[BC] dueling algorithms and audio quality

[Jeff Johnson]

I've been playing with algorithm stack for a number of months now. For 
source files, I recorded from older analog LPs - those produced before 
anything digital could have been used. I recorded with a good sound card at 
96kbps and 24 bits/sample. Then the files were converted to various flavors 
of MP3 and AAC, back to analog then back to the other compression 
algorithm, etc.

There is a bit of a difference depending on which was first. When I have 
enough various examples, I plan to do double-blind tests and we'll see.

Something that sounds a bit nasty is HD2 at 48kbps then into and out of 
ATRAC/MiniDisc.

Jeff.Johnson at rfproof.com

>The toughest problem I find lately is that there is a basic
>misunderstanding about data reduction algorithms.  I wish I had a nickel
>for each time someone told me "but I converted it to a WAV".  Trying to
>explain that once it is "compressed" (I hate that term) it is always
>compressed.  I try to use the screen door example.  If you stand far
>enough away from a screen door you do not even notice the missing
>parts.  Place several screen doors back to back and pretty soon you
>cannot get far enough away to not notice.  Conversely, converting a MP3
>(or whatever) to a WAV only makes the door and the screen bigger.  The
>screen is always there.
>
>Craig Bowman