I also commend the writings of Dan Lavry to those who want to sort out the science from the subjective opinion from the unsupportable and scientifically inaccurate claims.
Lavry, as many probably know, was a pioneer in sampler design, working first in medical equipment design in the 70s (a field where equipment marketing claims tend to be subjected to extremely rigorous testing processes) and moved into audio at the dawn of the (commercial) digital audio age.
Many may also be aware that he was intimately involved with the design of the first Apogee converters.
And, of course, his very expensive converters are among the most coveted in the audio world.
Lavry seems to be an inveterate straight-talker who does not suffer fools easily.
(I was personally -- if somewhat gently -- scolded in these very BB pages, causing me to say something to myself like, "Who's this Dan Lavry guy, anyhow, and what makes him think he's qualified to tell ME anything?"
When I researched the answer to that question, I decided that, yes, indeedy, he was very much qualified to tell me what was what and plenty more besides.)
Anyhow, he's got a forum at PSW that, unfortunately, has been compromised by PSW censoring/deleting at least one thread that ruffled the feathers of a large manufacturer whose reps had made some seemingly inaccurate claims: http://recforums.prosoundweb.com/index.php/f/38/15450/
Still, it has some lively -- and sometimes quite funny -- threads.
After the censorship/thread deletion issue, Lavry moved some of his discussion to the BB on his own company's site: www.lavryengineering.com
Lavry wrote what many feel is an excellent guide to sampling theory and its implications for us recording
engineers which is also available on his website along with some other articles and white papers.
Here's a bit pertinent to today's discussion. Lavry [who seems to be a hasty typist] is answering a question
in his PSW forum:
You are correct to say "Dan Lavry believes that external clocks can RARELY (if ever) make any system sound better than using a device's internal clock".
I would go much further then that: it is not about what I believe, it is about engineering.
I believe that people hear a difference between external clock boxes, and between internal and external clocking. Any and everyone is entitled to their subjective likes and dislikes.
It is a fact that jitter creates rather unpleasant distortions (non harmonic) and in most cases it increases the noise floor.
Yet, one can decide they like more noise, harsh distortions... all that is rather surprising, but not arguable because it is subjective.
But those that push the notion that an external clock will improve the sound, are not just saying that the sound will be altered, that some people will like it, others will not. The claim here is that there will be an improvement. That claim really translates to "more jitter is good". Of course no one is ready to state publicly that "more jitter is good", and most often the proponents of external clocking will go out of their way to convince you that they are providing with less jitter.
Less jitter the internal clock? A good internal clock provides a lot less jitter then the best of all external clocks.
In the case of internal clocking, one relays on a local fixed crystal aimed at keeping the frequency constant. That crystal circuit would be positioned near the AD, to minimize interference.
For external clocking, one may Begin with an equal quality crystal circuit inside the clock box, but the signal has to go through a driver, cable, receiver and yet another crystal circuit, all with some frequency locking circuitry...All of that is a lot more jitter.
So in order to overcome the above stated reality, there are stories floating around, that a clock box sends signals that improve the AD clocking. This is of course impossible, because one can not improve without knowing what needs to be improved. The clock box gets to send the signal to an AD, but it is a one way communication, and the AD does not tell the clock box what it is doing. In other words, a doctor can not heal a patient without seeing talking or having any communication with the patient.
The clock box can not send a better word clock then a simple square wave with the lowest possible jitter. Putting the same clock inside, eliminates many of the elements piled up in series between the clock source and the AD.
Of course there are times when one needs to use external clocks, and in such cases, a mediocre clock (say 25ps jitter is about as good as a theoretical 0ps clock, because by the time the clock shows up the AD most of the performance will be determined by the ability of the receiving AD to reject the jitter. Given the same quality level of design, an internal clock will be better, so external clocking it is a trade-off, not an improvement. With good gear the damage is small, but it should not be viewed as an improvement.