Ion-Mihai Tetcu wrote:
[ ... ]
You'd have to obtain the glass master used to press the original to
get an absolutely exact copy.  "bit-by-bit" accuracy of the raw CD is
not very easy or even desirable to obtain due to ECC hardware fixing
minor errors: for audio, you want a "bit-by-bit" accurate copy of the
data after ECC processing, not before.

OK, do you know any method to do it under BSD ? There are some cdtools in the ports that claim to do DAE - but I haven't the time to look in the code to see how it is done and if there is any difference between using a SCSI an an IDE drive.

Modern CD-ROM burners use the SCSI command set and thus behave much like a SCSI device modulo various per-device quirks: the ATAPICAM module connects IDE CD-ROM's ("ATAPI CD-ROM's") via the SCSI CAM system.

Better CD-burners let you do something known as DAO+96 ot TAO+96,
which lets you override the ECC hardware when reading or writing, and
thus copy even the errors from the original, which can be useful for
copy-protected data CD's rather than for audio....

We have a potential client who asked about studying both cases and make him a app that could say if it's an "original" or not so this sounds interesting (we're pretty new on this).

2048 bytes of "actual data" get turned into ~2536 bytes of Reed-Solomon ECC encoded data. There is also a 94 byte subchannel + 2 byte checksum (for "96") which is used to contain CD+G data and other information, which is probably a part of what you'll need to consider to have the CDDB diskid match.

For example, "man burncd" has the following:

     vcd           Set the write mode to produce VCD/SVCD tracks for the fol-
                   lowing image files on the command line.  This automatically
                   sets DAO (-d) and ``no gaps'' (-n) modes.

A good copy of a professionally mastered audio disk obviously will not have 2-second gaps between tracks (burned DAO and not TAO), will be closed ("fixate"), etc. I'm not an expert, however-- perhaps talk to <sos at> who wrote a goodly portion of this infrastructure for FreeBSD...


_______________________________________________ [EMAIL PROTECTED] mailing list To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Reply via email to