Will a good external DAC make an Average Transport sound fantastic? (1 Viewer)

[Chu Gai]

I'm quite sure that an application of one of these venerable and proven reptilian elixers will cure yer jitter

seriously though Lee, if there are truly audible differences, I think we need to focus our attention on something else.
[c] [/c]

 

[Chu Gai]

Really depends how you're doing the tests Mike.
Vibration of the transport mechanism isn't jitter and it'd have to be severe enough that the system couldn't read the data correctly. That much is implied in Dunn's findings that you're in good shape so long as the data can be read.
As far as noise on power lines, if you've got a link to a paper referencing this, I'd like to look at it. Most equipment is pretty good at dealing with everyday power however that hasn't stopped people from buying all sorts of things out of concerns, real or imagined, to provide 'better power'. Sadly for the consumer, legitimate work to substantiate claims of improvement is sorely lacking and harks of infomercials. Sloppy science ain't science...it's just sloppy.

 

[Michael R Price]

Chu,

The best explanation of jitter that I can remember reading is by Jon Risch: http://www.geocities.com/jonrisch/jitter.htm

Not that I can speak to any of these claims from actual experience.

 

[RichardHOS]

After having read a substantial amount of Jon Risch's writings, I have come to the conclusion that he is full of shit. There's so much pseudo-science in his writing that the factual statements and good points he raises are completely lost in the flood of nonsense. As always, YMMV.

 

[Chu Gai]

about 50 using the Toyota Hybrid...don't vary too much.

look, i can appreciate, from a technical point of view, companies such as Meridian, that strive to minimize distortions, improve upon things such as rotational stability, look to maintain signal integrity, in a sense apply a bit of overengineering. that's laudable in my estimation. jitter for some of their products is somewhere in the 200 ps area which i see as an honest attempt at quelling what's become the silliness with this jitter stuff. drive the numbers low enough. drive them below where they can reliably be heard using test tones and headphones. drive them low enough so that it's tantamount to us arguing about the number of angels that can fit on the head of a pin. i guess that depends on whether chris farley and john candy are angels huh?

Risch, with whom I've some issues, I don't think is 'always' full of excrement. However, he's very conveniently chosen to ignore psychoacoustics and related sciences and instead dabbles around in areas where he has absolutely no expertise. As an engineer, he's well aware of the scientific method, yet he refuses to disclose his methodology. And I dare say because he's well aware that his tests are flawed and won't stand up to scientific scrutiny. Risch, from AA fame where he moderates, is the defender of the crackpot theory that one will search in vain for the tiniest shred of scientific evidence. Whether it's defending superconductivity theories for Bybee devices or maxwellian postulations for the reasons why to his ears, wires sound different, Risch is there. This defense of the indefensible alone merits honorary membership into the ACLU!

One of the marks of the fake, the fraud, the quack, the charlatan, is to tell the truth to a point. Tell enough that what's said is not only known but has been verified experimentally. Then, after you're hooked, and nodding your head like some bobble head doll, the leap to the unknown and speculative occurs.

This and $5 will get you a cup of joe at Starbucks.

 

[Lee Scoggins]

That level of jitter could be audible (as shown by Eric Benjamin and Ben Gannon's paper at last years San Francisco AES) but the conclusion that then jitter audibility in the ps region is possible is misleading by a factor of (possibly) a thousand.

Chu, I spoke with the man a few months before his passing and he said:

"Myself and most researchers believe that the range of audibility is around 20 picoseconds of jitter."

This statement was made at the end of 2002, a full four years after that statement you referenced. So he clearly chnaged his mind.

Advances in jitter measurement have been substantial over the past four years and science has moved on to new conclusions about what is audible.

 

[Lee Scoggins]

look, i can appreciate, from a technical point of view, companies such as Meridian, that strive to minimize distortions, improve upon things such as rotational stability, look to maintain signal integrity, in a sense apply a bit of overengineering. that's laudable in my estimation. jitter for some of their products is somewhere in the 200 ps area which i see as an honest attempt at quelling what's become the silliness with this jitter stuff.

We record in hirez at 88.2 khz on an Alessis MasterLink at 200 picoseconds. Over time we experimented in lowering jitter by synching to an external Lucid Audio master clock which take the jitter down to 25 picoseconds.

