How many forms of Jitter can you identify?

There is a good summary of ten potential contributors to jitter in this paper by Steve Nugent of Empirical Audio, about a third of the way down the page.

Note that items 2 and 3, which pertain to the pits on the cd itself, and the electromechanical systems that spin the disk and read the data, are "not an issue ... for newer players that completely buffer the data at high-speed from a CDROM reader to a memory buffer."

IMO, assuming everything involved is reasonably well designed, by far the most major contributor to jitter is the S/PDIF or AES/EBU interface that is commonly used between source and dac. Making the timing that is applied to the dac chip itself ultimately dependent on a clock supplied by a different component, which is multiplexed together with the audio data into a single signal that is then transmitted between components, the clock subsequently having to be extracted from that signal, is an inherently compromised concept IMO. Several different items in Steve's list, btw, pertain to that interface.

At most it should a good cleaning product that could allow the laser to read the disk better eliminating any error correction from being necessary.

Assuming the player is reasonably well designed, error correction is not an issue, because it results in bit perfect recreation of the original data. An issue only arises when the data being read is so badly messed up, for instance due to a severe scratch on the disk, that the errors cannot be corrected. In that case error interpolation, or conceivably outright muting, will occur, which will affect sonics as it represents an approximation at best.

Best regards,
--Al

The CD disk is made of plastic & some organic dies that hold the digital data. The plastic structural material is not non-magnetic. Thus, when the laser light hits the plastic underside to read the data, over time the plastic gets mildly magnetized. For example, how many times have you experienced plastic packaging material "stick" to your fingers due to electro-static charge on it???

Bombaywalla, I'm not sure I follow that. Static electrical charges and magnetization are two different things. And why would exposure to light cause either one?

Good point by Shadorne. Seconding Bombaywalla's comment about that point, it should be kept in mind that the extent to which that effect may be significant will be highly dependent on the design of the particular component.

Best regards,
-- Al

Bombaywalla, thanks for the explanation.

The logical follow-up question would be how significant is all of this from a quantitative standpoint? The data on page 3 of the reference you provided addresses that question, although the various terms and quantities are not clearly defined. My interpretation of it is as follows:

1)Looking at the first two of the four tabulations, on average there were 0.00 uncorrectable errors in each 1 second block, both before and after demagnetization.

2)Looking at the last two of the four tabulations, the WORST CASE (across some unspecified number of disks, trials, etc., and averaged in some way) was that there were 0.22 uncorrectable errors per 1 second block, or in other words 1 interpolated sample approximately every 5 seconds (which would comprise approximately 450,000 16-bit samples for the two channels).

And what they are saying is that demagnetizing would reduce that worst case number of interpolated samples by about one-third.

The bottom line on audibility, IMO: I'll give it a "definite maybe," but from a personal standpoint I don't consider any of that to be sufficiently compelling to inspire me to go out and buy one.

Best regards,
-- Al