible type build warning
1492 - rtpjitterbuffer: limit max-dropout-time to maxint32
1493 - rtpjitterbuffer: Clear clock master before unreffing
1494 - qtdemux: Use empty-array safe way to cleanup GPtrArray
1495 - v4l2: Fix type compatibility issue with glibc 2.30
1496 - valgrind: suppress Cond error coming from gnutls and Ignore leaks
1497 caused by shout/sethostent
1498 - rtpfunnel: forward correct segment when switching pad
1499 - gtkglsink: fix crash when widget is resized after element
1501 - jpegdec: Don’t dereference NULL input state if we have no caps in
1503 - rtp: opuspay: fix memory leak in gst_rtp_opus_pay_setcaps
1504 - v4l2videodec: return right type for drain.
1505 - rtpssrcdemux: Avoid taking streamlock out-of-band
1506 - Support v4l2src buffer orphaning
1507 - splitmuxsink: Only set running time on finalizing sink element when
1508 in async-finalize mode
1509 - rtpsession: Always keep at least one NACK on early RTCP
1510 - rtspsrc: do not try to send EOS with invalid seqnum
1511 - rtpsession: Call on-new-ssrc earlier
1512 - rtprawdepay: Don’t get rid of the buffer pool on FLUSH_STOP
1513 - rtpbin: Free storage when freeing session
1514 - scaletempo: Advertise interleaved layout in caps templates
1515 - Support v4l2src buffer orphaning
1519 - hls: Make crypto dependency optional when hls-crypto is auto
1520 - player: fix switching back and forth between forward and reverse
1522 - decklinkaudiosink: Drop late buffers
1523 - srt: Add stats property, include sender-side statistics and fix a
1525 - dshowsrcwrapper: fix regression on device selection
1526 - tsdemux: Limit the maximum PES payload size
1527 - wayland: Define libdrm_dep in meson.build to fix meson configure
1528 error when kms is disabled
1529 - sctp: Fix crash on free() when using the MSVC binaries
1530 - webrtc: Fix signals documentation
1531 - h264parse: don’t critical on VUI parameters > 2^31
1532 - rtmp: Fix crash inside free() with MSVC on Windows
1533 - iqa: fix leak of map_meta.data
1534 - d3dvideosink: Fix crash on WinProc handler
1535 - amc: Fix crash when a sync_meta survives its sink
1536 - pitch: Fix race between putSamples() and setting soundtouch
1538 - webrtc: fix type of max-retransmits, make it work
1539 - mxfdemux: Also allow picture essence element type 0x05 for VC-3
1540 - wasapi: fix symbol redefinition build error
1541 - decklinkvideosrc: Retrieve mode of the ancillary data from the frame
1542 - decklinkaudiosrc/decklinkvideosrc: Do nothing in
1543 BaseSrc::negotiate() and…
1544 - adaptivedemux: do not retry downloads during shutdown.
1545 - webrtcbin: fix GInetAddress leak
1546 - dtls: fix dtls connection object leak
1547 - siren: fix a global buffer overflow spotted by asan
1548 - kmssink: Fix implicit declaration build error
1549 - Fix -Werror=return-type error in configure.
1550 - aiff: Fix infinite loop in header parsing.
