Hi there,
I wanted to test the VBR of Lame 3.82 so I made this wav in cooledit pro:
Initial Settings:
Base Frequency 20 Hz
All other settings are same as with default A 440 preset (don't want to type
them all here)
Final Settings: like initial, except: Base Frequency 20,000Hz and Log Sweep
turned on.
length - 10 seconds
For encoding I used lame 3.82 from http://www.chat.ru/~dkutsanov/. For
decoding, Exact Audio Copy with Radium fhg codec 1.2.
First I encoded the file with lame -V 1 -b 128 -h -m j
After decoding and opening in CoolEdit there are clearle visible volume ups
and downs in waveform view. Freqency analysis that the frequency curve is
not almost almost straight like in original file but there are also ups and
downs above 16 khz. When listening, there are very significant chirps (as
vdbj decribes them in his mail from 13.5.).
I also tried following configurations:
-V 1 -b 128 -h -m s
-V 1 -b 128 -h -m j --strictly-enforce-ISO
with same results.
So the problem is not in ISO specs or joint stereo.
I also tried to encode with -V 0 -b 256 -B 256 -m j -h -F so the file was
not even VBR but contained only 256 frames, and was almost same size as
normal CBR256 file. It produced the same bad results, while cbr 256 was
without errors.
Using winamp 2.62 to decode i got also the same results (+ errors that are
already known with this plugin)
I tried LAME 3.81 and it was also the same.
The most interesting thing is that vdbj's sweep produced with Sound Forge
(described on http://www.r3mix.net/) was clean! And it has the same
frequencies as mine sweep, even higher.
I also found that -k switch used in VBR mode produces binary same mp3's as
without -k switch, so in VBR mode there should be no lowpass filtering.
Testing with some music source the analysis was also different above 16 khz
(as vdbj reports in his mail).
If anyone wants to send the sweep.wav or snaps from freq.analysis, mail me.
Bye,
Vlastimil Babka (Caster)
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