The result? more clarity across the FR, more natural midrange, tighter and deeper soundstage. The clock was the only thing we changed and we have noticed this on transfering analog tapes, recording live events, studio work, etc. Clearly even low levels of jitter can make a difference. Michael Bishop and Bob Ludwig and others have also been exposed and understood the effects of master clocks and use them regularly in their acclaimed work.

This is not some audiophile dream, but something that is theoretically discussed by academics and applied in the real world by the top recording and mastering engineers.

 

[Chu Gai]

Dunn was well aware of the differences between beliefs and controlled testing. There is though the difference between test tones and actual musical material listened to in an environment where the noise floor is substantially higher. This always makes differences harder to discern.
Again, if I take your statement at face value that 20 ps of jitter is audible and we look at that in terms of Dunn's paper, there's nothing to dispute. However you're failing to grasp the difference of Dunn's 20 ps. 120 dB levels up at the limits of human hearing using a sine wave.
Insofar as your particular finding with regards to the CD's, I can only add that you'll need to conduct tests with the same rigorous attention to detail as the aforementioned papers before you can draw any conclusions that mean anything. For all we know, the ultimate levels were different or the lower jitter clock resulted in less corrections being sent to the motor which may've resulted in improved stability...I just don't know and neither do you. Besides, the jitter concerns from where i sit have more to do with consumer products and not issues from a mastering point. Recording engineers, and I don't mean you Lee, should worry a bit more about making decent recordings and it ain't because of jitter!
BTW, if you've got those two different pressings how about burning the same selection on the same CD and send it to me?

 

[Michael R Price]

"Recording engineers, and I don't mean you Lee, should worry a bit more about making decent recordings and it ain't because of jitter!"

Oh man, that is far more true than anything else said in this thread.

Well, eventually I'll have to put in a new clock circuit and tell you if it makes any difference because of lower jitter...

 

[Chu Gai]

and you'd be relying upon what to make the assessment? acoustic memory?

 

[Lewis Besze]

Chu, I spoke with the man a few months before his passing and he said:"Myself and most researchers believe that the range of audibility is around 20 picoseconds of jitter.

Did he present this to the AES in a same form as his previous papers?
Since he's not around we can only rely on your "testemony",which I must say is, far less convincing.

 

[Michael R Price]

Chu,

That's exactly why I would rather not judge this sort of thing... if a new clock makes me feel better, it's worth it (at least at the cost to me). If it actually sounds better, that's wonderful. Not that I'd be able to prove it, though...

 

[Craig_Kg]

Chu, I said before that you were looking at (listening to) jitter in the wrong manner. Not hearing a warble on a test tone does not mean that MUSIC is not affected. I don't agree with Lee on some issues but he is correct about the degradation caused by jitter (although I can't verify the numbers he quotes).

I have heard differences in transports into an unbuffered external DAC and I was astounded. A fully buffered DAC will eliminate this, of course, so that any jitter is solely due to the stability of the DAC master clock. I also know people who have installed new clock circuits in their CD players (with no other changes) and found huge differences vs an unmodded player when comparing the 2 directly.

I take it you're familiar that modern DAC's have anti-jitter circuitry built in by the manufacturers.

This works to a degree REDUCING the incoming jitter but the PLL used to reclock the incoming bitstream will still pass on some of the jitter - it must to adjust to bitrate variations in the incoming bitstream since the inbuilt buffer is so small.

I suggest anyone interested in this issue listen for themselves in as close to a DBT environment as they can arrange.

 

[Chu Gai]

Craig, I'm not stating that jitter is inaudible under any circumstances, simply that if we consider Dunn's J-Test, which I believe is how the aforementioned papers did it, that there is a point where it becomes a boogyman. As manufacturers embrace this, and I hope they do, and the results are published, to me, consumers will be able to make intelligent, informed decisions regarding their purchases. The point of mentioning the incorporation of anti-jitter circuitry into a DAC was not meant to imply that doing so drives jitter to 0. No more than a shock absorber in a car does not imply you won't feel any shocks.

The numbers that Lee quotes, and which I also do, are based on some careful work by some notable people who've gone to enormously great lengths to vary jitter and then determine at which point it becomes statistically audible. In Dunn's case, it was with test tones. In the other paper, this is extended to music. If the jitter is inaudbile or can't be reliably detected on a test tone, it won't be detected when playing music. Not today, not tomorrow.

