Repository navigation
[FIX] Honor 23.976 fps when importing SCC - #2452
Open
harshalsp07 wants to merge 2 commits into
Open
harshalsp07 wants to merge 2 commits into
harshalsp07 wants to merge 2 commits into
Conversation
Contributor
Author
|
sample 31 is byte-identical on master and this branch across 3 runs each. the approved first timestamp differs on both. could you rerun it? |
cfsmp3
requested changes
Nov 11, 2026
cfsmp3
left a comment
cfsmp3
left a comment
Contributor
There was a problem hiding this comment.
This is Carlos' claude doing the first pass. Before Carlos takes a look, please address these issues.
The fix is correct, and its timing model (wall seconds plus frames × 1001/24) matches our SCC writer in ccx_encoders_scc.c, so import and export round-trip. One thing left:
tests/samples/scc_23_976.sccis added but nothing references it. Either drop it, or use it in a test (for example,include_str!it in the newtest_parse_scc_uses_23_976_framerate_from_ffiinstead of the inline string).
The sample file was unreferenced. include_str! it in the FFI framerate test and assert the parsed 608 byte pairs so both timing and content are covered.
Collaborator
CCExtractor CI platform finished running the test files on linux. 172/238 tests matched the approved output:
66 tests do not match the approved output. That is the pass/fail verdict. Whether this branch caused it is a separate question, answered below.
Compared with the tip of master — test 9984, commit 24d1e0a:
Compared with the commit this branch was cut from — test 9878, commit 185fe3f:
No test changes behaviour relative to the tip of master: every failure above fails there too, byte for byte. The approved output for those tests is out of date, which is a baseline to review rather than a regression in this branch. |
Collaborator
CCExtractor CI platform finished running the test files on windows. 130/238 tests matched the approved output:
108 tests do not match the approved output. That is the pass/fail verdict. Whether this branch caused it is a separate question, answered below.
Compared with the tip of master — test 9985, commit 24d1e0a:
Pass there, fail here:
Fail on both but produce different output — behaviour moved even though the verdict did not:
Compared with the commit this branch was cut from — test 9693, commit af4b28d:
Fail on both but produce different output — behaviour moved even though the verdict did not:
Fail there, pass here:
This branch changes the behaviour of 52 test(s) relative to the tip of master. Those are the ones worth looking at; anything else in the list fails the same way on both sides. |
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
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.
In raising this pull request, I confirm the following (please check boxes):
Reason for this PR:
Sanity check:
--scc-framerate 23.976and its23.98alias are accepted by the CLI, but SCC import silently uses 29.97 fps. Both spellings become integer value 4;SccFrameRate::from_int()doesn't handle it and falls back to the default.Add the missing decoder rate and use
1001 / 24milliseconds per frame. The other rate mappings stay the same.Reproduce from the repository root with the included 90-byte sample:
The sample contains
HIbetween00:00:01:12and00:00:03:12. On masterd7eb132e, the SRT timing is00:00:01,401 --> 00:00:03,399. With this change it is00:00:01,501 --> 00:00:03,499. Repeating with23.98gives the same corrected timing.Tested on Linux:
cargo test --lockedinsrc/rust: 427 passed. The new SCC parsing test fails without the fix.cargo fmt --all -- --checkandcargo clippy -- -D warningspass.This fixes SCC input parsing. #2145 and #2406 concern CEA-708 SCC output timing. Windows/macOS and the external sample platform haven't been tested.
Follow-up local check on sample
b22260d065(platform regression 31): verified its SHA-256 and ran--autoprogram --out=srt --latin1three times on master's source and three times on this branch, with the same compiler and dependencies. All six runs exit 0 and produce identical 161,874-byte SRT files (SHA-256f5758a49bbe5fb46536c8a94b598f585ac0f92ac18bd451d862a8b3b57ecf561). Both builds start the first caption at00:00:04,838, while the platform's approved output expects00:00:02,635, so this does not establish an approved-output pass. The PR-specific difference reported by the hosted run was not reproduced locally; the platform test still needs a rerun.CI note: The initial Linux build stops before compilation because
libgpac-devhas no installation candidate on Ubuntu 26.04; #2428 addresses the runner change. The C format check also flagsccx_decoders_isdb.candmp4.c, which are unchanged from master.