Syntrillium Backup Forums :: View topic

HanzZ · Location: Netherlands

Hi all,

Last weekend I transferred a lp to cd for a friend of mine. By the time I was ready to burn it on a cd, I had listened to the songs so much that I couldn't stand them any more Smile

So I thought: is there a way to verify the burned cd (other than listen to it)?

Then it came to me: how about creating a delta file? With CEP2 you can import cd tracks, right? So what if I burn the cd, import all tracks back to my HD and flip the phase of the imported tracks and mix them with the versions before burning? The result should be the difference between the two. The difference should be zero. If it isn't zero something went wrong during the burning proces.

Am I right?

How do you all go about verifying burned CD's other then listen?

-- HanzZ

OBuckley · Posted - Fri Feb 21, 2003 8:54 am

Hans. Nero, for instance, has a check box which allows you to verify bit for bit what went into the CD with the source files. It just takes a couple of minutes longer at the end of the burn process.

Andrew Rose · Location: United Kingdom

Neither method -listening or inverting - should work, as error correction in the CD player or CD-ROM should correct any minor errors on the disc. If you have Nero you should be able to use the Scan Disc function in the Nero CD Speed program (in the Nero Toolkit) to check the disc - it gives a visual readout of good, damaged and unreadable sectors on an audio disc.

Using this I realised my old Philips burner was useless. Even though I'd never heard it produce a disc that skips, other people were having problems and I was blaming the media. Then I found it was leaving errors across the discs. Once I replaced it with LiteOn burners I've had no problems - and no returns either!

MusicConductor · Location: USA

ozpeter · Location: Australia

This post is slightly off-topic because I realise the question relates to a single CD burn, but here goes anyway.....

My procedure is as follows: I do listen to through the first CD copy for anything untoward - usually any problems are of my own creation! Perhaps therefore I omit one previous listen before the material is burned, knowing that it will get this final check, and I have of course the option of using a rewritable disk, which I can burn at x10 these days, if I want to avoid wasting a write-once disc. (My copy of Nero only allows for data files to be verified, by the way). But having got an acceptable 'master' CD, I then obtain an overall checksum of its contents using 'CDR Verifier' from Arrowkey. I keep a note of this checksum on a particular text file on my machine, so I've now got a list of completed jobs and their checksums. Then I use that checksum to verify (using the Arrowkey software again)any further copies I make. This takes as long as it takes the drive to read the disc at maximum speed, and doesn't tie up the computer exclusively. The checksum will not match if there is a single sample different (I have, of course, tested it!).

What I do find is that any form of disc to disc copying will change the checksum - and this has implications for invertion checks - the only way I have found to clone the CD exactly is to make the subsequent copies from the original data files, restored to the PC if necessary from the data file version of the job which I always make. I add to the data file version the text file of checksums, and also the file Nero uses to compile the CD (.nra I think), especially as that contains the CD text information.

Using these procedures I find that I can confidently make CD copies without having to listen to a single note, for which I am often grateful! The failure rate is extremely low, even when using x16 copy speed, but let's not start that subject again!

- Ozpeter (he of the ever lengthening posts)

MusicConductor · Location: USA

Oz, that's great, and not at all off-topic IMO. Yes, always important to do a confidence check before the final "printing out." I haven't heard of using checksums in this way, so thanks for a valuable idea. Even if it doesn't get us down to the BLER level, at least it's a great deal more efficient than rip-and-invert!

BillG · Location: USA

I'll often use Exact Audio copy to verify a CD that I've burned. Simply open EAC with your CD in the drive, select that tracks you wish to verify, and press F8. Then DON'T TOUCH YOUR COMPUTER WHILE IT RUNS. Even moving the mouse risks increasing the reported error rate (i. e. reducing the Track Quality %). When EAC is done analysing, click on OK, and a track by track report is displayed. This includes a "Track Quality" percentage that ideally is 100%. Any time EAC has to reread a track to obtain a perfect copy, it lowers the Track Quality %age. Ordinarily, the Quality %age should be between 99.6% and 100.0%, and if some part of a track is imperfectly readable after several tries, EAC will report an error with that track. Often, if a track is reported with errors, rescaning it will report it as OK.

