View Single Post
  #6   Report Post  
Old January 16th 04, 03:24 PM
malcolm
 
Posts: n/a
Default


"Ben Bradley" wrote in message
...
In rec.audio.misc,rec.audio.tech,rec.audio.pro,rec.ra dio.broadcasting,
"William Sommerwerck" wrote:

The Exact Audio Copy program checks for corruption when ripping a file

from a CD
to WAV format. But I don't think it can analyze existing files.


Reading from the CD is different, because the drive gives errors
when it can't read/reconstruct a sector from the CD. Also, the actual
data from the CD is "raw" whereas a .wav file is not.
As Dick Pierce said, the .wav file is not "raw" but has a header
with sample rate, number of channels, number of bytes per sample, file
length and such. This can be checked for "reasonable" values and the
claimed length against the actual file length, things like that.

In principle, you could write a program that opened and read the WAV

file,
looking for such things as excessively long all-zero or full-scale

sequences, or
abrupt jumps in sample values.


This is a way to discriminate in the .wav data or a raw file
whether the content is "musical" or not, but may not be foolproof.
All-zero sequences might last for over four and a half minutes. g
An LP declicking algorithm would find abrupt jumps with no trouble.
-----
http://mindspring.com/~benbradley


ah modern art, wonder if they would hand out a grant for a new version g