Important

We recommend running CasparCG 2.5 on CPUs which support AVX2. Officially Chrome claims to require AVX2, and it is required for some of our in-progress HDR support.
Intel CPUs based on Haswell or later support this, which were first released to consumers in 2013, or 2014 for servers.

Starting with CasparCG 2.6, this will become a requirement.

Make sure to install the latest MSVC runtime, the compiler update requires a newer version than 2.4 https://aka.ms/vc14/vc_redist.x64.exe

Core

Improvements

• Initial support for HDR. This is limited to a subset of producers and consumers at this stage.
• Build for Windows with VS2022
• Rework linux builds to produce ubuntu deb files
• Update ffmpeg to 7.0
• Reimplement mixer transforms, to handle routes correctly
• Support more pixel formats from ffmpeg, to preserve colour accuracy better
• Support running on headless linux
• Transitions: Additional behaviours

Fixes

• Build with boost 1.85/1.86/1.87/1.88
• Build with ffmpeg 7.1
• Only produce mixed frames on channels which have consumers
• Routed channels not compositing correctly when channel used a MIXER KEY
• Handle audio for fractional framerates properly
• Gracefully exit on SIGINT and SIGTERM

Producers

Improvements

• FFmpeg: Support loading with a scaling-mode, to configure how clips get fit into the channel
• FFmpeg: Support more pixel formats without cpu conversion
• FFmpeg: Enable alpha for webm videos
• Image: Support loading with a scaling-mode, to configure how images get fit into the channel
• Image: Replace freeimage with ffmpeg
• HTML: Update CEF to 142
• HTML: Support audio

Fixes

• Route: Use full field rate when performing i->p channel route
• HTML: Gracefully handle page load errors
• HTML: Always set cache path

Consumers

Improvements

• Screen: Set size and position from AMCP
• Screen: Improve performance
• Image: Propagate AMCP parameters from PRINT command
• FFmpeg: Remove unnecessary forced conversion to YUVA422
• Decklink: Support explicit yuv output (requires AVX2)
• Decklink: Allow selecting device by hardware persistent id

Fixes

• FFmpeg: Correctly handle PTS on frame drop

Full Changelog: v2.4.3-stable...v2.5.0-stable

A new major release, FFmpeg 8.0 "Huffman", is now available for download. Thanks to several delays, and modernization of our entire infrastructure, this release ended up being one of our largest releases to date. In short, its new features are:

• Native decoders: APV, ProRes RAW, RealVideo 6.0, Sanyo LD-ADPCM, G.728
• VVC decoder improvements: IBC, ACT, Palette Mode
• Vulkan compute-based codecs: FFv1 (encode and decode), ProRes RAW (decode only)
• Hardware accelerated decoding: Vulkan VP9, VAAPI VVC, OpenHarmony H264/5
• Hardware accelerated encoding: Vulkan AV1, OpenHarmony H264/5
• Formats: MCC, G.728, Whip, APV
• Filters: colordetect, pad_cuda, scale_d3d11, Whisper, and others

A new class of decoders and encoders based on pure Vulkan compute implementation have been added. Vulkan is a cross-platform, open standard set of APIs that allows programs to use GPU hardware in various ways, from drawing on screen, to doing calculations, to decoding video via custom hardware accelerators. Rather than using a custom hardware accelerator present, these codecs are based on compute shaders, and work on any implementation of Vulkan 1.3.
Decoders use the same hwaccel API and commands, so users do not need to do anything special to enable them, as enabling Vulkan decoding is sufficient to use them.
Encoders, like our hardware accelerated encoders, require specifying a new encoder (ffv1_vulkan). Currently, the only codecs supported are: FFv1 (encoding and decoding) and ProRes RAW (decode only). ProRes (encode+decode) and VC-2 (encode+decode) implementations are complete and currently in review, to be merged soon and available with the next minor release.
Only codecs specifically designed for parallelized decoding can be implemented in such a way, with more mainstream codecs not being planned for support.
Depending on the hardware, these new codecs can provide very significant speedups, and open up possibilities to work with them for situations like non-linear video editors and lossless screen recording/streaming, so we are excited to learn what our downstream users can make with them.

The project has recently started to modernize its infrastructure. Our mailing list servers have been fully upgraded, and we have recently started to accept contributions via a new forge, available on code.ffmpeg.org, running a Forgejo instance.

As usual, we recommend that users, distributors, and system integrators to upgrade unless they use current git master.