I'm sure this is far from the last word in CD verification, and it's certainly not a detailed or hardware level verification. But it seems to have worked well for me, and best of all, it's FREE. http://www.exactaudiocopy.de/index.html

ozpeter · Location: Australia

Very true, BillG - so long as any error is on the CD rather than something wrong with the data coming off your hard drive - and it also includes a track-by-track checksum that can be used to compare the first and second copies - use 'test' on the first, then rip the second and the two checksum columns should be the same. However, I've not found any elegant way to keep a record of those columns for future reference. (Any suggestions?)

- Ozpeter

post78 · Location: USA

Bill, that method doesn't work. EAC is checking the quality of the rip, not the burn. It's useful in determining if there were any errors when ripping what's already on the CD, but not the amount of errors present on the CD itself.

post78 · Location: USA

Hi, Oz. I didn't see your reply.
Am I missing a function of EAC or something? I thought that bill was talking about the error rate displayed when ripping a CD. Question

ozpeter · Location: Australia

I'm talking about 'Read CRC' and 'Test CRC'. If you use F8 (in EAC) it fills in the 'Test CRC' column, and if you actually rip the disc it fills in the 'Read CRC' column - using two copies of the same disc you should get the same figure, and I think it fills in a tick or cross in the CRC column. Right at this moment I'm not able to fully verify what I'm saying but I think I've got it right! And yes, that's different from the track quality info that BillG was mentioning. (Trust me to complicate matters...)

- Ozpeter

post78 · Location: USA

But does it compare the file on your HDD (master or mixdown) to the file that was burned onto disk? It sounds like the EAC method is only testing the quality of the rip; what was ripped compared to what's on the CD.

HanzZ · Location: Netherlands

Yep, that's what I want to know too...

I'm looking for a way to compare the HD file with the burned file. Comparing copies is, although of course important, not my first concern...

-- HanzZ

ozpeter · Location: Australia

The problem with comparing the file on disk with the file on the CD arises, I think, with offsets and/or padding applied by the writing operation. In a previous thread on the subject, some kind person (Graeme maybe?) pointed me in the direction of 'windiff', which can do the job, but is as user-friendly as a cornered rat and requires a bit of interpretation. So that's why I usually just grit my teeth and play through the CD, one more time folks....

- Ozpeter

(Check out the arrowkey site - they've got other products including burning software which I think is relevant).

BillG · Location: USA

Well, as I indicated, using EAC is less than perfect for CD verification. It indeed only checks the "readability" of the CD, and does not do a "file compare" with the original .WAV file. And, of course, since a .WAV file and a CD track are completely different formats, one cannot in any case do a dirct bit by bit compare with the original file. Moreover, as a top US CD mastering technician once explained to me, it's pretty much impossible to do a direct "hardware level" verification using an ordinary CD drive because one cannot bypass the drive's built in buffering and error corretion.

So EAC seems to work well enough to verify the "playability" of a CD. I also usually "verify" a CD in a different drive (my DVD drive) or on a different computer than the one that I burned it in. Works well enough for me. But we might consider asking the creator of EAC to further enhance its verification capabilities. It would make a great piece of freeware even better.

MusicConductor · Location: USA

ozpeter · Location: Australia

Another installment concerning my quest for CD verification perfection. I typed most of this offline, and now that I re-read the thread I realise it kind of duplicates material above, but perhaps it's still useful as I've spelt out the procedure somewhat.

I've now spent more time on a simple method for verification using EAC. I've also verified the EAC verifyer by making a single-sample alteration to a file, which EAC does pick up as a difference.
Bear in mind that I'm talking about bit-by-bit verification of one CD against another, not about EAC's quality and error readouts.

Assuming you've already set EAC up to get the best out of your CDRW drive, and generally chosen options in the program, the procedure is...

Insert your "master" CD (the one you've actually listened through, so you know it's OK) and highlight all the tracks on EAC's main page. Press Shift/F5, which will copy the contents to your HD in compressed form. To save disk space (if that's a concern) you can compress this copy to an unlistenable extent - to a mono 8K mp3 if you want - that won't affect the outcome at all. If disk space is of no concern at all, you can use F5 instead to copy the contents uncompressed, which might be quicker.

