On Tuesday 10 July 2001 05:37, you wrote:
> [...]
>
> > > If I'm not completely wrong the gnome cd player supports CDDB.
> > > The `ID' of a CD is derived from its total play length - taken as a
> > > hash value - and is used to look up the CD's data (volume and track
> > > names) at the CD data base on a CDDB server on the internet.
> > >
> > > Therefore if you don't change the total length of a copied CD you will
> > > get the same ID.
> > >
> > > Did I miss something?
>
> The checksum algorithm they are using will also use the numbers of the
> sectors where each track starts for it's computation.
>
> There should be an exact explanation of what they are doing on
> gracenote.com (or whatever the name may be today), at least there was back
> in the days when they were cddb.com.
>
> > WOW! I thought I was going crazy. So my empirical
> > experiment was telling me the truth, the wav files
> > are all that is needed to derive the lyrics, but because
> > I had no explanation, I could not believe the empirical
> > results.
> >
> > So, if I make an audio CD of my friend's .wav files,
> > I can "submit" the values to the CD database.
>
> don't do that. You personal best-of or remix CD should not go
> into the cddb, as it's useless for most people. If, however,
> it is an exact copy of the original CD, and that would include the
> offsets of the tracks, then mybe you could submit that CD...

No, this isn't a "personal best". My friend is a live music performer.
This is a CD of her songs, it's completely and totally original.
No track on her CD exists on any other CD.

It'll be useful for anyone who buys her CD just like any other.
She has a master CD, and someone else made copies for her.
When I get the master, sometime this week, I'll see if lyrics
show up on g-cdplayer. If yes, then ok, if not, then I can
submit them, improving the copies she has already sold.

>
> > I would think that total length would be insufficient for a signature,
> > but perhaps some hash value generated from some function
> > across the tracks. That is, if you substitute the 7th track with
> > another track with the same length, the total lenght is the same
> > but a function of the whole CD would be different.
>
> you don't mean the contents of the whole CD do you? Calcualting a hash
> over that would be both time-consuming and unreliable. using the track
> times is faster and yields reproduceable results.
>
> rainer


--  
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to