Core

Fixes

• Improve error handling for invalid config files #1571
• Flush logs before exit #1571
• Check audio cadence values look sane before accepting format #1588
• Cross-channel routes from progressive to interlaced showing lots of black #1576
• Transition: ignoring some transforms of input frames #1602

Producers

Fixes

• FFmpeg: fix crash on invalid frame header
• Decklink: Crash with ffmpeg 7 #1582
• HTML: Fix crash during uninit on exit
• Image: update state during init #1601

Consumers

Fixes

• FFmpeg: set frame_rate for rtmp streams #1462

Full Changelog: v2.4.2-stable...v2.4.3-stable

Consumers

Fixes

• Decklink: fix support for driver 14.3 and later

Full Changelog: v2.4.1-stable...v2.4.2-stable

Core

Fixes

• Fix bad config file examples
• Fix casparcg_auto_restart.bat not starting scanner
• Revert removal of tbbmalloc, due to notable performance loss on windows
• Supress some cmake build warnings
• Build failure when doxygen installed on system
• Build failures with ffmpeg 7.0
• Revert RPATH linking changes

Producers

Fixes

• FFmpeg: Ignore ndi:// urls
• FFmpeg: Using both in and seek could result in incorrect duration
• Route: Race condition during destruction
• Image: Update freeimage on windows with some CVE fixes and failures with certain pngs
• Image: Respect EXIF rotate flag
• NDI: list local sources

Consumers

Fixes

• Decklink: subregion copy not respecting frame height
• Decklink: subregion vertical offset
• Decklink: subregion height limited with some formats

Full Changelog: v2.4.0-stable...v2.4.1-stable

Changes since 2.4.0 RC 1:

Core

Improvements

• Support 4K DCI frames
• Build: Allow configuring diag font path at build time
• AMCP: Add CALLBG command to perform CALL on background producer

Fixes

• OSC: Background state not being produced
• Scanner: resolve issues with database not being updated

Producers

Improvements

• HTML: Expose cache-path setting

Fixes

• FFmpeg: Unable to play files with extra dots in filenames
• FFmpeg: Support parameters with name containing a dash
• FFmpeg: Audio channels being swapped unexpectedly

Changes since 2.3.3

Core

Improvements

• Custom resultions can be specified in casparcg.config
• Interlaced mixer pipeline to ensure field accuracy
• Preserve unicode characters in console input/output
• Producers to be run at startup can be defined in casparcg.config
• Support 8K frames
• Support 4K DCI frames
• Remove undocumented CII and CLK protocol implementations
• Config parameter can be an absolute system path, not just relative to the working directory
• AMCP: Add CLEAR ALL command
• AMCP: Command batching syntax
• AMCP: LOAD/LOADBG/PLAY commands accept a CLEAR_ON_404 parameter, to instruct the layer to be cleared when the requested file was not found
• AMCP: Add commands to subscribe and unsubscribe to OSC on any port number
• AMCP: Add CALLBG command to perform CALL on background producer
• Build: Require C++17 for building
• Build: Support newer versions of Boost
• Build: Support newer versions of TBB
• Build: Disable precompiled headers for linux
• Build: Support VS2022
• Build: Replace nuget and locally committed dependencies with direct http downloads
• Build: Allow configuring diag font path at build time
• Linux: Support setting thread priorities
• Linux: Initial ARM64 compatibility
• Linux: Rework build to always use system boost
• Linux: Rework build process to better support being build as a system package
• Logging: add config option to disable logging to file and to disable column alignment
• Transitions: Support additional audio fade properties for STING transition

Fixes

• Crash upon exiting if HTML producer was running
• AMCP: Ensure all consumers and producers are reported in INFO commands
• AMCP: Deferred mixer operations were not being cleared after being applied
• AMCP: LOAD command would show a frame or two of black while new producer was loading
• OpenGL: Fix support for recent Linux drivers
• Linux: Fix endless looping on stdin
• Route: Fix error when clearing layer
• Transitions: Fix wipe duration

Producers

Improvements

• Decklink: Require driver 11.0 or later
• Decklink: Scale received frames on GPU
• FFmpeg: Update to v5.1
• FFmpeg: Improve performance
• FFmpeg: Allow specifying both SEEK and IN for PLAY commands
• HTML: Update to CEF 117
• HTML: CALL 1-10 RELOAD to reload a renderer
• HTML: Expose cache-path setting
• NDI: Upgrade to NDI5
• System Audio: Allow specifying output device to use