After the contents have thus been ripped, and you've pressed OK a couple of times to clear the messages (which might make worthwhile reading before clearing), you should now have values in the "Read CRC" column for each track. You can now delete the tracks you've ripped - it's those "Read CRC" values you actually wanted.

Now insert the second copy, the one you want to verify without listening. Highlight all the tracks again, but this time press F8. When the procedure finishes and you've cleared the messages, you'll see values in the "Test CRC" column and the "CRC" column to its right should have "OK" all the way down, if the "Read CRC" column and the "Test CRC" column are the same. If they are not the same, the CDs were not the same (or you've got an inaccurate CDRW drive, in which case you are in trouble anyway). On my system EAC runs at full read speed in this situation as the CD quality is of course pristine (or should be!).

To check any subsequent copies, just use the "F8" bit of the procedure to obtain new "Test CRC" values. EAC maintains a database of CDs and you shouldn't ever need to obtain the "Read CRC" values again. But you might want to back up the EAC database from time to time.

If you want, you can clear the "Test CRC" column using a menu option which isn't hard to find, so you have the comfort factor of EAC visibly re-entering the values each time you do a verification run.

I've also now managed to get EAC to make verifyably identical copies of CDs, having got the "Read sample offset correction value" right. The way to accomplish that is well documented on EAC sites.

I've also found that EAC can be used (in a manner similar to the above) to check a Nero image file mounted in the Nero ImageDrive against a copy burned to CDR from that image, except the last track always comes up with a difference for some reason. That's not quite the same as verifying against the HD contents, but it's close.

- Ozpeter

cooknkpl · Location: USA

While we're on the subject, riddle me this Batman...

I burned 4 CD's via Nero at 40x using the multiple burner function (yea, finally figured it out). Tried playing the CD's in a boom box, car stereo, and walkman, no audio but the display showed the playback countdown. Put the Cd's in another PC & got audio...what's up?...kpl

ozpeter · Location: Australia

Something wrong with finalisation it would appear? Bizarre. I've experienced CDs with a TOC but no files, such that if put in a CD player it looks OK but there's nothing actually to play. Another reason to trust a burned CD no further than you could throw it. (Bad example, you could probably frisbee a CD quite a distance. Any claims for fame in that department here?)

- Ozpeter

RossW · Location: USA

kpl--

CD readers in PCs generally accept the widest variety of media. For example, I have (accidentally) burned music onto CDRW discs, and found them unplayable on audio-only systems, especially older units. Wouldn't even play on an "antique" PC CD-ROM reader. Could the choice of media be the cause? That might include discs not up to 40x burning, CDRWs, etc. Have you tried burning at a slower speed?

Andrew Rose · Location: United Kingdom

If you made a data CD by mistake you'll find an audio player will show the time and a single track but will mute the sound. PC software should be intelligent enough to interpret WAV file(s) and play them directly. This is one possible explanation...

pwhodges · Location: United Kingdom

Arrowkey do programs for checking CDs; one to compare disks (made from the same files, say) to see if they are identical, and another which turns the error-checking of the drive as low as it will go and then counts the errors - not quite measuring BLER but still useful.

Paul

tptman · Posted - Fri May 09, 2003 10:21 am

I've had problems doing multiple burner runs in Nero. Using my old 16x yamaha, which I'd had zero problems with, and the new 40x Yamaha, which has had zero problems on its own, I did a run of about 120 CDs. I usually spot check every tenth cd or so, and I didn't find any problems doing that. But I got about a dozen or so complaints about scratchy CDs or blank CDs from customers (all of which I promptly replaced, of course). So maybe the multiple burner thing isn't as bulletproof as Nero usually is.

cooknkpl · Location: USA

Naaahhh...it's much simpler than that - I'ma idiot. I burned them as data files. Works now...thx for the input...kpl

ozpeter · Location: Australia

Just another dataday problem... we all have these moments...

- Ozpeter