ffmpeg | branch: master | Fei Wang | Thu Aug 15 10:19:50
2024 +0800| [be7ab635522adf1331a2aa105a49045ba8e65192] | committer: Haihao Xiang
lavc/qsvdec: Add vvc_mp4toannexb bsf for QSV VVC decoder
Fix error:
$ ffmpeg -hwaccel qsv -i input.mp4 -f null -
..
[vvc_qsv @ 026890D966C0] Error decodi
ffmpeg | branch: master | Lynne | Fri Aug 16 01:12:16 2024
+0200| [a797317ab19d12df8c953dab3de548d14c4ba545] | committer: Lynne
vulkan_filter: don't require the storage flag for the base frames format
We check for whether subformats support storage immediately below.
Those are the ones we requi
ffmpeg | branch: master | Lynne | Thu Aug 15 00:54:48 2024
+0200| [b165f144e7426eb90467158c65218353ee9da36a] | committer: Lynne
vulkan_filter: allow reusing frame contexts with DRM tiling
There's no reason not to permit this, particularly if a user wants
to manipulate images which will be expor
ffmpeg | branch: master | Lynne | Thu Aug 15 00:52:33 2024
+0200| [604dfdb44c7261b8a8608d55727d2a7e848d3d1e] | committer: Lynne
hwcontext_vulkan: align host mapping size to minImportedHostPointerAlignment
This was left out of the recent rewrite of the system.
> http://git.videolan.org/gitweb.c
ffmpeg | branch: master | Lynne | Wed Aug 14 17:25:00 2024
+0200| [18d964fc2c520dd690aa3d737aea62e60fc3786b] | committer: Lynne
vulkan: enable encoding of images if video_maintenance1 is enabled
Vulkan encoding was designed in a very... consolidated way.
You had to know the exact codec and prof
ffmpeg | branch: master | Lynne | Wed Aug 14 15:52:13 2024
+0200| [c3cbaf39bb30bfaf5849def2020bdb6c3b2d5bc6] | committer: Lynne
hwcontext_vulkan: don't enable deprecated VK_KHR_sampler_ycbcr_conversion
extension
It was added to Vulkan 1.1 a long time ago.
Validation layer will warn if this is
ffmpeg | branch: master | Lynne | Wed Aug 14 16:31:58 2024
+0200| [97e947a2a729915759d511fdd2816249731bb41d] | committer: Lynne
hwcontext_vulkan: setup extensions before features
The issue is that enabling features requires that the device
extension is supported. The extensions bitfield was set
ffmpeg | branch: master | Lynne | Wed Aug 14 17:10:42 2024
+0200| [46c13834b6a08d4e7d9b2c457805bfdca55f4613] | committer: Lynne
hwcontext_vulkan: enable VK_KHR_video_maintenance1
We require it for encoding.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=46c13834b6a08d4e7d9b2c45780
ffmpeg | branch: master | Lynne | Wed Aug 14 15:50:03 2024
+0200| [3f65d24075b5c0d8c657b94fc74156b44159a8ba] | committer: Lynne
hwcontext_vulkan: fix user layers, add support for different debug modes
The validation layer option only supported GPU-assisted validation.
This is mutually exclusive
ffmpeg | branch: master | Lynne | Wed Aug 14 14:22:28 2024
+0200| [869f4aec484be200ea676d7d45c5b12c9b914434] | committer: Lynne
vulkan_decode: use the correct queue family for decoding ops
In 680d969a305c0927480573a1b455024088b51aeb, the new API was
used to find a queue family for dispatch, but
ffmpeg | branch: master | Anton Khirnov | Tue Aug 6
08:50:21 2024 +0200| [42cbf66fffed5a8bba2f14b5247ce60d788b9e01] | committer:
Anton Khirnov
lavfi: move AVFilterLink.{frame,sample}_count_{in,out} to FilterLink
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=42cbf66fffed5a8bba2f1
ffmpeg | branch: master | Anton Khirnov | Sun Aug 11
16:35:52 2024 +0200| [d566a37003d0257229f06436bef8062fee3af6e9] | committer:
Anton Khirnov
lavfi: move AVFilterLink.graph to FilterLink
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d566a37003d0257229f06436bef8062fee3af6e9
---
ffmpeg | branch: master | Anton Khirnov | Sun Aug 11
16:31:11 2024 +0200| [fb3efef1dbc96915edeb1df8027d1f1502324fc0] | committer:
Anton Khirnov
lavfi: move AVFilterLink.frame_wanted_out to FilterLinkInternal
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=fb3efef1dbc96915edeb1df802
ffmpeg | branch: master | Anton Khirnov | Mon Aug 5
18:00:30 2024 +0200| [7f17e0e6dd0b28b29295c1eeb1a38546eebc3946] | committer:
Anton Khirnov
lavfi: move AVFilterLink.hw_frames_ctx to FilterLink
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7f17e0e6dd0b28b29295c1eeb1a38546eebc3
