使用FFmpeg保存UDP组播流至文件时遇Invalid UE golomb code报错

Let's break down your problem and walk through practical fixes step by step. The Invalid UE golomb code error points to corrupted H.264/H.265 bitstream data (UE golomb codes are part of the video codec's entropy encoding), and this corruption is almost certainly what's causing FFmpeg to freeze after a few minutes. The unknown SEI type 229 warning is less critical—it just means the stream has a custom SEI (Supplemental Enhancement Information) frame that FFmpeg doesn't recognize, but this shouldn't be the root cause of the freeze.

Potential Fixes #

1. Mitigate UDP Multicast Packet Loss & Corruption #

UDP multicast is inherently prone to packet loss and out-of-order delivery, which can corrupt the video stream. Add these robustness parameters to your UDP input:

ffmpeg -i "udp://238.0.0.1:1234?buffer_size=10485760&reorder_queue_size=1024&max_delay=500000" -c copy -f segment -segment_time 900 -segment_atclocktime 1 -reset_timestamps 1 -strftime 1 $datestamp.mp4 -report

• buffer_size=10485760: Boosts the UDP receive buffer to 10MB, reducing the chance of buffer overflow and lost packets.
• reorder_queue_size=1024: Lets FFmpeg reorder out-of-order packets before processing, which addresses common multicast delivery quirks.
• max_delay=500000: Sets a 500ms maximum delay to wait for out-of-order packets before proceeding, balancing latency and stream integrity.

2. Force FFmpeg to Ignore Bitstream Errors #

Tell FFmpeg to skip corrupted frames instead of freezing up by adding error-handling flags:

ffmpeg -err_detect ignore_err -fflags +genpts+igndts -i "udp://238.0.0.1:1234" -c copy -f segment -segment_time 900 -segment_atclocktime 1 -reset_timestamps 1 -strftime 1 $datestamp.mp4 -report

• -err_detect ignore_err: Makes FFmpeg ignore bitstream errors and continue processing rather than halting.
• -fflags +genpts+igndts: Generates new presentation timestamps (PTS) if the stream's timestamps are missing or corrupted, preventing timing-related freezes.

3. Re-encode the Stream Instead of Copying #

If the source stream has persistent corruption, using -c copy (stream copy) will propagate the corruption and cause FFmpeg to choke. Re-encoding lets FFmpeg discard invalid frames entirely:

ffmpeg -i "udp://238.0.0.1:1234" -c:v libx264 -crf 23 -c:a aac -b:a 128k -f segment -segment_time 900 -segment_atclocktime 1 -reset_timestamps 1 -strftime 1 $datestamp.mp4 -report

• -c:v libx264: Re-encodes the video with the standard H.264 codec, which automatically skips unreadable frames.
• -crf 23: Sets a balanced video quality level (adjust between 18-28 for better/worse quality as needed).
• -c:a aac -b:a 128k: Re-encodes audio to AAC for broad compatibility.

4. Verify the Source Stream & Network #

• Test the multicast stream with another player (like VLC) to see if it also glitches or freezes—this will confirm if the issue is with the source or your FFmpeg setup.
• Use tcpdump to check for excessive packet loss or malformed traffic:

  tcpdump -i eth0 udp port 1234 and host 238.0.0.1
  

内容的提问来源于stack exchange，提问作者Stuart