Hi,

It's been a while since I did any serious listening, but the CVS
version I just downloaded is working astoundingly well here. The
classical record company for which I do all the RealAudio may want
to look at higher-quality methods of distributing its demo audio, so
I can't help suggesting MP3/LAME with its lack of reliance on a
single proprietary vendor, etc... (oh, how I'd hate to see them
using MS Audio! or RealNetworks with its ever-present advertising
and requests for paid-for upgrades).

For demo purposes, I'm using the latest CVS version at 96kbit/s,
with the flags:

-d -h -X 4 -Y -Z

This may be completely wrong, but the code (which I don't really
understand) connected with these flags looked as if it might add
some extra sophistication to the encoding process -- please tell me
if I'm error!

On most `classical' music including opera, piano + orchestra, solo
piano, strings, organ and harpsichord, your latest coder is
WONDERFUL.

I have a possible test case that doesn't sound so good, though: a
beautiful setting by Monteverdi of the `Cantate Dominum canticum
novum...' sung by Westminster Cathedral Choir on Hyperion Records,
CDA 66850, track 18. I find that at 96kbit/s there is an
upper-mid-range fluttering sound that isn't on the original CD, and
I think that the incoherent sound of so many voices in a big,
reverberant setting is giving the coder much hard work to do.

Perplexingly, when I direct the coder to use VBR at quality 0
(best), 96kbit/s max, some of that flutter goes away. I've converted
the non-VBR and the VBR MP3s to WAV files (using the -w option to
mpg123), subtracted one resultant file from the other, and get a
most amazing difference-sound, but that is probably beside the
point.

I've posted some of this Monteverdi test case (in WAV format) at
http://www.bbcradio.demon.co.uk/test.wav if you wanted to try it,
and to see if it's worth adding to your own test cases.

It's 3.5 MB, and lasts around 20 sec.

Thanks again for helping me promote (what I hope is) `good' music.

John Hayward-Warburton
[EMAIL PROTECTED]


--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )

Reply via email to