Jitter Audibility -- Robert Harley and Keith Johnson comment

Tom Martin -- Sat, 05/24/2008 - 09:06

Robert Harley posted this a few days ago, and it deserves to be in this section too:

The current issue of TAS (June/July, Issue 182) includes a letter from reader David Sanford questioning whether slight timing errors in digital-to-analog conversion (jitter) is really audible. The letter was sent in reponse to my review of the Esoteric G-0Rb rubidium clock that improves the timing accuracy of D/A conversion. Mr. Sanford found it hard to believe that a shifting of the sample timing (jitter) by several picoseconds would be audible. In my reply, I mentioned that Keith Johnson once told me he could hear the difference between 8 and 15 picoseconds of jitter. (Keith Johnson is the co-inventor of HDCD, recording engineer of all the Reference Recordings projects, designer for Spectral, and has perhaps the most insightful audio mind of anyone working today.)

Keith read the letter and my reply, and sent me this analysis:

David referred to perception of arrival times of sound and to distortion from jitter in a digital system. These are different things not best compared on a pico-second time scale My comments were about perceptual consequences when timing errors are introduced to digital conversion. Sigma delta methods, particularly one bit, introduce much more destructive artifacts. We hear the consequence of jitter rather than jitter by itself.

The time scale argument is interesting:

We can move forward one loudspeaker of a stereo pair 6 inches – an action creating roughly one half millisecond difference in arrival times to create audible consequence.

When delay or arrival times modulate as fast audio range jitter, one finds reasonable perceptual limits require at least one hundred fold smaller moving displacements. Early magnetic recorders demonstrate this disturbance as their unsupported tape paths were long and produced high-speed irregular motion that imparted a sliding paper character to string instrument sound. Manufacturers introduced idler wheels and loop drives to remove sub micro-inches of this scrape flutter and microseconds of modulation noise. We hear the improved clarity along with reduced coloration in the historic development of phonograph records.

At a million times smaller, pico-second changes to sound paths would pose nano-inch listener and speaker constraints – clearly well beyond reasonable contemplation of audibility.

We tend to think of digital processing as on – off and nothing in between. However, modern sigma delta DACs are configured with analog comparators and filters that operate with an environment of continuously changing voltages or currents. When jitter is introduced to clocks or the internal timing of the DAC is not good, the outputs and decision states from these analog functions become cast in stone before or after the time when responses would have been correct. Now a tally or sampling command produces an artifact or error. The jitter not only propagates as itself to cause some distortion but also produces conversion error and both are memorialized in filters involved with ongoing samples. This compound degradation increases at very high sampling rates a factor where larger performance specification numbers may not always provide better performance. We observe spurious signals, cross modulation, pulse response widening and signal coherent noise modulation – things that have unpleasant names and unpleasant sonic consequence.

I think the example of “audible 8 to 15 pico-second jitter” comes from describing a transport evaluation platform at Spectral. Like many digital playback devices, the fixture has multi-bit sigma delta DACs that require very low jitter to work properly. Hence, we provide direct crystal clocking and other technological supports that one should find in any serious product. We can degrade a synchronization and re-clock circuit intended to isolate optical tracking and servo activity of a CD transport. Then, an RF analyzer shows presence of new jitter related to CD condition and oscilloscope traces reveal moving peaks and spikes riding on clock waveforms. An audio analyzer will display upper band noises as well as widened base lines for continuous sine wave spectra. These artifacts might appear inconsequential except that we must consider the averaging nature of analyzer measurements. We have unsteady nasty waveforms that require experimentation and interpretation to reveal the full damage from the tracking interference. A slight increase of noise is observed and careful manipulation of test signals reveals rapidly changing spurious error in the parts per million – the possibly perceptual category. Similar examination methods reveal peak jitter to be well over 5 times the RMS or specification sheet type values produced from the analyzer. Consequences are audible though more so to others at Spectral than to me.

Keith O. Johnson
Reference Recordings

Robert Harley -- Sat, 05/24/2008 - 11:17

I'd like to add that Keith Johnson has remarkable insight into correlating measured distortions with specific sonic degradation. He talks in detail about this approach in my interview with him, which will appear in an upcoming issue of The Absolute Sound.

Robert Harley

Robert Harley -- Mon, 06/09/2008 - 09:46

TAS reader David Sanford sent me this e-mail in response to my reply to his letter in TAS 182:

Thanks for publishing my letter in the June/July 2008 issue.
I don't really question the audibility of jitter. However, I have to be skeptical when someone tells me that timing variations between clock cycles of 7 picoseconds, (which is 7 billionths of a millisecond) are audible. I need a bit more scientific evidence than what one person claims he can hear.

