So I took an Audio CD and a blank, stuck both in a philips dual-deck CD recorder (the consumer kind) and made a copy. Then loaded both disks in my PC and captured 10 copies of the CD image of each disk on to harddisk at high speed. That resulted in 10 significantly different files for each disk. Captured again at single speed (That did take about a day indeed :) and had 10 identical files for the original and 10 identical for the copy, but differences between the original and copy... SCMS!!??!! So I repeated the copy using a machine without SCMS, a stand-alone duplicator I use for software mastering... After the read and compare test, again two groups of 10 images were read at single speed. 10 identical images for the original (Unchanged from the previous set of 10), and 10 identical images for the copy, but the original and copy were significantly different. The final test I did was to write one of the original images back to CD using a CD-R in the computer, at double and single speed. These, again, were read back in the computer 10 times and compared. Both would get groups of consistant readbacks at single speed, but the images were different from eachother and from the original CD. The single speed writes were closer to the original than the double speed writes. So, I burned some more copies at single speed and did the same readback test. All copies were slightly different. All copies were very close to eachother, closer to eachother than to the original CD. Finally, I used a different CD drive to read back the original and the copy... This is where it got interesting. The copies read back binary identical, ie. a copy read on the HP it was written on was identical when read on the HP at single speed and on the companion Mitsumi drive in the same machine. However, the original disk did not work this way! It was consistant on the Mitsumi, and it was consistant on the HP, but the two sets of images differed.... I found that VERY odd!! Anyhow, all in all it resulted in over 60GB of CD images, and actually analysing that much data proofed to be hard, so I did not actually try to correlate the differences. What it proofed to me though is that there is a difference. As to the sound quality of the copies, which is what I was really interested in: The copies made on the PC at more than single speed and in the stand alone duplicator used for software sounded imperfect, clearly audible audio defects like ticks and static. The Philips made copy sounded best, I could not hear a difference. The computer made copy at single speed, to me, sounded almost as good as the original, there were some minor audible differences in the form of static. I used Easy CD Creator Pro to capture CD images and burn CD's on the computer side of things. That was a very long way of saying: No matter what I did, the copies were not identical! That was not what I expected...
CD Copies...why do they sound worse?
I had a theory that I haven't discarded yet that not all CD blanks are equal in terms of composition. Yes, they all are made of aluminum and polycarbonate, and when you burn a CD you are creating small holes, or dents in the blank. There is the red book standard that must be adhered to, but as in anything else, I'm sure there are better grades of aluminum and poly available, you get what you pay for. Since the laser reads the digital stream by optically scanning the surface of the CD and interpreting either a one or zero, you'd think it's a go/no-go operation. The original and copies do not sound the same, even to the uncritical ear. I thought for a while it may have had something to do with the relative quality of the CD blanks I was using to copy, in other words, the pressing plants simply use a better grade of master CD's. My friend has a contact and we were able to acquire bulk CD blanks from Saturn Disc that makes CD's. No difference, copies still aren't right. I guess we can eliminate the CD blanks for now. Here's where things get a little outside normal thinking in my twisted logic: we know there are error detection and correction schemes used in intrepreting the data on the CD, employed when the bit being read isn't immediately recognizable to the player. Is it possible the home-made copy that was burned using a cheap consumer grade burner, contains more errors? Are the pits burnt in the CD either irregular in shape or depth? Does the laser in these consumer grade CD burner introduce errors? If so, the EDAC is pretty busy, and doesn't always get it right, which would explain a general lack of quality due to latency delays in the data stream while the EDAC does it's work, and in the process is bound to mis-interpret zeros and ones, there is no 100% accurate EDAC. To me, this is a good place to start in terms of understanding the obvious differences in sound quality.