Feed of "FFmpeg" https://code.ffmpeg.org/FFmpeg <p dir="auto">A complete, cross-platform solution to record, convert and stream audio and video.</p> Sun, 15 Mar 2026 13:40:49 +0000 mkver commented on pull request FFmpeg/FFmpeg#22509 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509#issuecomment-31971 swscale/ops_backend: avoid UB (null pointer arithmetic) <p dir="auto">This patch presumes that <code>(uintptr_t)(ptr + x) == (uintptr_t)ptr + x</code> for pointer types with <code>sizeof(*ptr) == 1</code>. This is not guaranteed by the spec (although I don&#39;t know a system where it is wrong).</p> <p dir="auto">Also: Do you ever add something nonzero to NULL pointers?</p> This patch presumes that (uintptr_t)(ptr + x) == (uintptr_t)ptr + x for pointer types with sizeof(*ptr) == 1. This is not guaranteed by the spec (although I don't know a system where it is wrong).

Also: Do you ever add something nonzero to NULL pointers?

]]>
mkver 923230: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509#issuecomment-31971 Sun, 15 Mar 2026 13:02:40 +0000
michaelni commented on pull request FFmpeg/FFmpeg#22509 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509#issuecomment-31969 swscale/ops_backend: avoid UB (null pointer arithmetic) <p dir="auto">though in general its more robust to not add to null pointers because the resulting pointers are no longer detectable as null yet they are invalid</p> though in general its more robust to not add to null pointers because the resulting pointers are no longer detectable as null yet they are invalid ]]> michaelni 923159: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509#issuecomment-31969 Sun, 15 Mar 2026 12:58:35 +0000 michaelni approved FFmpeg/FFmpeg#22509 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509#issuecomment-31968 swscale/ops_backend: avoid UB (null pointer arithmetic) michaelni 923088: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509#issuecomment-31968 Sun, 15 Mar 2026 12:54:51 +0000 Kenaz123 commented on pull request FFmpeg/FFmpeg#22505 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22505#issuecomment-31967 avcodec/xxan: zero-initialize y_buffer <p dir="auto"><a href="/michaelni" class="mention" rel="nofollow">@michaelni</a> I&#39;ve added the commit message. Thanks for looking!</p> @michaelni I've added the commit message. Thanks for looking! ]]> Kenaz123 923017: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22505#issuecomment-31967 Sun, 15 Mar 2026 12:38:37 +0000 Kenaz123 commented on issue FFmpeg/FFmpeg#22420 https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/22420#issuecomment-31963 xxan: uninitialized y_buffer read when first frame is type 1 <p dir="auto"><a href="/michaelni" class="mention" rel="nofollow">@michaelni</a> I rebuilt FFmpeg with debug info in a Valgrind-friendly configuration, and that resolved the missing line numbers:</p> @michaelni I rebuilt FFmpeg with debug info in a Valgrind-friendly configuration, and that resolved the missing line numbers: ]]> Kenaz123 922945: https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/22420#issuecomment-31963 Sun, 15 Mar 2026 12:24:03 +0000 Niklas Haas commented on pull request FFmpeg/FFmpeg#22463 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22463#issuecomment-31961 swscale: filter branch subset <p dir="auto"><a href="/haasn" class="mention" rel="nofollow">@haasn</a> wrote in <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22463#issuecomment-31806" class="ref-issue" rel="nofollow">#22463 (comment)</a>:</p> @haasn wrote in #22463 (comment): ]]> Niklas Haas 922874: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22463#issuecomment-31961 Sun, 15 Mar 2026 12:01:57 +0000 Niklas Haas created pull request FFmpeg/FFmpeg#22510 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22510 22510#swscale: various more more minor fixes and preparatory commits# In anticipation of the main filtering branch. ]]> Niklas Haas 922803: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22510 Sun, 15 Mar 2026 12:01:02 +0000 mkver commented on pull request FFmpeg/FFmpeg#20961 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20961#issuecomment-31948 Add myself as maintainer to the various ogg files. <p dir="auto">Sorry for not responding earlier; I no longer look at my mails that often since development moved to forgejo and therefore I have not seen the GA mail in time. I know that the decision has been made, but given that I have been asked to explain myself by both toots and michaelni I think I should do so.</p> <p dir="auto">I think that anyone who wants to be a maintainer needs to have a track record of good PRs, i.e. PRs that basically can be applied as-is, without multiple rounds of review and that do not cause regressions. Also the good PRs should be the clear majority of the PRs of the aspiring maintainer.</p> <p dir="auto">This is not so for toots: In <a href="/FFmpeg/FFmpeg/issues/21787" class="ref-issue" rel="nofollow">#21787</a> you did not update FATE test results before submitting a PR; when you saw the runners falling, you then sent ever more updates that also did not pass instead of running FATE locally and updating the references. At one point (that commit is now 404) did you even intend to commit one of the test tool binaries.</p> <p dir="auto">Your chained ogg PR <a href="/FFmpeg/FFmpeg/issues/20876" class="ref-issue" rel="nofollow">#20876</a> caused memleaks and needed to be reverted (as did your older &#34;header packet skip&#34; commit 574f634e49847e2225ee50013afebf0de03ef013).</p> <p dir="auto"><a href="/FFmpeg/FFmpeg/issues/21813" class="ref-issue" rel="nofollow">#21813</a> and <a href="/FFmpeg/FFmpeg/issues/21199" class="ref-issue" rel="nofollow">#21199</a> were based on misunderstandings of the underlying specifications and needed to be abandonded. The latter also would have introduced many more issues if applied (even if the understanding of the spec were correct).</p> <p dir="auto">In <a href="/FFmpeg/FFmpeg/issues/20867" class="ref-issue" rel="nofollow">#20867</a> you wrongly claimed that dump_dictionary() in lavf/dump.c contained no special handling for language tags.</p> <p dir="auto">And your <a href="/FFmpeg/FFmpeg/issues/22236" class="ref-issue" rel="nofollow">#22236</a>, <a href="/FFmpeg/FFmpeg/issues/21079" class="ref-issue" rel="nofollow">#21079</a> and <a href="/FFmpeg/FFmpeg/issues/21187" class="ref-issue" rel="nofollow">#21187</a> also needed multiple rounds of review.</p> Sorry for not responding earlier; I no longer look at my mails that often since development moved to forgejo and therefore I have not seen the GA mail in time. I know that the decision has been made, but given that I have been asked to explain myself by both toots and michaelni I think I should do so.