I have been skeptical of people's perceptions ever since my high school physics lab. The teacher passed around two objects and we were to write down which one we thought was heavier. The first object was a small clear glass jar containing mercury. The second object was a larger clear glass jar containing some solid objects. Both jars had strings attached and each student was allowed to inspect the jars simultaneously any way he liked including holding them by the strings. Just over three-fourths of the class wrote down that the smaller jar containing the mercury was heavier. After we all gathered around and weighed the jars we found that the larger jar in fact weighed twice as much as the smaller jar!
Why were most of the students wrong, when the weights of the jars weren't even close? I don't know, but I realized that day that you should question the accuracy of human perceptions. In high end audio logic, however, many would argue that the fact that the larger jar actually measured heavier isn't important. What is important is that the students perceived the smaller jar to be heavier and therefore it must be.

David Sanford

discman -- Mon, 06/09/2008 - 10:36

If I understand Johnson's point, then Sanford (who may not have seen Johnson's message) misses it. Sanford, I believe, is thinking analog. But this is digital. In digital signals, time has meaning that is different than what it means to humans in the analog domain. Thus, we are not talking about direct human perception of picosecond time, but the digital meaning of picosecond time.

If for example, we had a time-based encoding system operating in the terahertz range, then we might assign different frequencies to different digitally encoded numbers. A difference of even 1 picosecond in such a system would change the digital representation of the signal. A 7 picosecond variation might completely change the signal (close to 100% distortion, perhaps?).

Now in the case of CD, the frequencies are much lower, and I don't know enough about the encode/decode schemes to know whether the data error introduced by 7 picosecond timing variations is material. The point is, there are two issues: 1) is there a data error introduced by jitter and 2) is the error perceivable. The former seems at least plausible, and until we know the type and magnitude of the error I don't see why we should assume it would be hard to hear.

discman -- Mon, 06/09/2008 - 10:40

Sanford, btw, may make another interesting point (unintentionally): his student buddies were very accurate perceivers (of density) and bad explainers (their theory was out of alignment with their perceptual system). We should not confound these two.

dazzdax -- Sat, 06/28/2008 - 17:47

It seems to me (but I could be wrong) that jitter is almost a selfreplicating error within digital signal conversion and that this selfreplicating effect is potentially more present in delta sigma conversion with very high oversampling rates. If this is the case then signal processing using DSD is very tricky and requires the highest precision and other counter measures to prevent high incidence of jitter. It would be interesting if another expert who is as knowledgeable in the field of digital conversion as Keith Johnson (Ed Meitner?) would also give his response.


Eureka -- Sun, 06/29/2008 - 22:52

Perhaps I have missed it, but I don't think so. I am thinking digital, not analog. To understand my point I think you have to read Robert Harley's article on page 120 of the March issue of TAS and my letter printed on page 8 of the June/July 2008 issue.

When I reference human hearing being able to detect differences of time delay in milliseconds, I am trying to put into perspective the ability of humans to detect changes in sound. In this case it is audible if the sound arrival time varies by more than 1/1000 second. A picosecond is one billionth of a millisecond. I realize that jitter is not time delay, but inaccuracies of the sample point being taken of the analog waveform. This means that since there is a slight difference in the sample point on the wave form that the waveform amplitude will be slightly incorrect for that sample. This slight variation of course will result in a waveform reproduction that would be slightly different than if a 100% accurate clock was used. Samples on a CD are taken 44,100 times per second. What I question is whether shifting the sample points by +/- 7 to 15 picoseconds will make an audible difference. I find it hard to believe that moving the sample point by such a small percentage of the sample size (on the order of one billionth), would make an audible difference. Mr. Harley stated that "it's an amazingly small timing variation, but one that our aural decoding system can easily detect." What I surmise from Keith Johnson's response is that the final error is in "parts per million -the possibly perceptual category." Rather than say "easily detected," I would say that it is a variation that some people might be able to hear some of the time.

I am just an audio enthusiast that happens to be an engineer, and don't claim to know much about electronics engineering. I have all the respect for Mr. Harley and Mr. Johnson, but when I read the article, I just found it hard to believe. This is why I was asked if any scientific testing of it's audiblity has ever been undertaken.

Concerning my physics lab results, the students were asked to choose the heavier object and not the most dense object. Their perception was incorrect. I firmly believe that humans can fool themselves to believe they hear differences when in fact there are none, and that small distortions can be claimed to be audible when they are not. Is 7-15 ps of jitter audible? I still don't know, but I would certainly be interested in hearing a scientific demonstration of it.

