Andrew Church wrote: >>As an aside, it it possible to specify a multipart .nuv fileset as input >>to 'transcode'? >> >> > > Not at the moment, I'm afraid. >
This is how I discovered the glitch with multipart .nuv I figured that there was no direct support, so I transcoded both parts of a two-part .nuv into MPEG2 video and audio elemental streams, concatenated them with mpgtx and multiplexed the results. I had to lose a couple of frames off the start of the second part of the multipart .nuv file to avoid a brief black flash, but other than that it worked OK. Hardly simple-to-use though! :-) But it did the job as long as I didn't mind the 44100Hz audio stream. What *didn't* do the job was that I found that I had to resample the audio to 48000Hz for the .mp2 stream for a PAL DVD. Transcode tried to use ffmpeg for that and made a subtle mess of it. I'll get back to y'all on that: I've got to go and repeat the problem with logs and decent info. In a nutshell, the resulting 48000Hz .mp2 stream ran slow w.r.t the video. By the end of an hour long programme, the sound was several seconds out. The lip-sync was noticeably bad after about 10 minutes. I then tried transcoding the audio to a 48000Hz .wav file and found it was the same contents as an unresampled 44100Hz .wav file but with the header changed to claim 48000Hz! I had to use 'audacity' on the 44100Hz .wav file to convert it to 48000Hz and then use 'toolame' by hand to generate the .mp2 audio. Then multiplex it with 'mplex' by hand so to speak. As I said, I'll repost with details on that one when I have them. Steve
