Here's an interesting message I thought should be posted
to the mp3encoder list:



------- Start of forwarded message -------
From: "Aleksandar Dovnikovic" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Subject: MP3 and hi-frequences
Date: Fri, 14 Apr 2000 21:00:41 +0200

Hello.

First of all, congrats on a great job you guys are doing with LAME.

I just wanted to point out the problem that all MP3s (more or less) have
with high-frequencies (>16kHz). Interesting thing is that this problem is
practically in the left channel - you can hear frequencies over 16kHz being
"echoed", "swirled" in left channel. Now LAME indeed does the best job with
these frequencies, FhG & Xing do pretty worse while Bladeenc is horrible.
When I first discovered this, I thought it had something to do with joint
stereo, but it didn't. I mean, if you can get proper hi-freq. encoding in
the right channel, why can't you do it also in the left? Is it encoder using
some intensity stereo for those frequencies? Why doesn't stereo mode help
when the encoder then encodes channels separately (but with different
bitrates). Maybe you should try adding an option in LAME that will encode
everything in 16-20kHz range in dual mono...?
FhG AAC doesn't have these problems with hi-freq.

Also since LAME has worse pre-echo detection then FhG MP3 (the only thing in
which FhG still holds the edge over LAME) have you thought of reverse
engineering mp3enc31.exe - I know it can be pretty time consuming but most
of you LAME guys are quite experienced so you might be able to extract what
algorithm FhG uses...

Your comments will be very much appreciated.
Keep up the good work.
------- End of forwarded message -------
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )

Reply via email to