Fixes

• Decklink: Log spamming when using some input formats
• FFmpeg: Prevent loading unreadable files
• FFmpeg: Unable to play files with unicode filenames
• FFmpeg: Don't lowercase filter parameters
• FFmpeg: Support parameters with name containing a dash
• HTML: media-stream permission denied
• HTML: Expose angle backend config field, the best backend varies depending on the templates and machine
• HTML: Crash when multiple iframes were loaded within a renderer
• Image: Improve file loading algorithm to match the case insensitive and absolute path support already used by ffmpeg

Consumers

Improvements

• Artnet: New artnet consumer
• Decklink: Configure device duplex modes in casparcg.config
• Decklink: Output a subregion of the channel
• Decklink: Add secondary outputs in a consumer, to ensure sync when used within a single card
• iVGA: Remove consumer
• NDI: Upgrade to NDI5

Fixes

• Decklink: Fix stutter when loading clips
• FFmpeg: Fix RTMP streaming missing headers
• NDI: dejitter

Changes since Beta 1:

Important

There is a known bug in the version of media-scanner included in this release.
It is recommended to replace it with v1.3.2 or newer from https://github.com/CasparCG/media-scanner/releases

Core

Improvements

• Config parameter can be an absolute system path, not just relative to the working directory
• Linux: Rework build to always use system boost
• Linux: Rework build process to better support being build as a system package
• Audio: 16 channel audio

Fixes

• Crash upon exiting if HTML producer was running
• AMCP: LOAD command would show a frame or two of black while new producer was loading

Producers

Fixes

• FFmpeg: Unable to play files with extra dots in filenames
• FFmpeg: Don't lowercase filter parameters
• HTML: Debugger connection closes
• HTML: Expose angle backend config field, the best backend varies depending on the templates and machine
• HTML: Crash when multiple iframes were loaded within a renderer
• Image: Improve file loading algorithm to match the case insensitive and absolute path support already used by ffmpeg

CasparCG 2.4.0 Beta 1

Core

Improvements

• Custom resultions can be specified in casparcg.config
• Interlaced mixer pipeline to ensure field accuracy
• Preserve unicode characters in console input/output
• Producers to be run at startup can be defined in casparcg.config
• Support 8K frames
• Remove undocumented CII and CLK protocol implementations
• AMCP: Add CLEAR ALL command
• AMCP: Command batching syntax
• AMCP: LOAD/LOADBG/PLAY commands accept a CLEAR_ON_404 parameter, to instruct the layer to be cleared when the requested file was not found
• AMCP: Add commands to subscribe and unsubscribe to OSC on any port number
• Build: Require C++17 for building
• Build: Support newer versions of Boost
• Build: Support newer versions of TBB
• Build: Disable precompiled headers for linux
• Build: Support VS2022
• Build: Replace nuget and locally committed dependencies with direct http downloads
• Linux: Support setting thread priorities
• Linux: Initial ARM64 compatibility
• Logging: add config option to disable logging to file and to disable column alignment
• Transitions: Support additional audio fade properties for STING transition

Fixes

• AMCP: Ensure all consumers and producers are reported in INFO commands
• AMCP: Deferred mixer operations were not being cleared after being applied
• OpenGL: Fix support for recent Linux drivers
• Linux: Fix endless looping on stdin
• Route: Fix error when clearing layer
• Transitions: Fix wipe duration

Producers

Improvements

• Decklink: Require driver 11.0 or later
• Decklink: Scale received frames on GPU
• FFmpeg: Update to v5.1
• FFmpeg: Improve performance
• FFmpeg: Allow specifying both SEEK and IN for PLAY commands
• HTML: Update to CEF 117
• HTML: CALL 1-10 RELOAD to reload a renderer
• NDI: Upgrade to NDI5
• System Audio: Allow specifying output device to use

Fixes

• Decklink: Log spamming when using some input formats
• FFmpeg: Prevent loading unreadable files
• FFmpeg: Unable to play files with unicode filenames
• HTML: media-stream permission denied

Consumers

Improvements

• Artnet: New artnet consumer
• Decklink: Configure device duplex modes in casparcg.config
• Decklink: Output a subregion of the channel
• Decklink: Add secondary outputs in a consumer, to ensure sync when used within a single card
• iVGA: Remove consumer
• NDI: Upgrade to NDI5

Fixes

• Decklink: Fix stutter when loading clips
• FFmpeg: Fix RTMP streaming missing headers
• NDI: dejitter

Improvements:

• Image Scroll Producer

Producers

Fixes

• Packages: Update TBB library to v2021.1.1 - fixes CPU and memory growth when deleting threads
• FFmpeg: Fix possible deadlock leading to producer not being cleaned up correctly

Producers

Fixes

• Packages: Update TBB library to v2021.1.1 - fixes CPU and memory growth when deleting threads
• FFmpeg: Fix possible deadlock leading to producer not being cleaned up correctly

This is the first patch release for 2.3.

