crazzell
crazzell's picture
Offline
Last seen: 12 years 8 months ago
Joined: Dec 7 2005 - 5:08pm
What;'s up with "jitter" on pressed music CDs?
dcrowe
dcrowe's picture
Offline
Last seen: 4 years 11 months ago
Joined: Sep 1 2005 - 9:39am


Quote:
Hi all,

I was wondering if I could get some authoratiative answers to the following question:

Does sloppy manufacturing of a music CD result in degraded sound quality due to "jitter".

Much internet chatter from respected sources says that it does. See for example: Roger Nichol's story

However, engineering and signal processing knowledge tells me that the word clock in a CD player can only be minimaly impacted by read-channel jitter in the transport. Mechanisms such as power supply modulation are usually described as the main issue.

Is it not true that XRCD's value propostion is largely based on this phenomenon? Has this ever been verified, by measurements rather than by opinion?

Love to hear your thoughts,

--dsdreamer.

dsdreamer,

Certainly poorly manufactured, or well manufactured but dirty or damaged, CD's can result in data errors that may differ depending upon the machine that reads the data from the CD. The built-in error correction can then either entirely or partially fix the problem (or fail to fix it altogether). The action of the error correction circuits takes a finite time and might in some cases result in a jittery analog signal. I would not lump all of the audible differences between different CD's in different players under the term "jitter" (phase error). Amplitude errors are also possible. Some CD players may handle a particular disc better than others, and it is not necessarily true that more expensive players are always better than less expensive ones. Most clean and undamaged CD's will sound good on most modern players. When it doesn't sound good, the CD may be the root cause if it is below standard, but the "jitter" (if any) primarily occurs in the player.

Devon

Editor
Editor's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 1 2005 - 8:56am


Quote:

Quote:
Does sloppy manufacturing of a music CD result in degraded sound quality due to "jitter".

Certainly poorly manufactured, or well manufactured but dirty or damaged, CD's can result in data errors that may differ depending upon the machine that reads the data from the CD. The built-in error correction can then either entirely or partially fix the problem (or fail to fix it altogether). The action of the error correction circuits takes a finite time and might in some cases result in a jittery analog signal.

I think you are introducing a different issue. The original poster was referring to time-base error, ie, errors in the length of the pits, that propagates through to the clock controlling the player's DAC. My Plexstor PC drive came with a utility that allows the user to measure this error and some CDs indeed have much worse time-base error than others.

The question then becomes: By how much does the design of the player's data-recovery circuitry suppress this error? We published an article on this subject in 1994; see http://www.stereophile.com/features/827/ .

John Atkinson
Editor, Stereophile

dcrowe
dcrowe's picture
Offline
Last seen: 4 years 11 months ago
Joined: Sep 1 2005 - 9:39am


Quote:

I think you are introducing a different issue. The original poster was referring to time-base error, ie, errors in the length of the pits, that propagates through to the clock controlling the player's DAC.

The question then becomes: By how much does the design of the player's data-recovery circuitry suppress this error? We published an article on this subject in 1994; see http://www.stereophile.com/features/827/ .

John Atkinson
Editor, Stereophile

That is a very good and informative article, although the terminology statement that this effect is "not a data error" is unexpected. I suppose that is meant to say that the correct bits are present at the wrong time. I would include both the bits (amplitude) and the timing (phase) in the "data" set.

In any terminology convention, yes, the quality of the CD and of the player are both very important.

Jim Tavegia
Jim Tavegia's picture
Offline
Last seen: Never ago
Joined: Sep 1 2005 - 4:27pm

Forgot to long in. The above post is mine. Jim Tavegia

Jim Tavegia
Jim Tavegia's picture
Offline
Last seen: Never ago
Joined: Sep 1 2005 - 4:27pm

I've gone back and done some additional reading including the digital domain site to "go back to school" on digital. It is no wonder we do not have perfect sound forever...yet!!!

I am curious about the "shape" of the pits and do understand about the 9 different pit lengths corresponding HF data. Is the leading edge and following edge "shape" of a pit an issue at all?

It would seem that a "curve" or "rounded end" shape would be or possibly could be more detrimental to having basically a long rectangle shape with squared ends in terms of laser tracking accuracy. Without understanding the actual "burning" of the pits in the clear poly disc and subsequent shape of the lands, it may be impossible to have the pits be close to a 90 degree-angled edges and ends. Again, this "shape" may not be important in the retrival of data anyway.

It struck me that the discs Bob Katz talked about from music clubs vs retail that sounded different measured the same in the testing they did on bit accuracy, but sonic differences did remain. Bit depth also seemed to be an issue.

I have also found that some CD-Rs with differing reflectivity will not play back in certain players I have had at the house. Most of them have some version of the "blue/green tint" CD-rs. I've landed on Sony's for now and had great luck with no rejects. but, I digress. Any info appreciated.

crazzell
crazzell's picture
Offline
Last seen: 12 years 8 months ago
Joined: Dec 7 2005 - 5:08pm

Thanks for the reference to the early Stereophile article on the subject, although I had already found it via Google. The article certainly makes a convincing case that the quality of the pressing can vary. What is less clear is how important this can be for sound quality (the original subject of my post).

I for one, do not find the following conclusive:

Quote:
These sonic differences may be regarded as slight or even imperceptible to the casual listener through a low-quality stereo system. But to the audiophile with a keen ear, an open mind, and a high-resolution playback system, the differences are musically significant.

So apparently, one has to have golden ears, an open mind and very good equipment to hear this. Are we sure the king is wearing any clothes at all?

So the original purpose of my post was to see if we could go beyond the anecdotal, heresay evidence to something closer to a repeatable experiment, preferably a quantifiable measurement. To me it is less a question of how much can a piece of equipment reject the jitter in the signal from the CD transport, but how it could fail to reject it.

Suppose a pair of 16-bit DACs of high linearity has a crystal controlled 44.1kHz word clock applied to it, and pair of FIFO buffers is used to supply LPCM samples. Does the time of arrival of the words to the input of the FIFO have any bearing on the delivered sound quality (assuming there is no buffer overrun or underrun, and all LPCM words are present and correct)?

Would it be possible to come up with test CDs that contain the signal needed for the Miller jitter test, which have been deliberately doctored to simulate bad mastering? If so, could we actually measure the extent of the supposed coupling of media induced jitter to the audio output of a CD player?

--dsdreamer

Jim Tavegia
Jim Tavegia's picture
Offline
Last seen: Never ago
Joined: Sep 1 2005 - 4:27pm

In the Roger Nichols 2 v nat the sound was easily discernable on his portable CD player through his Sony 7506's. When he got home and played the discs on his $5K player that I believe had a ram buffer and reclocked the data stream it sounded much better.

I have written DenonDigital here near Atlanta to see if I can do a free-lance article on their pressing plant in Madison, GA and get their engineering take on what is happening on these 9 distinct sized pits that make up our CD Red Book discs. I am more than curious now about why disc vary from manufacturing plant to manufacturing plant. Maybe they'll be glad to share their insight with the audiopile community without giving away any trade secrets. This Madison, GA plant is well regarded.

Log in or register to post comments
-->
  • X