Hello,

I\'ve done some listening tests (again) of some things.

I concentrated only at stereo 128kbps (overkill) and  mainly at 2 clips (main_theme 
and joe satriani)

I concentrated at reducing all artifacts as much as possible and not at overall 
high-end sound quality.


1)  Leonid\'s fix 

Twinkling artifacts don\'t gone, but sound a little bit different.

Some time ago I tried -Y option and it produced sometimes really bad twinkling.
With Leonid\'s fix -Y option significantly reduces artifacts, but only when used with 
-X1.

Resume: I think Leonid\'s fix is right and with -Y -X1 it produces nice results 
(sometimes I got better results than Fraunhoffer).


2)  ATH

Without Leonid\'s fix and without -Y -X1 it produces audible artifacts. I tried to 
divide ATH and useful values was from 10 to 100 (it was depending on type of music) - 
I know it seems too much (20 to 40 dB).


  for (i=start ; i < end; i++) {
    freq = samp_freq*i/(2*576);     /* freq in kHz */
    ATH_l[sfb]=Min(ATH_l[sfb],ATHformula(freq));
  }

++  ATH_l[sfb]=ATH_l[sfb]/20;

}
same for short blocks

If I let ATH unchanged, joe satriani sounds worse (more flanger) than with lame v3.1x

3)  Decoders - I tried winamp2.5, MS Media Player and mpg123 0.59q. Each decoder 
sounded another way (and produces another kind of artifacts). Winamp is buggy/broken 
(I\'m sure about it, with joe sat. it produces artifacts in left channel regardless I 
tried to swap chanels in original wav and reencoded it - again artifacts in left 
channel). Best results I get with mpg123 but I can\'t listen under Linux (damned Li(v 
|| f)e!).



When I played with ATH coefficients a bit, near all artifacts gone on first clip. But 
when I tried another song and tuned ATH, I broke first settings. Should ATM be 
variable/adaptive?

I\'m sure with ATH we can outperform Fraunhoffer, but it needs _MUCH_ tuning.


My new top quality lame at 128kbps is: 

v3.27 with Leonid\'s patch, with ATH divided by 20, and with swiches -X1 -Y.

Please try it with hard-to-encode samples at lower bitrates, because my test was 
limited to a few clips.


Bye, macik

PS: I was surprised, but try to encode \"fatboy.wav\" using Lame and any mpeg1layer2 
encoder - layer2 produces much better results at any bitrate. Is there serious bug in 
layer3?


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

Reply via email to