Primarily this contains a crtical fix to provide support for flash templates to be used after the flash player EOL at the end of 2020.
For more information on the issue see: #1352

Thanks to @silid for developing the fix and @didikunz for documenting the process and being a key member of chasing for these releases to be made

See the attached PDF, or the copy in the Zip for steps on how to prepare your system.

For any help or questions, check the forum

Changes

Producers

Fixes

• Flash: Use proper file urls when loading templates, to allow it to work after Flash Player EOL
• FFmpeg: Various HTTP playback improvements

This is a critical fix for the 2.0.7 release, to provide support for flash templates to be used after the flash player EOL at the end of 2020

For more information on the issue see: #1352

This has been built to be as close as possible to the previous 2.0.7 release, to minimise any risk of deployment.

Thanks to @silid for developing the fix and @didikunz for documenting the process and being a key member of chasing for these releases to be made

See the attached PDF, or the copy in the Zip for steps on how to prepare your system.

For any help or questions, check the forum

CasparCG Server v2.3.0 LTS Stable 🚀

We're are really thrilled to announce that we are releasing first LTS version ever of CapsarCG Server, version 2.3.0 LTS. We have focused on stability, reliability and the code base in this version so we have a solid base to stand on for future versions. For a complete list of changes please check out the changes below:

A version of CasparCG Client that works with this version of the server is found as attachment below

Changes

Producers

Features

• FFmpeg: Add more common file extensions to the supported list
• NDI: Require minimum of NDI v4.0
• Flash: Disable by default, requires enabling in the config file
• FFmpeg: Remove fixed thread limit to better auto select a number
• Decklink: Detect and update input format when no format is specified in AMCP
• Decklink: Improve performance (gpu colour conversion & less heavy deinterlacing when possible)
• Decklink: LOAD DECKLINK will display live frames instead of black
• FFmpeg: Update to 4.2.2
• HTML: Better performance for gpu-enabled mode
• HTML: window.remove() has been partially reimplemented
• NDI: Native NDI producer
• Route: Allow routing first frame of background producer
• Route: zero delay routes when within a channel, with 1 frame when cross-channel
• Transition: Add sting transitions
• Add frames_left field to osc/info for progress towards autonext

Fixes

• HTML: Minimise performance impact on other producers
• Decklink: Downgrade severity of video-format not supported
• FFmpeg: Correctly handle error codes. Ignore exit errors during initialisation
• Route: Detect circular routes and break the loop
• Colour: parsing too much of amcp string as list of colours
• FFmpeg: Always resample clips to 48khz
• FFmpeg: Ensure frame time reaches the end of the clip
• FFmpeg: RTMP stream playback
• FFmpeg: SEEK and LENGTH parameters causing issues with AUTONEXT
• FFmpeg: Ensure packets/frames after the decided end of the clip are not displayed
• FFmpeg: Incorrect seek for audio when not 48khz
• FFmpeg: Some cases where it would not be destroyed if playing a bad stream
• HTML: unlikely but possible exception when handling frames
• HTML: set autoplay-policy
• HTML: animations being ticked too much
• Route: Sending empty frame into a route would cause the destination to reuse the last frame

Consumers

Features

• Bluefish: Various improvmements including support for Kronos K8
• Audio: Fix audio crackling
• Audio: Fix memory leak
• Bluefish: Various improvmements including supporting more channels and UHD.
• NDI: Native NDI consumer
• Screen: Add side by side key output
• Screen: Add support for Datavideo TC-100/TC-200

Fixes

• Decklink: Tick channel at roughly consistent rate when running interlaced output
• Possible crash when adding/removing consumers

General

Features

• Add mixer colour invert property
• Restore INFO CONFIG and INFO PATHS commands
• Linux: Update docker images to support running in docker (not recommended for production use)

Fixes

• Diag not reflecting channel videoformat changes
• NTSC audio cadence
• Ignore empty lines in console input
• Fix building with clang on linux
• Fix building with vs2019
• Better error when startup fails due to AMCP port being in use
• Backslash is a valid trailing slash for windows

CasparCG Server v2.3.0 LTS RC (Release Candidate)

Continued work with the first LTS version ever of CasparCG Server. With that said, we're are happy to announce that we are releasing CapsarCG Server v2.3.0 LTS Release Candidate (RC). As mentioned in previous release, we are working and focusing on stability and reliability and a stable release is planned at the end of Q2.

CasparCG Server v2.3.0 LTS RC can be downloaded directly from this release (see attachments).

Changes

Producers

Features

• Flash: Disable by default, requires enabling in the config file
• FFmpeg: Remove fixed thread limit to better auto select a number

Fixes

• Decklink: Downgrade severity of video-format not supported
• FFmpeg: Correctly handle error codes. Ignore exit errors during initialisation
• Route: Detect circular routes and break the loop

Consumers

Features

• Bluefish: Various improvmements including support for Kronos K8

General

Fixes

• Diag not reflecting channel videoformat changes

FFmpeg 7.1 "Péter", a new major release, is now available! A full list of changes can be found in the release changelog.

The more important highlights of the release are that the VVC decoder, merged as experimental in version 7.0, has had enough time to mature and be optimized enough to be declared as stable. The codec is starting to gain traction with broadcast standardization bodies.
Support has been added for a native AAC USAC (part of the xHE-AAC coding system) decoder, with the format starting to be adopted by streaming websites, due to its extensive volume normalization metadata.
MV-HEVC decoding is now supported. This is a stereoscopic coding tool that begun to be shipped and generated by recent phones and VR headsets.
LC-EVC decoding, an enhancement metadata layer to attempt to improve the quality of codecs, is now supported via an external library.

Support for Vulkan encoding, with H264 and HEVC was merged. This finally allows fully Vulkan-based decode-filter-encode pipelines, by having a sink for Vulkan frames, other than downloading or displaying them. The encoders have feature-parity with their VAAPI implementation counterparts. Khronos has announced that support for AV1 encoding is also coming soon to Vulkan, and FFmpeg is aiming to have day-one support.

In addition to the above, this release has had a lot of important internal work done. By far, the standout internally are the improvements made for full-range images. Previously, color range data had two paths, no negotiation, and was unreliably forwarded to filters, encoders, muxers. Work on cleaning the system up started more than 10 years ago, however this stalled due to how fragile the system was, and that breaking behaviour would be unacceptable. The new system fixes this, so now color range is forwarded correctly and consistently everywhere needed, and also laid the path for more advanced forms of negotiation.
Cropping metadata is now supported with Matroska and MP4 formats. This metadata is important not only for archival, but also with AV1, as hardware encoders require its signalling due to the codec not natively supporting one.

As usual, we recommend that users, distributors, and system integrators to upgrade unless they use current git master.

The number of issues FFmpeg has in Coverity (a static analyzer) is now lower than it has been since 2016. Our defect density is less than one 30th of the average in OSS with over a million code lines. All this was possible thanks to a grant from the Sovereign Tech Fund.

FFmpeg now implements a native xHE-AAC decoder. Currently, streams without (e)SBR, USAC or MPEG-H Surround are supported, which means the majority of xHE-AAC streams in use should work. Support for USAC and (e)SBR is coming soon. Work is also ongoing to improve its stability and compatibility. During the process we found several specification issues, which were then submitted back to the authors for discussion and potential inclusion in a future errata.

The FFmpeg community is excited to announce that Germany's Sovereign Tech Fund has become its first governmental sponsor. Their support will help sustain the maintainance of the FFmpeg project, a critical open-source software multimedia component essential to bringing audio and video to billions around the world everyday.

A new major release, FFmpeg 7.0 "Dijkstra", is now available for download. The most noteworthy changes for most users are a native VVC decoder (currently experimental, until more fuzzing is done), IAMF support, or a multi-threaded ffmpeg CLI tool.

This release is not backwards compatible, removing APIs deprecated before 6.0. The biggest change for most library callers will be the removal of the old bitmask-based channel layout API, replaced by the AVChannelLayout API allowing such features as custom channel ordering, or Ambisonics. Certain deprecated ffmpeg CLI options were also removed, and a C11-compliant compiler is now required to build the code.

As usual, there is also a number of new supported formats and codecs, new filters, APIs, and countless smaller features and bugfixes. Compared to 6.1, the git repository contains almost ∼2000 new commits by ∼100 authors, touching >100000 lines in ∼2000 files — thanks to everyone who contributed. See the Changelog, APIchanges, and the git log for more comprehensive lists of changes.