ffmpeg | branch: master | Anton Khirnov | Tue Aug 6
08:50:21 2024 +0200| [a23d565ea7d41e61f160578f9714a23e695f3bfd] | committer:
Anton Khirnov
lavfi: move AVFilterLink.frame_rate to FilterLink
Co-developed-by: James Almer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a23d565ea
ffmpeg | branch: master | Anton Khirnov | Tue Aug 6
07:08:48 2024 +0200| [d6318a244d1dde48c1429c48b4076df47c2c7fe9] | committer:
Anton Khirnov
lavfi: move AVFilterLink.current_pts(_us) to FilterLink
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d6318a244d1dde48c1429c48b4076df47c
ffmpeg | branch: master | Anton Khirnov | Fri Aug 2
13:55:21 2024 +0200| [426e33c75853c395b39e216e7ecde0103ff48086] | committer:
Anton Khirnov
lavfi: set AVFilterLink.graph on link creation
There is no reason to delay this.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=426e33c7
ffmpeg | branch: master | Anton Khirnov | Mon Aug 5
12:13:52 2024 +0200| [ce24b5ba8f8d61eae518809e1188f74614db73d5] | committer:
Anton Khirnov
lavfi/vf_*_cuda: do not access hw contexts before checking they exist
The checks are performed in init_processing_chain().
> http://git.videolan.org/
ffmpeg | branch: master | Anton Khirnov | Mon Aug 5
11:41:53 2024 +0200| [99ee7a948fd5bc014ce5a4160a71ffb068fd21e2] | committer:
Anton Khirnov
lavfi: move AVFilterLink.m{ax,in}_samples to FilterLink
Also, document who sets these fields and when.
> http://git.videolan.org/gitweb.cgi/ffmpeg.gi
ffmpeg | branch: master | Anton Khirnov | Mon Aug 5
10:49:18 2024 +0200| [54754eec1e4b4882e7bbe1fbf4374c11d955eb36] | committer:
Anton Khirnov
lavfi: add a new struct for private link properties
Specifically those that should be visible to filters, but hidden from
API callers. Such properties
ffmpeg | branch: master | Wu Jianhua | Sun Aug 11 04:33:04
2024 +0800| [ca5c9e810ae4f70b6212bfaa8d61ee0a5150db80] | committer: Nuo Mi
avcodec/vvc/dsp: prefix TxType and TxSize with VVC
See
https://patchwork.ffmpeg.org/project/ffmpeg/patch/tyspr06mb64337c4a9adf5312e6648543aa...@tyspr06mb6433.ap
ffmpeg | branch: master | Wu Jianhua | Sun Aug 11 04:33:02
2024 +0800| [ae1a9cfd52c72e6d2703275b258cb2ded86a519d] | committer: Nuo Mi
avcodec/vvc_parser: move avctx->has_b_frames initialization to dec
>From Jun Zhao :
> Should we relocate this to the decoder? Other codecs typically set this
> p
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:04
2024 +0800| [0c7106f618f75461386e485db814c3e8c761b0c4] | committer: Nuo Mi
avcodec/vvcdec: do not zero frame cpm table
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=0c7106f618f75461386e485db814c3e8c761b0c4
---
libavcodec/vvc
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:02
2024 +0800| [4f882c453ef61c6c308f0e9fb8ff1d91d138f3de] | committer: Nuo Mi
avcodec/vvcdec: do not zero frame mvf table
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=4f882c453ef61c6c308f0e9fb8ff1d91d138f3de
---
libavcodec/vvc
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:07
2024 +0800| [80af195804fdef9ccd5a48251fc366d28bceb437] | committer: Nuo Mi
avcodec/vvcdec: move frame tab memset from the main thread to worker threads
memset tables in the main thread can become a bottleneck for the decoder.
For example, i
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:05
2024 +0800| [ca2caeb21d18c75dd43cfb5111629dc2d8de4c87] | committer: Nuo Mi
avcodec/vvcdec: do not zero frame msf mmi table
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=ca2caeb21d18c75dd43cfb5111629dc2d8de4c87
---
libavcodec
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:03
2024 +0800| [3e497249ebd0634e6fe6b044896f883a608d51ec] | committer: Nuo Mi
avcodec/vvcdec: check_available, use && instead of &= for shortcut evaluation
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3e497249ebd0634e6fe6b044896
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:06
2024 +0800| [daf6fcd816783dfd1f82a795ed774933880abcc1] | committer: Nuo Mi
avcodec/vvcdec: do not zero frame qp table
For luma, qp can only change at the CU level, so the qp tab size is related to
the CU.
