Re-clocking digital audio? Benefit or snakeoil???

[tbewick]

I had a dealer mention to me yesterday that re-clocking digital audio signals provides a huge improvement in SQ. Apparently, there is a fellow around here modding very high end CD players (Moon, etc) with external clocks. Anybody ever hear of this? Is this a valid way of improving the digital processing of the audio signal?

The internal clock of a good quality, modern CD player, combined with the jitter rejection and performance of the CD player's DAC, is probably adequate (see [1], [2], and [3] for info on jitter and jitter audibility). Professional studios may use an external master clock, which can be used to prevent excess jitter from building up between digital audio devices. You can have this trouble in large digital audio installations [4].

[1] Dunn, J. (2003). "Measurement Techniques for Digital Audio". Jitter audibility is briefly discussed on page 34. Audio Precision Application Note #5, Audio Precision, Inc. USA.
http://ap.com/library/technotes.htm
[2] http://forums.audioholics.com/forums/showthread.php?t=4547
[3] Ashihara, K. et al (2005). "Detection threshold for distortions due to jitter on digital audio", Acoustical Science and Technology, Vol. 26 (2005) , No. 1 pp.50-54.
http://www.jstage.jst.go.jp/article/ast/26/1/26_50/_article
[4] 'Engineering Guidelines – The EBU/AES Digital Audio Interface'. Page 43-44. John Emmett, 1995. European Broadcasting Union.
http://www.ebu.ch/CMSimages/en/tec_AES-EBU_eg_tcm6-11890.pdf

 

[mtrycrafts]

IMHO reclocking the converter is like upsampling a cd player, theoretically it'll sound better, but 98% of average unmutated humans will not hear the difference.

I would think that clocking issues cause jitter. But, surprisingly, jitter is difficult to detect unless it is really gross, by 100% of subjects

http://www.jstage.jst.go.jp/article/ast/26/1/50/_pdf

 

[stratman]

Companies have been doing this since the early to mid-90's. Check this out:

http://www.monarchyaudio.com/DIP4.htm

You can turn off the music at the bottom of the page!

Or this:

http://en.wikipedia.org/wiki/Audio_Alchemy

They were doing it more affordably than anyone else.

 

[MDS]

With digital audio, timing is everything...BUT it is not so grave as the 'modders' would have you believe. With s/pdif the clock signal is inherent in the data itself. You have to consider that even if the timing is off a small bit, how audible is it?

Everyone loves to talk about 'jitter' but in most cases it is a non-issue. If you think you can hear the difference between 30 and 50 picoseconds of jitter then no digital audio system in existence can satisfy you. For those not familiar with the terms, a picosecond is 1 trillionth of a second!.

For comparison, it has been shown that most people cannot distinguish between two pure tones at different levels unless they differ by at least 6 milliseconds. 6 milliseconds is two orders of magnitude greater than the typical jitter performance of even the cheapest cd players...so take from it what you will...

 

[mtrycrafts]

With digital audio, timing is everything...BUT it is not so grave as the 'modders' would have you believe. With s/pdif the clock signal is inherent in the data itself. You have to consider that even if the timing is off a small bit, how audible is it?

Everyone loves to talk about 'jitter' but in most cases it is a non-issue. If you think you can hear the difference between 30 and 50 picoseconds of jitter then no digital audio system in existence can satisfy you. For those not familiar with the terms, a picosecond is 1 trillionth of a second!.

For comparison, it has been shown that most people cannot distinguish between two pure tones at different levels unless they differ by at least 6 milliseconds. 6 milliseconds is two orders of magnitude greater than the typical jitter performance of even the cheapest cd players...so take from it what you will...

2 orders? millisecond is 1/1000 of a second, right? That would be 10^-3 pico is what, 10^-12?
250, 000 pico sec of jitter is not audible, according to recent testing:

http://www.jstage.jst.go.jp/article/ast/26/1/50/_pdf

 

[MDS]

2 orders? millisecond is 1/1000 of a second, right? That would be 10^-3 pico is what, 10^-12?
250, 000 pico sec of jitter is not audible, according to recent testing:

http://www.jstage.jst.go.jp/article/ast/26/1/50/_pdf

Yes, it is way more than 2 orders of magnitude - it's actually 12. So I left off the leading digit.

 

[sparky77]

I would equate this debate to someone telling the difference between a cd player with 8x oversampling and a 1 bit da converter to one with 64x over sampling and a 16 bit da converter, very very very few people will ever be able to tell the difference.