[arch-general] ejecting after cdparanoia

[Joerg Schilling]

Jorge Almeida <jjalmeida at gmail.com> wrote:

> On Wed, Sep 14, 2011 at 9:44 AM, Joerg Schilling
> <Joerg.Schilling at fokus.fraunhofer.de> wrote:
>
> >
> > cdparanoia is a cdda2wav version from 1997 with some modifications.
>
> Even the recent version?
>
> Cdparanoia

Correct, Monty did take a cdda2wav release from before the first 
major rewrite has been done.

There was a major rewrite starting in 1999 another one starting around 2006.

> > is Linux/gcc only.
>
> Tha'ts fine by me. I understand devs will think otherwise, but for a
> linux-only user portability is not a plus nor  a minus.

All my software is portable to virtually all platforms and this is why I need 
to make code portable before I can ise it.


> Joerg: I've been using your software, and I intend to keep using it.
> Yes, I know about the  -paranoia option. But sometimes cdda2wav report
> some (usually minor) problems. Example:
>
> 100%  track  4 recorded with minor problems (0.2% problem sectors)
> 100%  0 rderr, 0 skip, 0 atom, 1 edge, 81 drop, 1 dup, 0 drift
> 100%  466 overlap(0.5 .. 0.8274)

"minor problems" means that these problems are not expected to cause audible 
results. If you like to most agressive parameters, I recommend to call:

	cdda2wav paraopts=proof


> I suppose this is essentially harmless, but, being somewhat
> perfectionist, I would rather have 0 problems. Most tracks usually
> come out without errors. So I tried cdparanoia, and it did the ripping
> with two "+" signs at full speed. So I tried at speed=1 and it came
> out completely clean.

This just verifies that cdparanoia doesn't inform you about the problems.

> Now, I'm not that naif as to assume that "no errors reported" <=> "no
> errors at all". (Maybe 0 errors from cdparanoia means that the 81
> drops are still there but deemed irrelevant?) I did visit the
> cdparanoia site, but I didn't leave much wiser.
> So, for someone who wants to rip a brand new, duly purchased, never
> played before, CD with the maximum quality, what would be useful to
> know (with proper arguments) is:

I fixed some problems in the praranoia code related to error reporting....

> -- When one of the programs ("cdda2wav -paranoia"  or "cdparanoia")
> reports less than optimal results, is it worth  to try the other one?

My ststistic experience shows that you will usually not get a better overall 
result if you repeat the extract with all tracks as usually one track will be 
worse then before. I thus recommend to repeat extracting single tracks.

Note: cdda2wav supports MD5 sums on the audio data since 2008. This allows an 
easy check on whether an identical result may give different paranoia 
statistics.


> Note that this wouldn't mean that one is better than the other: they
> might use different algorithms, and it might happen that one algorithm
> performs better for a particular CD. Then again, what I'm saying may
> be complete nonsense. I am not qualified to read the source of either
> program.

Before 2006, cdda2wav did not enable dynamic overlap with libparanoia.
This may cause different results.

Jörg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily