On 30-Sep-99 Mike Oliphant wrote:
> I think the resampling should be part of LAME, as it is critical for LAME
> to be able to produce good low-bitrate mp3s.
I agree. Sort of.
I once had a *really* bad resampling algorithm to get from 44.1 > 32khz. The
sampling was quick enough to be instantaneus, and I was able to get realtime
encoding out of lame on my p166 box (quality was awful though :)
I'm concerned that we'll basically be taking the SRC functionality of sox and
putting it in lame. Why not just use "sox | lame" ? If you have a mp3 that
you need at a different rate, why not "mpg123 | sox | lame" ?
Both Xing and Fhg have resampling code as part of their encoders afaict. And
the idea that you can feed in any stream and get out any mp3 stream seems like
a good idea.
> This is why I plan to allow LAME to
> optionally be compiled into Grip (as soon as I get some time...)
BTW grip rocks :)
later
mike
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
