Where do you take that time offset value from?
I play the simplified ts file using smplayer -ss is the start in the ts file, -t is the duration to add to the -ss value (I would have prefered the absolute end avoiding a substract...).
I'm also working on such kind of scripts right now, because I want to remux especially HD recordings to MKV. In the case of movies broadcasted by ARD HD or ZDF HD (720p@50Hz), they are full of so-called filler-NALUs, which can be removed just by muxing into MKV not directly by ffmpeg, but by mkvmerge, which shrinks a 10GB movie to about 3GB.
You mean you extract the various component (audio, video) into separate files and then reassemble them using mkvmerge?
My problem is still A/V sync, all the original TS chunks (I also record into 200MB pieces, because of fast editing in VDR) show video delays for each audio track of over 1 second, which is also very noticeable ion the final converted MKV. The delays are reported both by tsMuxeR and mediainfo see for example the "Video delay" value below:
Yes I know this is why the "async 2, (+ vsync 2 for récent ffmpeg version)" is there. Without them I have the audio and the video out of sync just like you.
ffmpeg -y -i simplified1.ts -vcodec copy -sameq -acodec copy -vsync 2 -async 2 -f matroska 1.mkv
I'll post the mediainfo on the ts and mkv when back at home.
-- eric