1551 - nvdec: Fix possible frame drop on EOS
1552 - srtserversrc: yields malformed rtp payloads
1553 - srtsink: Fix crash in case no URI
1554 - dtlsagent: Fix leaked dtlscertificate
1555 - meson: bluez: Early terminate configure on Windows
1556 - decklink: Correctly ensure >=16 byte alignment for the buffers we
1558 - webrtcbin: fix DTLS when receivebin is set to DROP
1559 - zbar: Include running-time, stream-time and duration in the messages
1560 - uvch264src: Make sure we set our segment
1561 - avwait: Allow start and end timecode to be set back to NULL
1562 - avwait: Don’t print warnings for every buffer passed
1563 - hls/meson: fix dependency logic
1564 - Waylandsink gnome shell workaround
1565 - avwait: Allow setting start timecode after end timecode; protect
1566 propeties with mutex
1567 - wayland/wlbuffer: just return if used_by_compositor is true when
1569 - proxy: Set SOURCE flag on the source and SINK flag on the sink
1570 - ivfparse: Check the data size against IVF_FRAME_HEADER_SIZE
1571 - webrtc: Add various Since markers to new types after 1.14.0
1572 - msdk: fix the typo in debug category
1573 - dtlsagent: Do not overwrite openssl locking callbacks
1574 - meson: Fix typo in gsm header file name
1575 - srt: handle races in state change
1576 - webrtc: Add g_autoptr() support for public types
1577 - openh264enc: Fix compilation with openh264 v2.0
1578 - meson: Allow CUDA_PATH fallback on linux
1579 - meson: fix build with opencv=enabled and opencv4. Fixes #964
1580 - meson: Add support for the colormanagement plugin
1581 - autotools: gstsctp: set LDFLAGS
1582 - nvenc/nvdec: Add NVIDIA SDK headers to noinst_HEADERS
1583 - h264parse: Fix typo when setting multiview mode and flags
1584 - Add various Since: 1.16 markers
1585 - opencv: allow compilation against 4.1.x
1586 - Backport of some minor srt commits without MR into 1.16
1587 - meson: fix build with opencv=enabled and opencv4
1588 - wasapisrc: fix segtotal value being always 2 due to an unused
1590 - meson: colormanagement missing
1591 - androidmedia amcviddec fail after 1.15.90 1.16.0 update
1595 - meson: Always require the gmodule dependency
1599 - docs: don’t include the type hierarchy, fixing build with gtk-doc
1601 - avvidenc: Correctly signal interlaced input to ffmpeg when the input
1603 - autotools: add bcrypt to win32 libs
1604 - gstav: Use libavcodec util function for version check
1605 - API documentation fails to build with gtk-doc 1.30
1609 - rtsp-client: RTP Info must exist in PLAY response
1610 - onvif-media: fix “void function returning a value” compiler warning
1611 - Add various Since: 1.16 markers
1615 - fix egl context leak and display creation race
1616 - pluginutil: Remove Mesa from drivers white list
1617 - Classify vaapidecodebin as a hardware decoder
1619 - vaapivideomemory: demote error message to info
1620 - encoder: vp8,vp9: reset frame_counter when input frame’s format
1622 - encoder: mpeg2: No packed header for SPS and PPS
1623 - decoder: vp9: clear parser pointer after release
1624 - encoder: Fixes deadlock in change state function
1625 - encoder: h265: reset num_ref_idx_l1_active_minus1 when low delay B.
1626 - encoder: not call ensure_num_slices inside g_assert()
1627 - encoder: continue if roi meta is NULL
1628 - decoder: vp9: Set chroma_ ype by VP9 bit_depth
1629 - vaapipostproc: don’t do any color conversion when GL_TEXTURE_UPLOAD
1630 - libs: surface: fix double free when dmabuf export fails
1631 - h264 colors and artifacts upon upgrade to GStreamer Core Library
1634 gst-editing-services
1636 - element: Properly handle the fact that pasting can return NULL
1637 - Add various missing Since markers
1638 - launch: Fix caps restriction short names
1639 - python: Avoid warning about using deprecated methods
1640 - video-transition: When using non crossfade effect use ‘over’
1642 - meson: Generate a pkgconfig file for the GES plugin
1646 - launcher: testsuites: skip systemclock stress tests
1647 - validate: fix build on macOS
1651 - Update win flex bison binaries
1652 - Update the flexmeson windows binary version
1653 - Don’t allow people to run meson inside the uninstalled env
1655 Cerbero build tool and packaging changes in 1.16.1
1657 - cerbero: Add enums for Fedora 30, Fedora 31 and Debian bullseye
1658 - gnutls.recipe: Fix crash when running on Android Q
1659 - recipes: Upgrade openssl to 1.1.1c
1661 - add support for vs build tools 2019, fixes #183
1662 - android: Adjust gstreamer-1.0.mk for NDK r20
1664 - bootstrap: Fix dnf usage on CentOS
1665 - Make _add_system_libs reentrant
1666 - meson.recipe: Fix setting of bitcode compiler options
1667 - cerbero: support Ubuntu disco dingo
1668 - cerbero: Set utf-8 to execution character set also on MSVC
1669 - git: simplify the reset of the source branch.
