5

I have a series of videos that I'm converting from .mov to .ts and then create an HLS playlist for. I'm able to figure out the ending pts for both the audio and video streams of any given video and am apply that ending (cumulative) offset when converting later videos in the sequence. For instance:

ffmpeg -y -i 1.mov \
  -filter:a "asetpts=PTS-STARTPTS+367534" \
  -filter:v "setpts=PTS-STARTPTS+363000" \
  -codec:v libx264 -crf 18 -preset veryfast \
  -acodec aac -muxdelay 0 1.ts

This works but requires a decent amount of CPU. I'd like to be able to ideally copy the video/audio streams. Is there any way to apply a uniform pts offset for the audio/video streams of a .ts file without re-encoding the whole thing?

Vinay
  • 6,204
  • 6
  • 38
  • 55

1 Answers1

4

The notional way to do this is

ffmpeg -y -i 1.mov -c copy -output_ts_offset 4.54 -muxdelay 0 -muxpreload 0 1.ts

If you need to apply different offsets to n streams, then you'll need to generate n outputs with 1 mapped stream per output with its unique offset. The remux all outputs together into one with -copyts added.

Gyan
  • 85,394
  • 9
  • 169
  • 201
  • This seems so obvious now that you've laid it out like this (facepalm). Going to try and get back to you! For the outputs, I can definitely figure out how to get it to where I create 2 separate .ts files (one with video only and one with audio only), but any idea how I might be able to do this in a single command with a complex filter? No worries if not - can definitely figure it out through trial and error! – Vinay Nov 22 '18 at 18:10
  • 1
    `ffmpeg -y -i 1.mov -map 0:v -c copy -output_ts_offset 4.54 -muxdelay 0 -muxpreload 0 1v.ts -map 0:a -c copy -output_ts_offset 4.41 -muxdelay 0 -muxpreload 0 1a.ts` – Gyan Nov 23 '18 at 05:08