> If one instance of lame were able to encode multiple output files from the
> same input file/data (eg "$ lame in.wav -b 128 out128.mp3 -b 192
out192.mp3"),
> could it possibly be substancially more efficient than several independent
> instances of lame? Like are there some cpu intensive calculations that
> would have the same or similiar results for different bitrates (perhaps
> having the same output sampling rates) that could be reused instead of
> multiple processes recalculating the same info?
>
I'm not sure that it would be really interesting. Most of the time, when you
stream multiple bitrates, they are also of different sampling rates, so you
won't gain anything. Moreover, if the sampling rate is the same, if you
change the bitrate, the frequency cut-off will be different. Computations
could be re-used in this case, but not so much, mainly a few parts of fft
and mdct, but they do no use a lot of the global computation power needed.
Regards,
--
Gabriel Bouvigne - France
[EMAIL PROTECTED]
icq: 12138873
MP3' Tech: www.mp3-tech.org
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