For chroma, considering the joint
ffmpeg | branch: master | Nuo Mi | Sun Jan 14 11:58:37
2024 +0800| [93b0d5f9ce906b622431a4c4a17f8e1695306e52] | committer: Nuo Mi
avcodec/vvcdec: refact out is_available from is_a0_available
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=93b0d5f9ce906b622431a4c4a17f8e1695306e52
---
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:17:59
2024 +0800| [4f609614983bcabc277d354ef8bc4e65e52faaf6] | committer: Nuo Mi
avcodec/vvcdec: remove unnecessary perframe initializations
deblock, sao, alf
skip, imtf, ipm, cqt_depth, cb_pos_x, cb_pos_y, cb_height, cp_mv,
tb_pos_x0, tb_pos_y0,
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:17:58
2024 +0800| [210bf1b1544fc141765c20dcec9acb55dba40b3e] | committer: Nuo Mi
avcodec/vvcdec: refact, combine bs tab with tu tab
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=210bf1b1544fc141765c20dcec9acb55dba40b3e
---
libavco
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:17:57
2024 +0800| [bdb79fe60afaf54505f1ffa8b5324bf05c700929] | committer: Nuo Mi
avcodec/vvcdec: thread, ensure the parse stage gets the highest priority
The parser stage is not parallelizable.
We need to schedule it as soon as possible to create
ffmpeg | branch: master | Nuo Mi | Sun Jul 28 11:18:00
2024 +0800| [e8ac76129365302de7e2885b76459ad890b7ecc7] | committer: Nuo Mi
avcodec/vvcdec: split ctu table to zero init and no zero init parts
cus need to init to zero, other parts are not
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a
ffmpeg | branch: master | Nuo Mi | Thu Jul 25 21:35:44
2024 +0800| [1c17520e7987e1e3d93d18d4c9b59541e4815826] | committer: Nuo Mi
avcodec/vvcdec: Use av_image_copy_plane for DMVR 10-bit integer pixels
It's no need to shift and interpolate for 10-bit integer pixels,
av_image_copy_plane is enough
ffmpeg | branch: master | Nuo Mi | Thu Jul 25 21:35:46
2024 +0800| [7eb1df44ae6daea1fa68ff2fdd96a13d3111be3e] | committer: Nuo Mi
checkasm: add tests for vvc dmvr
dmvr_8_12x20_c: 186.2
dmvr_8_12x20_avx2: 25.7
dmvr_8_20x12_c: 181.7
dmvr_8_20x12_avx2: 25.2
dmvr_8_20x20_c: 283.2
dmvr_8_20x20_avx2:
ffmpeg | branch: master | Nuo Mi | Thu Jul 25 21:35:45
2024 +0800| [6464c157d588e0a9f32c9a1b228d5a30106e93e3] | committer: Nuo Mi
x86/vvcdec: add dmvr avx2 code
Decoder-Side Motion Vector Refinement is about 4~8% CPU usage for some clips
here is the test result for one time
clips
ffmpeg | branch: master | eaphone | Wed Aug 7 23:58:07
2024 +0800| [7897b0beedc822129f7e8e429995f9b6b3860d74] | committer: Zhao Zhili
libavdevice/gdigrab: change hwnd tail check fail logic to !=null
Signed-off-by: Zhao Zhili
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7897b0b
ffmpeg | branch: master | gnattu | Sat Aug 10 14:13:39 2024
+0800| [a1976e963fa441b6f645c425186f272d05def3f9] | committer: Zhao Zhili
avutil/hwcontext_videotoolbox: silence warning for RGB
Hardware frames with RGB colorspace will not have a YCbCrMatrixKey.
Currently, it will spam the console wi
ffmpeg | branch: master | Araz Iusubov | Thu Aug 15
12:57:59 2024 +0200| [42a40538f3138ab4ebeb99513865baf3824660e5] | committer:
Dmitrii Ovchinnikov
avcodec/amfenc: new encoder features support
Implemented:
New usage modes for AV1 encoder.
Latency mode for H264, HEVC and AV1 encoders.
Adaptive
ffmpeg | branch: master | Gyan Doshi | Sun Aug 11 13:09:50
2024 +0530| [1f801dfdb5066aadf0ade9cb5e94d620f33eacdc] | committer: Gyan Doshi
lavc/libx265: unbreak build for X265_BUILD >= 210
x265 added support for alpha starting with build 210.
While doing so, x265_encoder_encode() changed its fif
40 matches
Mail list logo