I think that anyone who wants to be a maintainer needs to have a track record of good PRs, i.e. PRs that basically can be applied as-is, without multiple rounds of review and that do not cause regressions. Also the good PRs should be the clear majority of the PRs of the aspiring maintainer.

This is not so for toots: In #21787 you did not update FATE test results before submitting a PR; when you saw the runners falling, you then sent ever more updates that also did not pass instead of running FATE locally and updating the references. At one point (that commit is now 404) did you even intend to commit one of the test tool binaries.

Your chained ogg PR #20876 caused memleaks and needed to be reverted (as did your older "header packet skip" commit 574f634e49847e2225ee50013afebf0de03ef013).

#21813 and #21199 were based on misunderstandings of the underlying specifications and needed to be abandonded. The latter also would have introduced many more issues if applied (even if the understanding of the spec were correct).

In #20867 you wrongly claimed that dump_dictionary() in lavf/dump.c contained no special handling for language tags.

And your #22236, #21079 and #21187 also needed multiple rounds of review.

]]>
mkver 922730: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20961#issuecomment-31948 Sun, 15 Mar 2026 11:59:40 +0000
Niklas Haas created pull request FFmpeg/FFmpeg#22509 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509 22509#swscale/ops_backend: avoid UB (null pointer arithmetic)# Just use uintptr_t, it accomplishes the exact same thing while being defined
behavior.

Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas [email protected]

]]>
Niklas Haas 922658: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22509 Sun, 15 Mar 2026 11:22:28 +0000
mkver commented on pull request FFmpeg/FFmpeg#22236 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22236#issuecomment-31944 tests/fate/ogg-*.mak: fix dependency on $(FFMPEG) <p dir="auto"><a href="/toots" class="mention" rel="nofollow">@toots</a> wrote in <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22236#issuecomment-31876" class="ref-issue" rel="nofollow">#22236 (comment)</a>:</p> @toots wrote in #22236 (comment): ]]> mkver 922585: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22236#issuecomment-31944 Sun, 15 Mar 2026 10:10:57 +0000 Gyan Doshi pushed to master at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/482395f830a18686d23c12f783b7ea927c2f2bdb <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/482395f830a18686d23c12f783b7ea927c2f2bdb">482395f830a18686d23c12f783b7ea927c2f2bdb</a> avfilter/af_whisper: Add translate parameter 482395f830a18686d23c12f783b7ea927c2f2bdb avfilter/af_whisper: Add translate parameter]]> Gyan Doshi 922440: https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/482395f830a18686d23c12f783b7ea927c2f2bdb Sun, 15 Mar 2026 06:53:24 +0000 Gyan Doshi merged pull request FFmpeg/FFmpeg#22450 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22450 avfilter/af_whisper: Add translate parameter Gyan Doshi 922370: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22450 Sun, 15 Mar 2026 06:53:23 +0000 Zhao Zhili approved FFmpeg/FFmpeg#22485 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22485#issuecomment-31937 aarch64: Add Armv9.3-A GCS support <p dir="auto">LGTM, Learned something new today.</p> LGTM, Learned something new today. ]]> Zhao Zhili 922299: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22485#issuecomment-31937 Sun, 15 Mar 2026 03:36:55 +0000 Zhao Zhili pushed to release/8.1 at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/7df9a56f0d2bdbf45e7868f7b4e1f06a6fc01be1 <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/7df9a56f0d2bdbf45e7868f7b4e1f06a6fc01be1">7df9a56f0d2bdbf45e7868f7b4e1f06a6fc01be1</a> avformat/rtmpproto: fix listen_timeout conversion for special negative values 7df9a56f0d2bdbf45e7868f7b4e1f06a6fc01be1 avformat/rtmpproto: fix listen_timeout conversion for special negative values]]> Zhao Zhili 922228: https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/7df9a56f0d2bdbf45e7868f7b4e1f06a6fc01be1 Sun, 15 Mar 2026 03:15:41 +0000 Zhao Zhili merged pull request FFmpeg/FFmpeg#22506 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22506 [8.1] avformat/rtmpproto: fix listen_timeout conversion for special negative values Zhao Zhili 922158: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22506 Sun, 15 Mar 2026 03:15:40 +0000 Zhao Zhili approved FFmpeg/FFmpeg#22506 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22506#issuecomment-31934 [8.1] avformat/rtmpproto: fix listen_timeout conversion for special negative values Zhao Zhili 922088: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22506#issuecomment-31934 Sun, 15 Mar 2026 03:15:15 +0000 dana-feng commented on pull request FFmpeg/FFmpeg#22261 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31927 vf_mpdecimate: Fix keep option logic for keep > 0 <p dir="auto"><a href="/michaelni" class="mention" rel="nofollow">@michaelni</a> wrote in <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31043" class="ref-issue" rel="nofollow">#22261 (comment)</a>:</p> @michaelni wrote in #22261 (comment): ]]> dana-feng 921597: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31927 Sun, 15 Mar 2026 00:59:28 +0000 dana-feng commented on pull request FFmpeg/FFmpeg#22261 