The libavcodec library now contains a native VVC (Versatile Video Coding) decoder, supporting a large subset of the codec's features. Further optimizations and support for more features are coming soon. The code was written by Nuo Mi, Xu Mu, Frank Plowman, Shaun Loo, and Wu Jianhua.

The libavformat library can now read and write IAMF (Immersive Audio) files. The ffmpeg CLI tool can configure IAMF structure with the new -stream_group option. IAMF support was written by James Almer.

Thanks to a major refactoring of the ffmpeg command-line tool, all the major components of the transcoding pipeline (demuxers, decoders, filters, encodes, muxers) now run in parallel. This should improve throughput and CPU utilization, decrease latency, and open the way to other exciting new features.

Note that you should not expect significant performance improvements in cases where almost all computational time is spent in a single component (typically video encoding).

FFmpeg 6.1 "Heaviside", a new major release, is now available! Some of the highlights:

• libaribcaption decoder
• Playdate video decoder and demuxer
• Extend VAAPI support for libva-win32 on Windows
• afireqsrc audio source filter
• arls filter
• ffmpeg CLI new option: -readrate_initial_burst
• zoneplate video source filter
• command support in the setpts and asetpts filters
• Vulkan decode hwaccel, supporting H264, HEVC and AV1
• color_vulkan filter
• bwdif_vulkan filter
• nlmeans_vulkan filter
• RivaTuner video decoder
• xfade_vulkan filter
• vMix video decoder
• Essential Video Coding parser, muxer and demuxer
• Essential Video Coding frame merge bsf
• bwdif_cuda filter
• Microsoft RLE video encoder
• Raw AC-4 muxer and demuxer
• Raw VVC bitstream parser, muxer and demuxer
• Bitstream filter for editing metadata in VVC streams
• Bitstream filter for converting VVC from MP4 to Annex B
• scale_vt filter for videotoolbox
• transpose_vt filter for videotoolbox
• support for the P_SKIP hinting to speed up libx264 encoding
• Support HEVC,VP9,AV1 codec in enhanced flv format
• apsnr and asisdr audio filters
• OSQ demuxer and decoder
• Support HEVC,VP9,AV1 codec fourcclist in enhanced rtmp protocol
• CRI USM demuxer
• ffmpeg CLI '-top' option deprecated in favor of the setfield filter
• VAAPI AV1 encoder
• ffprobe XML output schema changed to account for multiple variable-fields elements within the same parent element
• ffprobe -output_format option added as an alias of -of

This release had been overdue for at least half a year, but due to constant activity in the repository, had to be delayed, and we were finally able to branch off the release recently, before some of the large changes scheduled for 7.0 were merged.

Internally, we have had a number of changes too. The FFT, MDCT, DCT and DST implementation used for codecs and filters has been fully replaced with the faster libavutil/tx (full article about it coming soon).
This also led to a reduction in the the size of the compiled binary, which can be noticeable in small builds.
There was a very large reduction in the total amount of allocations being done on each frame throughout video decoders, reducing overhead.
RISC-V optimizations for many parts of our DSP code have been merged, with mainly the large decoders being left.
There was an effort to improve the correctness of timestamps and frame durations of each packet, increasing the accurracy of variable frame rate video.

Next major release will be version 7.0, scheduled to be released in February. We will attempt to better stick to the new release schedule we announced at the start of this year.

We strongly recommend users, distributors, and system integrators to upgrade unless they use current git master.

A few days ago, Vulkan-powered decoding hardware acceleration code was merged into the codebase. This is the first vendor-generic and platform-generic decode acceleration API, enabling the same code to be used on multiple platforms, with very minimal overhead. This is also the first multi-threaded hardware decoding API, and our code makes full use of this, saturating all available decode engines the hardware exposes.

Those wishing to test the code can read our documentation page. For those who would like to integrate FFmpeg's Vulkan code to demux, parse, decode, and receive a VkImage to present or manipulate, documentation and examples are available in our source tree. Currently, using the latest available git checkout of our repository is required. The functionality will be included in stable branches with the release of version 6.1, due to be released soon.

As this is also the first practical implementation of the specifications, bugs may be present, particularly in drivers, and, although passing verification, the implementation itself. New codecs, and encoding support are also being worked on, by both the Khronos organization for standardizing, and us as implementing it, and giving feedback on improving.