Eureka -- Mon, 06/30/2008 - 13:39

I just wanted to close this up. I again reviewed jitter in Robert Harley's TAS article and his "The Complete Guide to High End Audio."

Perhaps this subject is just too complicated for a layman, such as myself, to understand. His book states: "the specific pattern of ones and zeros creates a clock signal at the data rate read from the disc, 4.3218 million bits per second." Everything here seems to be in millions of bits per second, and the claim is that an 8 picosecond variation in the clock can be easily detected. Given that a millionth is 1x10-6 and a pico is 1x10-12, I'm just amazed that human hearing can detect it at all, let alone easily.

Thanks to Robert Harley and Keith Johnson for their time, and for everyone elses input. It has been most interesting.

sdfewf (not verified) -- Fri, 03/27/2009 - 02:00

The teacher passed around two objects and we were to write down which one we thought was heavier. The first object was a small clear glass jar containing mercury. The second object was a larger clear glass jar containing some solid objects. Both jars had strings attached and each student was allowed to inspect the jars simultaneously any way he liked including <a href="http://www.roomfurniturechina.com">bedroom furniture</a> holding them by the strings. Just over three-fourths of the class wrote down that the smaller jar containing the mercury was heavier. After we all gathered around and weighed the jars we found that the larger jar in fact weighed twice as much as the smaller jar!

Anonymous1 (not verified) -- Wed, 11/25/2009 - 00:18

Sorry but jitter in current gear is not audible, 30,000ps for sighted and 250,000ps for blind.

Steven Stone -- Wed, 11/25/2009 - 14:49

 Quote from the text linked above "Experiments were carried out in the listening booth or studio that each listener had offered. The examiner only brought there a personal computer with a digital audio interface and a mouse and each listener provided his or her favorite DAC, amplifiers and loudspeakers."
Knowing what we know about USB DACs and interfaces between portable computers and DACs, it's no wonder that jitter differences were masked by this inferior and highly variable interface methodology. Also these tests were done in 2005 when the quality of USB DACs was far inferior to today.

Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications

Keladrin -- Mon, 04/23/2012 - 10:23

Hm, an interesting discussion.
Steven - you are right but Jitter is more likely to be masked by the loudspeaker accuracy as this seesm to be the weakest link in the audion chain, as explained in the paper. I assume the reason why listeners were allowed to bring their own DAC's was tio cancel out the effect of differing DAC's. If the DAC has some significant effect on the signal perhaps we should not be worrying so much at the Jitter, than the DAC. Presumably you could reanalyse the test data to see if the different DAC's had a significant effect on the outcome also.
Hm, so why doesn't Robert Harley interview the writers of this paper of at least feature it in Stereophile instead of talking to a guy who has a vested interest in a company who is selling special low jitter recordings (even if he has some expertise)? Surely the analogy is talking to a health food store salesperson rather than an independant medical doctor when finding some cure to an illness?

earwaxxxer -- Thu, 05/17/2012 - 11:33

Very interesting stuff for sure! Highly debatable as well, along the lines of different sounding USB cables etc. I do believe in our perceptions in the above mentioned instances, as difficult as it may be to quantify.
I was experimenting with a new interconnect I got yestereday. Switching back and forth between my pure silver 'Silver Bullets 4' and my new JPS Labs 'UltraConductor 2'. I switched positions back and forth on my preamp. A clear difference in sound on weather the JPS was in the DAC to pre or pre to amp position. Of course we are talking analog here, but these are both .5m high end cables and its 'just' wire. So, I dont see how the same logic would not apply in digital realm as well. There are going to be differences perceived when digital is processed in different ways, yeilding different end products.

Maggie MMG's (modded), Squeezbox Transporter, Emotiva XPA-2, HSU Research sub, Sub amp - Crown, MSB Link DAC lll, Bottlehead Quickie preamp (modded), JPS Labs Digital powercord, MIT Z-cord, JPS Labs Ultraconductor 2 interconnect and speaker wire

Keladrin -- Fri, 05/18/2012 - 03:23

'A clear difference in sound on weather the JPS was in the DAC to pre or pre to amp position'
The first thing here is to get somebody else to change the interconnect (or not) and see if you can clearly identify if it has been changed and which one is in. Then it would be interesting to see what differences you are actually hearing. Interconnects are a bit tricky as if they are not shielded properly or there is another issue there can be an issue with dynamic range. Turn the volume right up with both but with no music playing - is there a difference in background noise? The issue won't be to do with the metal or conductor in the inconnect, but could be some other issue around implementation or system issue - worth investingating further.

All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.