After conducting some tests with the new VBR j-stereo code I am quite
pleased.

I replaced (1+fac)/(1-fac) with fac, and I was encoding with -X4 -V 0 -h

(btw- I'm working on a -Q for better VBR quality control)

I was using track 2 from Extreme III, which I often use for listening
tests.

I encoded three streams:

VBR -ms -v -V 0 -X4 -h
VBR_J -mj -v -V 0 -X4 -h
Fixed -mj -b 112 -X4 -h (most common bitrate of VBR_J)

We tested this along with the orignal.

My two subjects could easily tell the Fixed from the orignal, the VBR ones
sounded much better then the fixed. Still the VBR ones were noticably
less-good then the orignal (this is IMHO because the current VBR -X4 VBR
quality setting is too lax, The quality code I'm working on should help).

They though the VBR_J didn't sound as alive as VBR. They couldn't state
HOW but they said both were pleasent to listen to.

After I explained the difference between the files they imeaditly pointed
out 'lack of stereo imaging'.  

So it appears that VBR-j lessened stereo imaging, but we did not find it
to cause any noticable or annoying artifacts. Which is very good
considering that VBR-j was 34% *smaller* then VBR!

Does anyone know of any samples which have done very poorly with VBR
j-stereo in the past? 

--
The comments and opinions expressed herein are those of the author of this 
message and may not reflect the policies of the Martin County Board of
County Commissioners.

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

Reply via email to