A new major release, FFmpeg 6.0 "Von Neumann", is now available for download. This release has many new encoders and decoders, filters, ffmpeg CLI tool improvements, and also, changes the way releases are done. All major releases will now bump the version of the ABI. We plan to have a new major release each year. Another release-specific change is that deprecated APIs will be removed after 3 releases, upon the next major bump. This means that releases will be done more often and will be more organized.

New decoders featured are Bonk, RKA, Radiance, SC-4, APAC, VQC, WavArc and a few ADPCM formats. QSV and NVenc now support AV1 encoding. The FFmpeg CLI (we usually reffer to it as ffmpeg.c to avoid confusion) has speed-up improvements due to threading, as well as statistics options, and the ability to pass option values for filters from a file. There are quite a few new audio and video filters, such as adrc, showcwt, backgroundkey and ssim360, with a few hardware ones too. Finally, the release features many behind-the-scenes changes, including a new FFT and MDCT implementation used in codecs (expect a blog post about this soon), numerous bugfixes, better ICC profile handling and colorspace signalling improvement, introduction of a number of RISC-V vector and scalar assembly optimized routines, and a few new improved APIs, which can be viewed in the doc/APIchanges file in our tree. A few submitted features, such as the Vulkan improvements and more FFT optimizations will be in the next minor release, 6.1, which we plan to release soon, in line with our new release schedule. Some highlights are:

• Radiance HDR image support
• ddagrab (Desktop Duplication) video capture filter
• ffmpeg -shortest_buf_duration option
• ffmpeg now requires threading to be built
• ffmpeg now runs every muxer in a separate thread
• Add new mode to cropdetect filter to detect crop-area based on motion vectors and edges
• VAAPI decoding and encoding for 10/12bit 422, 10/12bit 444 HEVC and VP9
• WBMP (Wireless Application Protocol Bitmap) image format
• a3dscope filter
• bonk decoder and demuxer
• Micronas SC-4 audio decoder
• LAF demuxer
• APAC decoder and demuxer
• Media 100i decoders
• DTS to PTS reorder bsf
• ViewQuest VQC decoder
• backgroundkey filter
• nvenc AV1 encoding support
• MediaCodec decoder via NDKMediaCodec
• MediaCodec encoder
• oneVPL support for QSV
• QSV AV1 encoder
• QSV decoding and encoding for 10/12bit 422, 10/12bit 444 HEVC and VP9
• showcwt multimedia filter
• corr video filter
• adrc audio filter
• afdelaysrc audio filter
• WADY DPCM decoder and demuxer
• CBD2 DPCM decoder
• ssim360 video filter
• ffmpeg CLI new options: -stats_enc_pre[_fmt], -stats_enc_post[_fmt], -stats_mux_pre[_fmt]
• hstack_vaapi, vstack_vaapi and xstack_vaapi filters
• XMD ADPCM decoder and demuxer
• media100 to mjpegb bsf
• ffmpeg CLI new option: -fix_sub_duration_heartbeat
• WavArc decoder and demuxer
• CrystalHD decoders deprecated
• SDNS demuxer
• RKA decoder and demuxer
• filtergraph syntax in ffmpeg CLI now supports passing file contents as option values
• hstack_qsv, vstack_qsv and xstack_qsv filters

We strongly recommend users, distributors, and system integrators to upgrade unless they use current git master.

FFmpeg 5.1 "Riemann", a new major release, is now available! Some of the highlights:

• add ipfs/ipns protocol support
• dialogue enhance audio filter
• dropped obsolete XvMC hwaccel
• pcm-bluray encoder
• DFPWM audio encoder/decoder and raw muxer/demuxer
• SITI filter
• Vizrt Binary Image encoder/decoder
• avsynctest source filter
• feedback video filter
• pixelize video filter
• colormap video filter
• colorchart video source filter
• multiply video filter
• PGS subtitle frame merge bitstream filter
• blurdetect filter
• tiltshelf audio filter
• QOI image format support
• ffprobe -o option
• virtualbass audio filter
• VDPAU AV1 hwaccel
• PHM image format support
• remap_opencl filter
• added chromakey_cuda filter

We strongly recommend users, distributors, and system integrators to upgrade unless they use current git master.