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31925 vf_mpdecimate: Fix keep option logic for keep > 0 <p dir="auto"><a href="/michaelni" class="mention" rel="nofollow">@michaelni</a> wrote in <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31047" class="ref-issue" rel="nofollow">#22261 (comment)</a>:</p> @michaelni wrote in #22261 (comment): ]]> dana-feng 921525: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31925 Sun, 15 Mar 2026 00:55:22 +0000 dana-feng commented on pull request FFmpeg/FFmpeg#22261 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31923 vf_mpdecimate: Fix keep option logic for keep > 0 <p dir="auto">added the av_frame_clone!</p> added the av_frame_clone! ]]> dana-feng 921453: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31923 Sun, 15 Mar 2026 00:53:20 +0000 dana-feng commented on pull request FFmpeg/FFmpeg#22261 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31922 vf_mpdecimate: Fix keep option logic for keep > 0 <p dir="auto">ah, removed this!</p> ah, removed this! ]]> dana-feng 921382: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31922 Sun, 15 Mar 2026 00:52:28 +0000 dana-feng commented on pull request FFmpeg/FFmpeg#22261 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31921 vf_mpdecimate: Fix keep option logic for keep > 0 <p dir="auto">got it, added the enums DECIMATE_DROP, DECIMATE_KEEP_UPDATE, and DECIMATE_KEEP_NO_UPDATE</p> got it, added the enums DECIMATE_DROP, DECIMATE_KEEP_UPDATE, and DECIMATE_KEEP_NO_UPDATE ]]> dana-feng 921311: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22261#issuecomment-31921 Sun, 15 Mar 2026 00:52:14 +0000 FFmpeg pushed to release/8.1 at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/compare/c9a5f7c6cae515e4ee5dc6839538cbff24b664f4...6f890cb10423aa5835f55f0e52e46b5cdbf5b2cc <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/6f890cb10423aa5835f55f0e52e46b5cdbf5b2cc">6f890cb10423aa5835f55f0e52e46b5cdbf5b2cc</a> avcodec/exr: Check input space before reverse_lut() <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/42692d0f571f335174049e06c855b20340d73e6d">42692d0f571f335174049e06c855b20340d73e6d</a> avformat/mpegts: remove JPEG-XS early return on invalid header_size <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/259ee609acd124456f1c6c442843168b111ab262">259ee609acd124456f1c6c442843168b111ab262</a> avcodec/aac/aacdec_usac_mps212: Introduce a temporary array for ff_aac_ec_data_dec() <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/b6a617c8bfc1b3b1a661013c93727a708b7e1b13">b6a617c8bfc1b3b1a661013c93727a708b7e1b13</a> avcodec/cbs_h266_syntax_template: Check tile_y <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/a5696b44a6f692118f5ebf6e420f0158971e9345">a5696b44a6f692118f5ebf6e420f0158971e9345</a> avcodec/h264_slice: reject slice_num &gt;= 0xFFFF 6f890cb10423aa5835f55f0e52e46b5cdbf5b2cc avcodec/exr: Check input space before reverse_lut() 42692d0f571f335174049e06c855b20340d73e6d avformat/mpegts: remove JPEG-XS early return on invalid header_size 259ee609acd124456f1c6c442843168b111ab262 avcodec/aac/aacdec_usac_mps212: Introduce a temporary array for ff_aac_ec_data_dec() b6a617c8bfc1b3b1a661013c93727a708b7e1b13 avcodec/cbs_h266_syntax_template: Check tile_y a5696b44a6f692118f5ebf6e420f0158971e9345 avcodec/h264_slice: reject slice_num >= 0xFFFF]]> FFmpeg 921239: https://code.ffmpeg.org/FFmpeg/FFmpeg/compare/c9a5f7c6cae515e4ee5dc6839538cbff24b664f4...6f890cb10423aa5835f55f0e52e46b5cdbf5b2cc Sun, 15 Mar 2026 00:23:47 +0000 James Almer merged pull request FFmpeg/FFmpeg#22470 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22470 avformat/movenc: add support for LCEVC track muxing James Almer 921169: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22470 Sun, 15 Mar 2026 00:08:33 +0000 FFmpeg pushed to master