Add 12-bit profile (Main_444C_12_IP2) - #5112
Merged
urvangjoshi merged 9 commits intoAug 4, 2026
Merged
Conversation
xinzhao-apple
marked this pull request as draft
July 14, 2026 22:28
wantehchang
reviewed
Jul 14, 2026
wantehchang
left a comment
wantehchang
left a comment
Member
There was a problem hiding this comment.
Xin: Thank you for writing this PR. I suggest some simple changes.
xinzhao-apple
force-pushed
the
contrib/avm-12bit-fix
branch
5 times, most recently
from
July 16, 2026 20:28
9527f14 to
7493fd0
Compare
xinzhao-apple
marked this pull request as ready for review
July 28, 2026 20:08
xinzhao-apple
requested review from
Krishna-Rapaka,
leolzhao,
urvangjoshi and
wantehchang
July 28, 2026 20:31
xinzhao-apple
changed the title
xinzhao-apple
force-pushed
the
contrib/avm-12bit-fix
branch
from
July 29, 2026 16:51
0c15375 to
978195d
Compare
wantehchang
approved these changes
Jul 29, 2026
wantehchang
left a comment
wantehchang
left a comment
Member
There was a problem hiding this comment.
Xin: LGTM. Thanks! Note that I did not compare this PR with the corresponding spec PR (https://github.com/AOMediaCodec/av2-spec-internal/pull/776),
xinzhao-apple
force-pushed
the
contrib/avm-12bit-fix
branch
from
July 30, 2026 16:29
cfa60fb to
075f76c
Compare
wantehchang
approved these changes
Jul 30, 2026
wantehchang
reviewed
Jul 30, 2026
xinzhao-apple
force-pushed
the
contrib/avm-12bit-fix
branch
from
July 30, 2026 21:11
075f76c to
c194f43
Compare
wantehchang
approved these changes
Jul 31, 2026
wantehchang
approved these changes
Jul 31, 2026
urvangjoshi
reviewed
Aug 3, 2026
urvangjoshi
approved these changes
Aug 3, 2026
urvangjoshi
left a comment
urvangjoshi
left a comment
Contributor
There was a problem hiding this comment.
One comment related to CI job -- otherwise looks good.
xinzhao-apple
force-pushed
the
contrib/avm-12bit-fix
branch
from
August 3, 2026 21:57
08e7b0d to
a052daa
Compare
wantehchang
approved these changes
Aug 3, 2026
Flip the default for the 12-bit test-only profile from 0 to 1 so that 12-bit encode/decode is built in by default. Builds that explicitly opt out can still pass -DCONFIG_TESTONLY_12BIT_SUPPORT=0.
- Rename CONFIG_TESTONLY_12BIT_SUPPORT to CONFIG_12BIT_PROFILE and the profile enum to MAIN_4xx_12_IP2 (spec Main_4xx_12_IP2) - Accept 8/10/12-bit and all chroma formats for profile 5 - Set max mlayer count for profile 5 to 3 (IOP2 limit) - Give profile 5 its own PicSize/Bitrate factor row (36 / 3.0) per Table A.2 - Update Annex A profile/config tables and comments
Per the VCWG decision on 12-bit profile naming, rename the 12-bit profile from MAIN_4xx_12_IP2 to MAIN_444C_12_IP2 (spec Main_444C_12_IP2), where "C" stands for Comprehensive. - Rename the profile enum MAIN_4xx_12_IP2 to MAIN_444C_12_IP2 - Rename the multi-sequence configuration label C_MAIN_4xx_12 to C_MAIN_444C_12 to keep it aligned with the profile label - Update the Annex A profile/configuration tables and comments - Update the CONFIG_12BIT_PROFILE cmake help string Naming only; enum values (profile idc 5, configuration id 3) and all profile constraints and level factors are unchanged.
The spec defines MAXQ_12_BITS (MAXQ_8_BITS + 4 * MAXQ_OFFSET) as the maximum quantizer for bit depth 12, separately from MAXQ_BITS. No functional change: all values are identical before and after.
The function returns a row index into the PicSize/Bitrate factor tables, not the spec's ProfileScalingFactor from Table A.2. The two coincide for profiles 0..4, but MAIN_444C_12_IP2 (idc 5) shares ProfileScalingFactor 2 with MAIN_444_10_IP1 (idc 4) while needing different factors (36 / 3.0 vs 30 / 2.5), so it gets its own row index. Rename the function to get_profile_factor_table_row_index and the callers' locals to profile_factor_row, so the name states what the value is instead of relying on a comment to disclaim the old name. Trim the now-redundant sentence from the comment and replace the stale one-line description in annexA.h with a pointer to the full explanation. No functional change.
urvangjoshi
force-pushed
the
contrib/avm-12bit-fix
branch
from
August 3, 2026 22:31
a052daa to
0db1923
Compare
Contributor
|
@xinzhao-apple @wantehchang : the PR contains 9 commits. When merging, do we want Github to squash them into a single commit? Or keep 9 separate commits? |
Member
|
Urvang: Please wait for Xin's reply. I think the 9 commits shoud be squashed into a single commit when you merge this PR. |
Contributor
Author
Hi @urvangjoshi I think we should squash them into a single commit. Thanks. |
urvangjoshi
enabled auto-merge (squash)
August 4, 2026 00:18
urvangjoshi
disabled auto-merge
August 4, 2026 23:03
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR enables the 12-bit profile by default and brings its implementation into conformance with AV2 Spec Draft with 12-bit support: https://github.com/AOMediaCodec/av2-spec-internal/pull/776.