[MPlayer-dev-eng] 2 pass encoding - need explanation

[Michael Niedermayer]

Hi

On Sunday 23 February 2003 02:20, Arpi wrote:
> Hi,
>
> > > On Sat, Feb 22, 2003 at 04:23:55PM -0500, D Richard Felker III wrote:
> > > > BTW, you should try with vqscale=2 on the first pass and see if that
> > > > works well, rather than vqscale=5. The latter will encode the whole
> > > > first pass at very low quality so it probably won't estimate
> > > > complexity well.
> > >
> > > What about the encoding speed? Someone told me that it is the
> > > same for each quantizer (I doubt it). And what about quant=1? I
> >
> > lower quant will be slower, as more bits must be encoded
>
> Hmm. What about an option to drop headers, vlc etc encoding for
> 1st pass? it's enough to do the MV search and quantization, as we
> only want the statistics, don't need the actual bits.
> Counting the required bits (sth like: total_bits+=vlc[n].length) is
> far easier & faster than doing put_bits()... but it may mean many if()'s
> so it's maybe not as good idea as i thought 5 mins ago :)
its planed ...

[...]

Michael