Add fate tests for palette act and field #291

QSXW · 2025-06-03T18:37:40Z

The fix for DUAL_CHROMA_TREE coding unit was lost in our rebased commits. Please help review 0f9b47a

nuomi2021 · 2025-06-07T14:44:44Z

The fix for DUAL_CHROMA_TREE coding unit was lost in our rebased commits. Please help review 0f9b47a

thank you for the patch

For which specific clip are you seeing the issue? All ACT clips pass locally with the upstream code.

like:
ACTPIC_A_Huawei_3.bit ACTPIC_B_Huawei_3.bit ACTPIC_C_Huawei_3.bit ACT_A_Kwai_3.bit ACT_B_Kwai_3.bit

QSXW · 2025-06-08T14:18:27Z

The fix for DUAL_CHROMA_TREE coding unit was lost in our rebased commits. Please help review 0f9b47a

thank you for the patch

For which specific clip are you seeing the issue? All ACT clips pass locally with the upstream code.

like: ACTPIC_A_Huawei_3.bit ACTPIC_B_Huawei_3.bit ACTPIC_C_Huawei_3.bit ACT_A_Kwai_3.bit ACT_B_Kwai_3.bit

It's the one encoded by myself. See #224 (comment)

We should use the width and height of the start component. When the cu is DUAL_CHROMA_TREE, the start component is the first transform block, so we need to use the tb width and height.

nuomi2021 · 2025-06-09T12:54:22Z

It's the one encoded by myself. See #224 (comment)

We should use the width and height of the start component. When the cu is DUAL_CHROMA_TREE, the start component is the first transform block, so we need to use the tb width and height.

could you share the DUAL_CHROMA_TREE clip with me?
Thank you

nuomi2021 · 2025-06-09T12:55:19Z

It's better to remove the company name from the clip, just like the other FATE clips.

QSXW · 2025-06-09T13:20:20Z

It's the one encoded by myself. See #224 (comment)
We should use the width and height of the start component. When the cu is DUAL_CHROMA_TREE, the start component is the first transform block, so we need to use the tb width and height.

could you share the DUAL_CHROMA_TREE clip with me? Thank you

Sure. https://github.com/user-attachments/assets/91fce515-47bf-45e6-bc32-4473eba2260f

To upload the clip, I changed the file extension from bit to mp4. You need to change it back.

Open the link and use Ctrl + S to download the clip.

nuomi2021 · 2025-06-12T14:32:21Z

libavcodec/vvc/ctu.c

        const int scale    = ff_vvc_palette_derive_scale(lc, tu, tb);
-        const int hs       = sps->hshift[c];
-        const int vs       = sps->vshift[c];
+        const int hs       = tu->tbs[0].c_idx ? 0 : sps->hshift[c_idx];


How about defining a start_index? Then this can be simplified to sps->hshift[c_idx] - sps->hshift[start_index], which makes it more readable.

nuomi2021 · 2025-06-12T14:36:46Z

libavcodec/h274.c

+#if HAVE_BIGENDIAN
+        if (c->type != HASH_CHECKSUM) {
+            if (ps && !c->buf) {
+                av_fast_malloc(&c->buf, &c->buf_size,


do we need to free this?

Yes update.

nuomi2021 · 2025-06-12T14:40:09Z

libavcodec/h274.c

+#if HAVE_BIGENDIAN
+        if (c->type != HASH_CHECKSUM) {
+            if (ps && !c->buf) {
+                av_fast_malloc(&c->buf, &c->buf_size,


what if the frame resolution or ps changed?

Makes find output consistent. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

QSXW · 2025-08-04T22:24:03Z

@nuomi2021 Let's get this to mainline. Should I send it to the mailing list?

Fixes use-of-uninitialized-value when bp.len == 0. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

Commit f566032 added frame validation. Since then this decoder has been failing validation of sample rate value. Found by OSS-Fuzz. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

Printing dummy logs during fuzzing can significantly slow the process and blow the size of logs, making them both unredable and huge. Keep the loggging commented-out for easy local restore if needed. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

The check to return on EOF should not be inside a block that will not be entered after reaching EOF. Should fix "libavcodec/bytestream.h:144:27: runtime error: applying zero offset to null pointer". Signed-off-by: James Almer <jamrial@gmail.com>

Add allow extension name cmfa and cmfv test, this testcase only cover fragment mp4 named cmfa. ticket description in ticket/11526

This makes it possible to apply Adobe .cube files to inputs.

NVIDIA's support for it is a disaster. Of no benefit to other vendors. NVIDIA are working on fixing it, but it may take time.

This makes left_bits return useful data rather than overflowing, and also saves some 64-bit integer operations, which is still always a plus sadly.

