What are the specifications of the checksum used in AccurateRip's database?

Hello,

I have been trying for a few days to find the exact specifications of how to compute AccurateRip's chesksums. AccurateRip is a centralized database of ripped CD tracks checksums, that allows one to be sure that his CD rip was flawless. It is very useful when archiving audio cd losslessly.

The access to the database is free, but the specs are impossible to find. They were reverse-engineered in at least two free sofware utilities (one in perl, the other one in C++, both language I'm not fluent with): arcue.pl and arcue.exe. See:

http://perm.david256.com/2007/07/10/ARCue.pl
http://www.hydrogenaudio.org/forums/lofiversion/index.php/t60440.html

What I need to know is:

- How discs are recognized and matched to the DB;
- How track ranges are computed (the start and the end of the actual sound stream to be checksum'd);
- One you know that range, how do I compute its checksum (does not look like a CRC32) so that it is compatible with AccurateRip's.

I have a fairly good knowledge of how audio CD work, the problems they have when you convert them into a computer file (how to deal with offset, for instance), of the de-facto .cue sheet standard to model a CD, but how AccurateRip's tracks checksum work looks like black magic to me.

I'd be happy with a simple English description of the process (which should be 10 or 20 lines of text, maybe less, from what I have read/understood). I will chose what language I will use when I know the specs.

Thanks.Start Free Trial