Error rate in commercial CDs

[Stig Ø]

Apologies if this is posted in the wrong place. Does anybody have any information regarding error rates in current commercial CDs vs e.g. twenty years ago? I still buy cds, and I find I have more errors with unplayable cds now than I did before.

[Hugh Robjohns]

More likely to be a tired CD player... but I'll run some comparison checks using Plextools later and report back.

[Folderol]

Are they still putting in deliberate errors in an attempt to stop people ripping them on computers?

[Rich Hanson]

If they are then they're doing a pretty poor job of it as I have successfully ripped every single CD in my collection (apart from a couple of damaged ones).

In answer to the question, newer CDs don't appear to have more errors in my experience. Certainly not noticed anything significant when using EAC to rip.

[Hugh Robjohns]

Just ran four CDs through Plex Tools XL using a Plextor Premium2 drive, two purchased within the last month and two from the early 90s.

Old 1: C1 average =6.9/sec C2 =6 in total CU =0

Old 2: C1 average =7/sec C2 =129 in total CU =0

New 1: C1 average =1.5/sec C2 = 0 in total CU =0

New 2: C1 average = 0.9/sec C2 = 0 in total CU =0

These figures are exactly what I'd expect.

The two older discs both have higher C1 error rates, and the C2 errors occurred in bursts where the disc surface was damaged. But all errors are correctable, so there was no problem in playing or ripping those discs.

The two newer discs have obviously substantially lower C1 rates and no C2s at all, again exactly as expected.

So to answer your question, I see no evidence from that quick poll that modern pressed discs have higher error rates or are more problematic in any way.

[wireman]

Not sure what tests will prove. It seems I have PlexUTILS (forgot about that)

So trying a couple of old CDs
BER test

Code: Select all

                 C1 max        C1 ave   C2 max      C2 ave
HFN003             41          1.5          93          0.1
Brothers in Arms   7011         5.8         7185       4.6

The error rate on BiA started very bad (in the hundreds) and quickly droppend through the range bringing down the average as it went.

I have no idea what these numbers mean.

There was someone in the US who might have known the answer to this question but I thnk he shut up shop some time ago: https://web.archive.org/web/20110817123 ... index.html

[Hugh Robjohns]

The error correction system in the CD format is basically a three stage system. C1 and C2 represent the first and second layers of error detection/correction. There will inevitably be errors on the disc -- that was expected in the design and that's why it has an error detection/protection system.

So you would expect a fair few C1 errors. They are inherent in the disc manufacturing process. The average rate is important as the system can only handle so many at any one time, but as long as it's less than an average of 220/sec (in any 10 second window) there's no problem. So in my test discs C1 rates averaging under 10/sec are of no concern, although an average rate of 2/sec or less would be my preferred target.

Any more complicated errors that can't be detected/corrected by the C1 stage are passed to the C2 part of system. Again, C2 errors can be corrected so their presence is not a major cause for concern, although the number of C2 errors is more indicative of the state of the disc.

Ideally there should be no C2 errors at all, (as the two new discs in my quick check above showed), but a few won't be a problem, and they are usually the result of disc surface damage -- a careful clean and polish will usually remove them!

The third stage in the process for audio CDs is 'interpolation'. If an uncorrectable error is found (flagged as CU in the test above) then that sample is discarded and a simple average derived from the preceding and following data.

This is clearly a bodge, but it is deemed an acceptable one for consumer audio playback purposes. To be fair, it works fairly well because of the inherently cyclical nature and low frequency of most audio waveforms (especially the louder ones!), so the averaged sample recreation is rarely far from the real amplitude, and the momentary noise/distortion/click this process typically introduces in generally inaudible as long as the rate of interpolations is very low.

Obviously, this interpolation technique is not acceptable when dealing with raw data so CD-ROMs (and other disc formats) use an additional layer of error correction for better data security.

[Stig Ø]

Thanks a lot, Hugh. It seems like your cds don’t show any signs of increased error rate, although our respective sample sizes are small... The reason I asked is that I’ve had some errors show up on two different CD players, but with the exception of one disc that must have been damaged, I’ve been able to rip them all.

Thanks for the time and effort.