FFmpeg 5.0 "Lorentz", a new major release, is now available! For this long-overdue release, a major effort underwent to remove the old encode/decode APIs and replace them with an N:M-based API, the entire libavresample library was removed, libswscale has a new, easier to use AVframe-based API, the Vulkan code was much improved, many new filters were added, including libplacebo integration, and finally, DoVi support was added, including tonemapping and remuxing. The default AAC encoder settings were also changed to improve quality. Some of the changelog highlights:

• ADPCM IMA Westwood encoder
• Westwood AUD muxer
• ADPCM IMA Acorn Replay decoder
• Argonaut Games CVG demuxer
• Argonaut Games CVG muxer
• Concatf protocol
• afwtdn audio filter
• audio and video segment filters
• Apple Graphics (SMC) encoder
• hsvkey and hsvhold video filters
• adecorrelate audio filter
• atilt audio filter
• grayworld video filter
• AV1 Low overhead bitstream format muxer
• swscale slice threading
• MSN Siren decoder
• scharr video filter
• apsyclip audio filter
• morpho video filter
• amr parser
• (a)latency filters
• GEM Raster image decoder
• asdr audio filter
• speex decoder
• limitdiff video filter
• xcorrelate video filter
• varblur video filter
• huesaturation video filter
• colorspectrum source video filter
• RTP packetizer for uncompressed video (RFC 4175)
• bitpacked encoder
• VideoToolbox VP9 hwaccel
• VideoToolbox ProRes hwaccel
• support loongarch.
• aspectralstats audio filter
• adynamicsmooth audio filter
• libplacebo filter
• vflip_vulkan, hflip_vulkan and flip_vulkan filters
• adynamicequalizer audio filter
• yadif_videotoolbox filter
• VideoToolbox ProRes encoder
• anlmf audio filter

We strongly recommend users, distributors, and system integrators to upgrade unless they use current git master.

We have a new IRC home at Libera Chat now! Feel free to join us at #ffmpeg and #ffmpeg-devel. More info at contact#IRCChannels

FFmpeg 4.4 "Rao", a new major release, is now available! Some of the highlights:

• AudioToolbox output device
• MacCaption demuxer
• PGX decoder
• chromanr video filter
• VDPAU accelerated HEVC 10/12bit decoding
• ADPCM IMA Ubisoft APM encoder
• Rayman 2 APM muxer
• AV1 encoding support SVT-AV1
• Cineform HD encoder
• ADPCM Argonaut Games encoder
• Argonaut Games ASF muxer
• AV1 Low overhead bitstream format demuxer
• RPZA video encoder
• ADPCM IMA MOFLEX decoder
• MobiClip FastAudio decoder
• MobiClip video decoder
• MOFLEX demuxer
• MODS demuxer
• PhotoCD decoder
• MCA demuxer
• AV1 decoder (Hardware acceleration used only)
• SVS demuxer
• Argonaut Games BRP demuxer
• DAT demuxer
• aax demuxer
• IPU decoder, parser and demuxer
• Intel QSV-accelerated AV1 decoding
• Argonaut Games Video decoder
• libwavpack encoder removed
• ACE demuxer
• AVS3 demuxer
• AVS3 video decoder via libuavs3d
• Cintel RAW decoder
• VDPAU accelerated VP9 10/12bit decoding
• afreqshift and aphaseshift filters
• High Voltage Software ADPCM encoder
• LEGO Racers ALP (.tun & .pcm) muxer
• AV1 VAAPI decoder
• adenorm filter
• ADPCM IMA AMV encoder
• AMV muxer
• NVDEC AV1 hwaccel
• DXVA2/D3D11VA hardware accelerated AV1 decoding
• speechnorm filter
• SpeedHQ encoder
• asupercut filter
• asubcut filter
• Microsoft Paint (MSP) version 2 decoder
• Microsoft Paint (MSP) demuxer
• AV1 monochrome encoding support via libaom >= 2.0.1
• asuperpass and asuperstop filter
• shufflepixels filter
• tmidequalizer filter
• estdif filter
• epx filter
• Dolby E parser
• shear filter
• kirsch filter
• colortemperature filter
• colorcontrast filter
• PFM encoder
• colorcorrect filter
• binka demuxer
• XBM parser
• xbm_pipe demuxer
• colorize filter
• CRI parser
• aexciter audio filter
• exposure video filter
• monochrome video filter
• setts bitstream filter
• vif video filter
• OpenEXR image encoder
• Simbiosis IMX decoder
• Simbiosis IMX demuxer
• Digital Pictures SGA demuxer and decoders
• TTML subtitle encoder and muxer
• identity video filter
• msad video filter
• gophers protocol
• RIST protocol via librist

We strongly recommend users, distributors, and system integrators to upgrade unless they use current git master.