On Mon, 05 Jun 2006 12:17:46 JST [EMAIL PROTECTED] (Andrew Church) wrote: > >That's good for demuxers, but not so good for muxers. There is a couple of > >issues: > >1) encoder core code (src/encoder.c) uses single-instance multiplexor and > >runs it, as well as A/V encoders, in a single thread. This happens both on > >OMS and in NMS-powered code. I'm pretty satisfied of this situations > >because it's work quite well, it's clean enough and... We need to join > >streams at some point in processing pipeline! > > Why? If you're outputting the audio to a separate file, it doesn't > make sense to try and keep them in sync anyway.
Yep, that's true for this case. My key point on encoder it's just that our main encoder (I mean src/encoder.c) expects also to join streams, so separate A/V files is the exception, not the rule. That's the rule that trained the other conclusions so the rationale for my first point was... My second point (not quoted here) ;). > It's easy enough to create > a second multiplexor instance if we're outputting a separate audio stream, > and call the multiplexor twice instead of once, like the (incomplete) patch > below. That's a pretty nice idea indeed, but the the last one it's even better > Or we could just drop support for separate audio export completely; > the main reason for it (MPEG) will go away once we have MPEG multiplexing > support, and if anyone really needs it we can tell them to do the audio run > separately: transcode ... -o foo.wav > > So I'm still in favor of only one file argument (and no audio/video > flag) to open(). That's pretty fine excepts for the fact that MPEG multiplexor will not pop out soon, surely not for 1.1.0 (at least not from my side, unfortunately :( ). So we need a temporary solution and perhaps the better thing it's to add a second-instance multiplexor (triggered by -m usage, I think). Best regards, -- Francesco Romani - Ikitt ['people always complain, no matther what you do'] IM contact: (email-me, I have antispam default deny!) icq://27-83-87-867 some known bugs: http://www.transcoding.org/cgi-bin/transcode?Bug_Showcase
