[paranoia] Streaming Errors

[Monty]

> Monty,
> 
>      I (finally) got the chance to test the streaming error corrections.  I
> did two runs on a CD that created many "+"s in the progress bar (but no
> "V").  As Bill suggested, I converted the data to ascii and compared.  As
> the data below show, the two runs, while very similar, do have slight
> differences.  
> 
>      How does cdparanoia decide whether it's DAE is good enough?

When it gets two successive reads over an area of at least 64 samples (and 
overlapping with another good section by at least the same amount), and no 
other read disagrees (if it happens to read an area more than twice to get 
overlapping, and 'new' data consistently disagrees, it keeps reading).

In your case, the interleaved nature of the 'bad' samples is a dead giveaway 
of a real media error.  The drive itself is performing error correction, 
although these sections are unstable enough that subsequent reads are probably 
disagreeing by a bit or two with the probability making it impossible to tell 
what the correct data is.  So, the drive (and cdparanoia) has to guess (and 
the odds are close enough to even that subsequent reads could tip it either 
way).  Paranoia II would actually discard these 'unstable' values and 
interpolate through, but paranoia III doesn't have that feature (too hard to 
do reliably on all drives, although I'll give it another shot someday).

Monty

--- >8 ----
List archives:  http://www.xiph.org/archives/
Paranoia homepage: http://www.xiph.org/paranoia/
To unsubscribe from this list, send a message to 'paranoia-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.