[MEncoder-users] x264 2-pass, FAAC, smearing

[GM]

On Fri, 2009-01-16 at 16:59 +0100, Robert Ramiega wrote:
>  Of course not. It's about bitrate distribution.
>  Still I don't see how omitting audio can screw video bitrate calculation 
> (and thus bitrate distribution)
But it doesn't screw bitrate calculation.
It simply timeshifts the second pass by an amount of a couple of
milliseconds, I guess in an effort to sync A/V.


> I did just now. I don't belive that bad bitrate distribution comes from
> different audio options. You haven't written anything about constraints on
> output file (bitrate etc) but normally you would use video bitrate of say
> thousands of kbits and audio bitrate - hundreds of kbits. Maybe you've hit
> some bug =o)
It is not important. If it were, I would have outlined it.
Video is baseline, level 3, 1 ref frame, 750kbit. Audio is 128kbit
AAC-LC. But it can be anything else, I tested it extensively.

> 
>  It might be helpful when you post full command-line of both passes not just
> some tidbits as in first message in the thread.
It doesn't matter, as obviously you missed the point of the first
thread.
I can reproduce it EVERY time with ANY input video with ANY x264
settings if i use -oac copy for first pass and -oac faac for second
pass.

> When i do some transcoding i use hand-crafted batch created by h264enc. This
> batch encoded video and audio seperately later muxing them to mp4 or mkv
> (-noaudio on both passes then dump audio to a pipe and encode from it - note
> that i'm using neroaac for this not mencoder). Did you tried this approach?
That is always a bad idea, as sometimes mplayer has to drop or duplicate
frames to keep A/V synced.
Yes, I have tried, it works of course, but it causes A/V sync issues
with some videos.

> > _______________________________________________
> > MEncoder-users mailing list
> > MEncoder-users at mplayerhq.hu
> > https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
>