1670 - FORTIFY: %n not allowed on Android Q
1671 - Fails to build if there’s no license file for the given license
1672 (GPL/LGPL without Plus, Proprietary, …)
1674 Contributors to 1.16.1
1676 Aaron Boxer, Adam Duskett, Alicia Boya García, Andoni Morales Alastruey,
1677 Antonio Ospite, Arun Raghavan, Askar Safin, A. Wilcox, Charlie Turner,
1678 Christoph Reiter, Damian Hobson-Garcia, Daniel Klamt, Danny Smith, David
1679 Gunzinger, David Ing, David Svensson Fors, Doug Nazar, Edward Hervey,
1680 Eike Hein, Fabrice Bellet, Fernando Herrrera, Georg Lippitsch, Göran
1681 Jönsson, Guillaume Desmottes, Haihao Xiang, Haihua Hu, Håvard Graff, Hou
1682 Qi, Ignacio Casal Quinteiro, Ilya Smelykh, Jan Schmidt, Javier Celaya,
1683 Jim Mason, Jonas Larsson, Jordan Petridis, Jose Antonio Santos Cadenas,
1684 Juan Navarro, Knut Andre Tidemann, Kristofer Björkström, Lucas Stach,
1685 Marco Felsch, Marcos Kintschner, Mark Nauwelaerts, Martin Liska, Martin
1686 Theriault, Mathieu Duponchelle, Matthew Waters, Michael Olbrich, Mike
1687 Gorse, Nicola Murino, Nicolas Dufresne, Niels De Graef, Niklas
1688 Hambüchen, Nirbheek Chauhan, Olivier Crête, Philippe Normand, Ross
1689 Burton, Sebastian Dröge, Seungha Yang, Song Bing, Thiago Santos,
1690 Thibault Saunier, Thomas Coldrick, Tim-Philipp Müller, Víctor Manuel
1691 Jáquez Leal, Vivia Nikolaidou, Xavier Claessens, Yeongjin Jeong,
1693 … and many others who have contributed bug reports, translations, sent
1694 suggestions or helped testing. Thank you all!
1696 List of merge requests and issues fixed in 1.16.1
1698 - List of Merge Requests applied in 1.16
1699 - List of Issues fixed in 1.16.1
1704 - possibly breaking/incompatible changes to properties of wrapped
1705 FFmpeg decoders and encoders (see above).
1707 - The way that GIO modules are named has changed due to upstream GLib
1708 natively adding support for loading static GIO modules. This means
1709 that any GStreamer application using gnutls for SSL/TLS on the
1710 Android or iOS platforms (or any other setup using static libraries)
1711 will fail to link looking for the g_io_module_gnutls_load_static()
1712 function. The new function name is now
1713 g_io_gnutls_load(gpointer data). See Android/iOS sections above for
1719 Our next major feature release will be 1.18, and 1.17 will be the
1720 unstable development version leading up to the stable 1.18 release. The
1721 development of 1.17/1.18 will happen in the git master branch.
1723 The plan for the 1.18 development cycle is yet to be confirmed, but it
1724 is now expected that feature freeze will take place shortly after the
1725 GStreamer conference/hackfest in early November 2019, with the first
1726 1.18 stable release ready in late November or early December.
1728 1.18 will be backwards-compatible to the stable 1.16, 1.14, 1.12, 1.10,
1729 1.8, 1.6, 1.4, 1.2 and 1.0 release series.
1731 ------------------------------------------------------------------------
1733 _These release notes have been prepared by Tim-Philipp Müller with_
1734 _contributions from Sebastian Dröge, Guillaume Desmottes, Matthew
1735 Waters, _ _Thibault Saunier, and Víctor Manuel Jáquez Leal._
1737 _License: CC BY-SA 4.0_