[PR #1857] [MERGED] fix(avc): Handle streams that don't start with NAL start codes #2623

New Issue

claunia · 2026-01-29T17:23:08Z

claunia commented

2026-01-29 17:23:08 +00:00

📋 Pull Request Information

Original PR: https://github.com/CCExtractor/ccextractor/pull/1857
Author: @cfsmp3
Created: 12/20/2025
Status: ✅ Merged
Merged: 12/20/2025
Merged by: @cfsmp3

Base: master ← Head: fix/issue-1626-broken-avc-stream

📝 Commits (1)

73db3a2 fix(avc): Handle streams that don't start with NAL start codes (#1626)

📊 Changes

1 file changed (+161 additions, -46 deletions)

View changed files

📝 src/rust/src/avc/core.rs (+161 -46)

📄 Description

Summary

Fixes #1626 - [BUG] [EIA-608] Can't extract subtitles in TS file with "Broken AVC stream" error
Fixes #1582 - [BUG] Error: Broken AVC stream on Twitch.tv stream file
Fixes #1597 - [BUG] Subtitles not extracted from Twitch stream

Problem

The AVC parser would fail with "Leading bytes are non-zero" error when processing HLS/Twitch stream segments. These segments often start mid-stream without proper NAL unit headers at the beginning.

Root cause: When process_avc encountered non-zero leading bytes, it returned an error with 0 bytes processed. The C code would not remove any bytes from the buffer, causing subsequent data to accumulate with the corrupt beginning, leading to infinite errors and no caption extraction.

Solution

Add find_nal_start_code() function to search for valid NAL start codes (3-byte or 4-byte)
If buffer doesn't start with 0x00 0x00, search for first NAL start code
Skip garbage data before first valid NAL unit
Return full buffer length when no NAL found (clears the buffer, prevents accumulation)
Change forbidden_zero_bit error from fatal to skip-and-continue

Test plan

Tested with all 6 Twitch HLS sample files from #1626 - all now process correctly
Tested with sample from #1582 - processes to 100% with no AVC errors
Captions extracted successfully (see output below)
All 299 Rust tests pass (292 existing + 7 new unit tests for find_nal_start_code)

Before fix:

Error in process_avc: BrokenStream("Leading bytes are non-zero")
Error in process_avc: BrokenStream("Leading bytes are non-zero")
... (repeated for every PES packet)

After fix:

Number of NAL_type_7: 5
Total frames time:    00:00:10:000  (600 frames at 60.00fps)
Done, processing time = 0 seconds

Extracted captions:

1
00:00:04,066 --> 00:00:06,332
congratulations to papa plot and
their community for winning ■k71
million in a TV contest through 

2
00:00:06,334 --> 00:00:09,998
their community for winning ■k71
million in a TV contest through 
Twitch. I mean, just wild. Um

🤖 Generated with Claude Code

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/CCExtractor/ccextractor/pull/1857 **Author:** [@cfsmp3](https://github.com/cfsmp3) **Created:** 12/20/2025 **Status:** ✅ Merged **Merged:** 12/20/2025 **Merged by:** [@cfsmp3](https://github.com/cfsmp3) **Base:** `master` ← **Head:** `fix/issue-1626-broken-avc-stream` --- ### 📝 Commits (1) - [`73db3a2`](https://github.com/CCExtractor/ccextractor/commit/73db3a2c39f03006d89fc5d2f2a19fb27624e3f3) fix(avc): Handle streams that don't start with NAL start codes (#1626) ### 📊 Changes **1 file changed** (+161 additions, -46 deletions) <details> <summary>View changed files</summary> 📝 `src/rust/src/avc/core.rs` (+161 -46) </details> ### 📄 Description ## Summary - Fixes #1626 - [BUG] [EIA-608] Can't extract subtitles in TS file with "Broken AVC stream" error - Fixes #1582 - [BUG] Error: Broken AVC stream on Twitch.tv stream file - Fixes #1597 - [BUG] Subtitles not extracted from Twitch stream ## Problem The AVC parser would fail with "Leading bytes are non-zero" error when processing HLS/Twitch stream segments. These segments often start mid-stream without proper NAL unit headers at the beginning. **Root cause:** When `process_avc` encountered non-zero leading bytes, it returned an error with 0 bytes processed. The C code would not remove any bytes from the buffer, causing subsequent data to accumulate with the corrupt beginning, leading to infinite errors and no caption extraction. ## Solution - Add `find_nal_start_code()` function to search for valid NAL start codes (3-byte or 4-byte) - If buffer doesn't start with `0x00 0x00`, search for first NAL start code - Skip garbage data before first valid NAL unit - Return full buffer length when no NAL found (clears the buffer, prevents accumulation) - Change forbidden_zero_bit error from fatal to skip-and-continue ## Test plan - [x] Tested with all 6 Twitch HLS sample files from #1626 - all now process correctly - [x] Tested with sample from #1582 - processes to 100% with no AVC errors - [x] Captions extracted successfully (see output below) - [x] All 299 Rust tests pass (292 existing + 7 new unit tests for `find_nal_start_code`) **Before fix:** ``` Error in process_avc: BrokenStream("Leading bytes are non-zero") Error in process_avc: BrokenStream("Leading bytes are non-zero") ... (repeated for every PES packet) ``` **After fix:** ``` Number of NAL_type_7: 5 Total frames time: 00:00:10:000 (600 frames at 60.00fps) Done, processing time = 0 seconds ``` **Extracted captions:** ``` 1 00:00:04,066 --> 00:00:06,332 congratulations to papa plot and their community for winning ■k71 million in a TV contest through 2 00:00:06,334 --> 00:00:09,998 their community for winning ■k71 million in a TV contest through Twitch. I mean, just wild. Um ``` 🤖 Generated with [Claude Code](https://claude.com/claude-code) --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

claunia added the pull-request label 2026-01-29 17:23:08 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: starred/ccextractor#2623