ozpeter
Location: Australia
Posts: 3200
|
Posted - Wed Apr 02, 2003 7:47 am
|
|
|
Regular readers may recall that I use Arrowkey's "CD-R Verifier" software to ensure that, at the completion of a project, every CD I send out is a verified exact clone of the master CD I listened through. I can thus be confident that any of the copies is as good as any other for the critical purposes of (say) broadcast or mass duplication. If you are not interested in this, please stop reading now!
"CD-R Verifier" will detect if a single sample of the copy is different from the original - believe me, I've checked that for myself.
I've recently refined my procedures a little so that I can simply copy from a master CD in one drive to a copy in another, and still obtain a verifiable copy. You might think that with products such as 'Clone CD' or 'Exact Audio Copy' there is no problem about this, but due to differences in padding and other matters that I don't pretend to understand, 'CD-R Verifier' will fail CD copies made from a master audio CD. While for all practical purposes they may well be perfectly good copies, they are not verifiably identical copies.
What I've now tried, and found to work, is the following. Having used Cool Edit to prepare the wave files using the usual cue-merge-batch system, and Nero 5.5.10.7 to prepare the compilation, I burn it to an image file on the hard disk (which of course is a quick operation). I then use Nero's ImageDrive utility to replay the parts of the image which are most likely to give trouble, namely the beginning and end of each track to ensure that there are no inter-track clicks or omitted fades etc.
Then if all seems well, I burn the image to a CD-RW and play it all the way through looking for any audible defects. If it's OK, I use CD-R Verifier to obtain the checksum signature of the disk for comparison with later copies. Now I burn the copies which are immediately required and compare their checksum signatures with the monitored master.
Finally, to archive the finished project, I burn the image file as a data file onto a CD, with data verification turned on in Nero. (Nero will verify only data CDs, not audio). Now, if at a later date further copies are required, I can use Nero to burn them direct from the image file on the archived CD, drive-to-drive (in my case one is a DVD/CD reader, the other a CD burner). I then use CD-R Verifier to compare the checksum signatures of the new copies against that of the original master, of which I keep a text record.
The only catch is that an image file for a 79' CD will not fit on a 700Mb CDR. I have found that WinZip will compress the file enough to get around this problem - it does mean that you have to unzip the file from your archived image-file CD to your hard drive to restore the image file before burning it - this doesn't take long, and in my experiments did not detract from the verified accuracy of subsequent copies.
So, my recommendation is that if you want to make exact copies of your project from a master CD to a copy CD, use an image file burned as data as the master, rather than a normal audio CD - my tests indicate that you should obtain an exact clone by this method rather than the other, even if you are not going to the trouble of verification.
- Ozpeter
|
|
|
|
MusicConductor
Location: USA
Posts: 1524
|
Posted - Wed Apr 02, 2003 4:47 pm
|
|
|
Thanks for the update, Oz.
Quote: |
but due to differences in padding and other matters that I don't pretend to understand, 'CD-R Verifier' will fail CD copies made from a master audio CD. |
I've heard that's because the exact start position of a CD-Audio write and a CD-Audio read is very difficult to match; probably all the data's identical once it actually gets going.
I've been archiving CD masters to CD-ROM for years, but without the assurance of your method of verification. A big wave file or group of files and a CDRWIN cuesheet is how I've burned CDs CD-ROM to CD-R; if the disc was too long, I'd place the end of the CD in a separate file on a second CD and copy it to the harddrive before making the burn.
As I said in a previous discussion along these lines, I wish consumer drives and software would support block error rate readings.
Thanks again for the update.
|
|