> > Ok, but how about --athlower 100 ? Will it leads to an endless loop?
That
> > could be a problem. I think that when --athlower is used, the lowering
of
> > the ath in sfb 21 must be reduced for vbr.
>
> Did you notice any endless loop using --athlower ??
I tested the behaviour of lame vbr-old with --athlower 100. The good news is
that there is no endless loop. The bad news is the result.
I used lame -V1 -h -mj --athlower 100. With --lowpass 15000 the mp3 is
181.9kbps, wich is about what I expected, no problem here. But with -k the
mp3 reach an average bitrate of 320kbps because each frame is 320kbps. What
I suspected is happening here: an incredible amount of bits are used in
sfb21.
Why would any user would use such an high value of --athlower? (ok, 100 is a
little too high)
Because I think that the default ath for v1 is too high, and I don't want
the low intensity passages to be cut-off. But I still want to benefit of the
analog silence->32kbps for parts where the intensity is very low, ie no
instrument playing, only the potential noise of the recording material, wich
I know to be extremely low for some of my recordings. This is why I'd like
to use something like --athlower 50.
This is why I think that trying to put the sfb21's ath value in both masking
and ath would be good, if possible. This way, using -v1 the sfb21 masking
would still be reduced, but it would be possible to also use --athlower with
some high values.
Regards
--
Gabriel Bouvigne - France
[EMAIL PROTECTED]
icq: 12138873
MP3' Tech: www.mp3-tech.org
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