…eaders

Previously, it was assumed that `drmFormatModifierPlaneCount` was one for all modifiers when exporting, which is not always the case, in particular for AMD GPUs and maybe others. Fetch the number of memory planes and fill the structs appropriately in this situation. The encoded stream is still bad in the case whre modifers are involved, but I think this patch still stands on its own and I suspect that may be a driver bug. A potential improvement that could be make is to cache the format information, so we can avoid the two GetPhysicalDeviceFormatProperties2 calls for each export, as well as the allocation. I doubt this is very expensive, but seemed worth noting. v2 changes: query the format properties with the test image created in `vulkan_frames_init` to avoid allocating space for the query during export Signed-off-by: Russell Greene <russellgreene8@gmail.com>

There was implicit assumption that the $TMPC file is empty when doing --cpu=host checks. This breaks if any check is done before that. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

This avoids adding flags that cl.exe doesn't understand. Fixes cases where external libraries pkg-config file adds `-L` to the cflags, strip it before passing to cl.exe. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

This is important for feature checking to work correctly. It can happen that an unrecognized flag passes the compile test with only a warning, while failing in preprocessor-only check with an error. This causes all test_cpp calls to fail and silently produces arguably broken MSVC builds. Also, all check_* functions don't work as expected, because they assume the check passed, even though there was a warning. Additionally, this brings the behavior in line with GCC/Clang based builds, failing early on unrecognized flags instead of silently continuing with warnings in the log. The /options:strict option is available starting in Visual Studio 2022 version 17.0. Because of that, we cannot use check_cflags alone, as it would add this flag for older MSVC versions and produce warnings. So, we need to manually perform a version check. A bit of a chicken and egg problem. Perform the version check before adding extra flags from the user to ensure we don't silently fail the preprocessor check due to invalid flags on older MSVC versions. Note that behavior differs depending on whether we are compiling or only preprocessing. This fixes silent different between handling: `cl.exe -P foo c.c` c1: fatal error C1083: Cannot open source file: 'foo': No such file or directory `cl.exe -c foo c.c` cl : Command line warning D9024 : unrecognized source file type 'foo', object file assumed Where -P fails, while -c throws warnings only. Of course `foo` is completely bogus here, but depends on the flags or configuration this may be unsupported argument. Or even some converted path from MSYS when run inside it. The objective is to always error out instead of silently hiding this. Use check_cflags even after the _MSC_FULL_VER check, for non-MSVC compilers. For example Clang-CL impersonate MSVC, but does not support -options:strict flag currently. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

…hitectures This changes configure to stop disabling -ftree-vectorize on GCC versions 13 and newer, on major architectures. Background: - Original `-fno-tree-vectorize` was added in 2009 in commit 973859f to avoid compiler errors. - Re-enabled in 2016 in commit cb8646a but caused failures due to inline CABAC assembly issues and was disabled again in fd6dbc5. - Commit 182663a in 2023 fixed the inline CABAC assembly issues. - Recent versions of GCC, in particular 13 and newer, seem to generally work reliably with respect to vectorization, although bugs have been observed on Loongarch. Cautiously allow the GCC default of having vectorization enabled, on major architectures where we expect to see enough testing. If further issues are observed, they should be reported and noted here in configure, so the workarounds can be scoped and version limited.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

… it's not already in place If the first Sample references the first stsd entry, then setting it here is redundant. Signed-off-by: James Almer <jamrial@gmail.com>

Introduced by 307983b Use the following command line to reproduce the issue: ./configure --toolchain=msvc --disable-asm --enable-ffmpeg \ --disable-everything --enable-decoder=vvc --enable-parser=vvc \ --enable-demuxer='vvc,mpegts' --enable-protocol='file,pipe' \ --enable-encoder='rawvideo,wrapped_avframe' \ --enable-muxer='rawvideo,md5,null' Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

This commit fixed decoding the DUAL_TREE_CHROMA palette coding unit Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

…rrect It's helpful for developers and the same as the hevcdec. Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

This commit added 10b422_L_5 for testing palette mode. Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

nuomi2021 · 2025-08-08T03:43:23Z

+1. Better to group patches with related functions and send them to the mailing list separately.

QSXW requested a review from nuomi2021 June 3, 2025 18:46

QSXW changed the base branch from up to main June 3, 2025 19:14

QSXW changed the base branch from main to up June 3, 2025 19:14

QSXW force-pushed the fate/add_plt_act_field branch 2 times, most recently from d5d5f5f to 44b6d39 Compare June 8, 2025 16:12

QSXW force-pushed the fate/add_plt_act_field branch from 44b6d39 to 8ebea9f Compare June 10, 2025 19:48

nuomi2021 reviewed Jun 12, 2025

View reviewed changes

forgejo/workflows: sort file list for cache hash

efeddd1

Makes find output consistent. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

QSXW force-pushed the fate/add_plt_act_field branch from 8ebea9f to 4a0437d Compare August 4, 2025 22:20

kasper93 and others added 13 commits August 5, 2025 03:27

avformat/ffmetadec: don't compare undefined string

96ab006

Fixes use-of-uninitialized-value when bp.len == 0. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

avcodec/dsicinaudio: set missing sample rate

4718974

Commit f566032 added frame validation. Since then this decoder has been failing validation of sample rate value. Found by OSS-Fuzz. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

avcodec/evrcdec: set missing sample rate

8f0c146

Commit f566032 added frame validation. Since then this decoder has been failing validation of sample rate value. Found by OSS-Fuzz. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

avcodec/qcelpdec: set missing sample rate

7f5db34

Commit f566032 added frame validation. Since then this decoder has been failing validation of sample rate value. Found by OSS-Fuzz. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

avcodec/qcelpdec: set missing sample rate

8b5db38

Commit f566032 added frame validation. Since then this decoder has been failing validation of sample rate value. Found by OSS-Fuzz. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

avformat/iff: fix EOF check

8be539b

The check to return on EOF should not be inside a block that will not be entered after reaching EOF. Should fix "libavcodec/bytestream.h:144:27: runtime error: applying zero offset to null pointer". Signed-off-by: James Almer <jamrial@gmail.com>

tests/fate/hlsenc: add testcase of hls fragment mp4 named cmfa

7838648

Add allow extension name cmfa and cmfv test, this testcase only cover fragment mp4 named cmfa. ticket description in ticket/11526

vf_libplacebo: add support for specifying a LUT for the input

d328801

This makes it possible to apply Adobe .cube files to inputs.

hwcontext_vulkan: temporarily disable host_image_copy

53826f1

NVIDIA's support for it is a disaster. Of no benefit to other vendors. NVIDIA are working on fixing it, but it may take time.

hwcontext_vulkan: enable uniformBufferStandardLayout

50756b8

lavc/vulkan/common: sign-ify lengths

2c3315b

This makes left_bits return useful data rather than overflowing, and also saves some 64-bit integer operations, which is still always a plus sadly.

vulkan_encode_h264/5: Fix uninitialized return value in write_extra_h…

cc126af

…eaders

russelltg and others added 10 commits August 5, 2025 23:53

configure: don't use dangling temp file state

5b3d6c0

There was implicit assumption that the $TMPC file is empty when doing --cpu=host checks. This breaks if any check is done before that. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

configure: filter link flags separately for MSVC

50408c5

This avoids adding flags that cl.exe doesn't understand. Fixes cases where external libraries pkg-config file adds `-L` to the cflags, strip it before passing to cl.exe. Signed-off-by: Kacper Michajłow <kasper93@gmail.com>

avcodec/motion_est: Remove double protection from init_ref()

3b0ab68

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

.forgejo/CODEOWNERS: add myself to a few things

31bad42

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

forgejo/CODEOWNERS: add myself to docs

399c766

avformat/mov: set primary extradata based on the first Sample only if…

2681eed

… it's not already in place If the first Sample references the first stsd entry, then setting it here is redundant. Signed-off-by: James Almer <jamrial@gmail.com>

QSXW force-pushed the fate/add_plt_act_field branch from 4a0437d to bae25c8 Compare August 6, 2025 18:53

QSXW and others added 7 commits August 7, 2025 03:51

avcodec/vvc/ctu: should use the width and height of the start component

ecd1e39

This commit fixed decoding the DUAL_TREE_CHROMA palette coding unit Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

avcodec/h274: fix hash verification on BE

81c139d

Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

avcodec/vvc/dec: fix typo and also output log when the checksum is co…

c3680ee

…rrect It's helpful for developers and the same as the hevcdec. Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

fate/vvc: add vvc-conformance-10b422_L_5

49a60c3

This commit added 10b422_L_5 for testing palette mode. Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

fate/vvc: add vvc-conformance-ACT_A_3

5f49b88

Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

fate/vvc: add vvc-conformance-FIELD_A_4

8517e9e

Signed-off-by: Wu Jianhua <toqsxw@outlook.com>

add github workflow

ed06eda

QSXW force-pushed the fate/add_plt_act_field branch from bae25c8 to ed06eda Compare August 6, 2025 19:51

Add fate tests for palette act and field #291

Are you sure you want to change the base?

Add fate tests for palette act and field #291

Uh oh!

Conversation

QSXW commented Jun 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nuomi2021 commented Jun 7, 2025

Uh oh!

QSXW commented Jun 8, 2025

Uh oh!

nuomi2021 commented Jun 9, 2025

Uh oh!

nuomi2021 commented Jun 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

QSXW commented Jun 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nuomi2021 Jun 12, 2025

Choose a reason for hiding this comment

Uh oh!

QSXW Aug 4, 2025

Choose a reason for hiding this comment

Uh oh!

nuomi2021 Jun 12, 2025

Choose a reason for hiding this comment

Uh oh!

QSXW Aug 4, 2025

Choose a reason for hiding this comment

Uh oh!

nuomi2021 Jun 12, 2025

Choose a reason for hiding this comment

Uh oh!

QSXW Aug 4, 2025

Choose a reason for hiding this comment

Uh oh!

QSXW commented Aug 4, 2025

Uh oh!

nuomi2021 commented Aug 8, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

11 participants

QSXW commented Jun 3, 2025 •

edited

Loading

nuomi2021 commented Jun 9, 2025 •

edited

Loading

QSXW commented Jun 9, 2025 •

edited

Loading