at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/compare/77ddfcfeb10206aac8d0b54badad25c249c8ab8f...539fc854e74dcaf2e7ffdfe7ec05431c37f226d0 <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/539fc854e74dcaf2e7ffdfe7ec05431c37f226d0">539fc854e74dcaf2e7ffdfe7ec05431c37f226d0</a> fftools/ffmpeg_mux_init: add support for LCEVC Stream Group muxing <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/9f9db1f673183a5f50a3bab928da6978bcf0257e">9f9db1f673183a5f50a3bab928da6978bcf0257e</a> avformat/options: add missing AVOption for AVStreamGroupLCEVC <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/0878ae59f900f74f32b1bc86fd425c0d64dae539">0878ae59f900f74f32b1bc86fd425c0d64dae539</a> avformat/movenc: add support for LCEVC track muxing 539fc854e74dcaf2e7ffdfe7ec05431c37f226d0 fftools/ffmpeg_mux_init: add support for LCEVC Stream Group muxing 9f9db1f673183a5f50a3bab928da6978bcf0257e avformat/options: add missing AVOption for AVStreamGroupLCEVC 0878ae59f900f74f32b1bc86fd425c0d64dae539 avformat/movenc: add support for LCEVC track muxing]]> FFmpeg 921098: https://code.ffmpeg.org/FFmpeg/FFmpeg/compare/77ddfcfeb10206aac8d0b54badad25c249c8ab8f...539fc854e74dcaf2e7ffdfe7ec05431c37f226d0 Sun, 15 Mar 2026 00:08:30 +0000 FFmpeg pushed to master at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/77ddfcfeb10206aac8d0b54badad25c249c8ab8f <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/77ddfcfeb10206aac8d0b54badad25c249c8ab8f">77ddfcfeb10206aac8d0b54badad25c249c8ab8f</a> Changelog: move an entry wrongly put in the 8.1 section to next 77ddfcfeb10206aac8d0b54badad25c249c8ab8f Changelog: move an entry wrongly put in the 8.1 section to next]]> FFmpeg 921027: https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/77ddfcfeb10206aac8d0b54badad25c249c8ab8f Sat, 14 Mar 2026 23:45:09 +0000 michaelni merged pull request FFmpeg/FFmpeg#22490 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22490 avcodec/exr: Check input space before reverse_lut() michaelni 920956: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22490 Sat, 14 Mar 2026 22:26:49 +0000 FFmpeg pushed to master at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/70286d59f1bd413009c28dea22ab69649fa87490 <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/70286d59f1bd413009c28dea22ab69649fa87490">70286d59f1bd413009c28dea22ab69649fa87490</a> avcodec/exr: Check input space before reverse_lut() 70286d59f1bd413009c28dea22ab69649fa87490 avcodec/exr: Check input space before reverse_lut()]]> FFmpeg 920885: https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/70286d59f1bd413009c28dea22ab69649fa87490 Sat, 14 Mar 2026 22:26:47 +0000 Pierre-Anthony Lemieux pushed to master at FFmpeg/FFmpeg https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/dfc5d176c9fd2ce5bf1b40603c8fe2f87f13d0a9 <a href="https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/dfc5d176c9fd2ce5bf1b40603c8fe2f87f13d0a9">dfc5d176c9fd2ce5bf1b40603c8fe2f87f13d0a9</a> fuzzer: improve documentation dfc5d176c9fd2ce5bf1b40603c8fe2f87f13d0a9 fuzzer: improve documentation]]> Pierre-Anthony Lemieux 920811: https://code.ffmpeg.org/FFmpeg/FFmpeg/commit/dfc5d176c9fd2ce5bf1b40603c8fe2f87f13d0a9 Sat, 14 Mar 2026 21:37:03 +0000 Pierre-Anthony Lemieux merged pull request FFmpeg/FFmpeg#22460 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22460 fuzzer: improve decoder fuzzer target documentation Pierre-Anthony Lemieux 920740: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22460 Sat, 14 Mar 2026 21:37:02 +0000 michaelni closed pull request FFmpeg/FFmpeg#22256 https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22256#issuecomment-31907 avcodec/aac/aacdec_usac: clip sfo michaelni 920669: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22256#issuecomment-31907 Sat, 14 Mar 2026 21:22:55 +0000