What you heard, I can't say. Perhaps one unit had significantly audible levels of jitter and the difference was indeed that. Perhaps the difference was something other than jitter. Some of the work Dunn was investigating were those 'other' things and one of them was cross-talk in the electrical components.

I tell you though my friend, some of the things that people who play vinyl accept, cherish and raise to the levels of a unique signature, would never be tolerated by those playing CD's or the like.

BTW, my thanks to the Aussie gov't for supporting the US

 

[Craig_Kg]

Some of the work Dunn was investigating were those 'other' things and one of them was cross-talk in the electrical components.

The external DAC used was the same in the case I heard so I can't see how anything but jitter was the difference. I'll see if I can find measurements of the jitter for those players.
As an aside, the Chord DAC64 has a selectable buffer with no buffer, 1 sec and 4 sec options - the differences are quite audible with music and I can't explain the (smaller) differences between the 1 and 4 sec settings except that perhaps the larger buffer allows a much greater comparator period for the PLL so transport jitter is further reduced.

I'm an engineer, too, and understand what you are getting at but the problem is that the psychoacoustic model of our hearing is incomplete and some things can't be easily quantified yet. The study with test tones is a case of looking in the wrong area.

Vinyl? Not for me, pal. Whatever the shortcomings of digital recordings (which are damned few IMO), CD is the greatest mass market audio medium ever to succeed and I have severe doubts as to whether the hi-res formats will be successful or are even necessary (a major point of contention between Lee and myself).

I'm glad our government decided to support yours and Great Britain's - means I voted for the right mob.

Where are the moderators? This thread really belongs in the AV sources forum.

 

[Lee Scoggins]

For all we know, the ultimate levels were different or the lower jitter clock resulted in less corrections being sent to the motor which may've resulted in improved stability...I just don't know and neither do you.

Actually I do know...here's how:

1. Lucid calibrates to a jitter manufacturing standard of around 20PS.

2. We have heard this effect across several Lucid master clocks.

3. We do not change ANYTHING else in the recording chain so it has to be the effect of the clock.

4. Volumes were matched.

Here I have recounted real world experience Chu and you seem to want to blame the difference on everything but that of jitter. We have scientifically found the answer though - lower jitter from 200PS to 20PS. We have eliminated any other possible variation and we have reproduced the results over several recording sessions.

Craig, you are right about test tones. These often substitute for musical passage because the brain has an easier time finding them. But real music is a complex aggregation of same and such tones better produced at any frequency will add to the music's clarity and definition.

 

[Chu Gai]

have to part ways with you regarding test tones as those are more sensitive to anomalies than music and while science's understanding can always be said to be incomplete, the science of psychoacoustics is surprisingly well understood and like all the sciences, an area of active research. maybe this is one of these things you and I can explore a bit via PM...game?
the phenomenon of jitter was explored though to a large extent long before audio and it's probably to audio's detriment that designers and coders didn't fully appreciate it. Kind of like the first few solid state amplifiers that just didn't get a variety of things right at the getgo or those super thin long lengths of speaker wire that used to be used which gave birth to Monster and all the little Monsterettes.
The distortion due to jitter is proportional to both the frequency and amplitude of the signal. Now the smaller the signal, the lower the distortion. Since we know that above several kHz, the amount of music is rolling off. Probablistically, that translates to there's likely to be more small signals at the upper frequencies rather than larger ones. This then implies that distortion of that low intensity signal is quite likely to be inaudible. When signals are very low, the ear behaves in a non-linear fashion. Hence small movements of the head, a change in our perceptual focus, can alter our perception of a signal even though that signal hasn't changed. Further, the ear's sensitivity decreases rapidly once we go past ~5 kHz. It takes larger threshold differences in order for us to reliably discern them. Now if you've listening to music, any musical energy up in the frequencies where there's any jitter can mask it's audibility.
All these reasons and more are why jitter becomes less audible by orders of magnitude, as found in the Gannon paper, when one's talking music.
As an engineer Craig, I know you appreciate throughness and the importance of investigating other phenomena before one begins to draw conclusions. In the personal example you cited, same everything except you varied the amount of buffering, it might be premature to ascribe what you heard to jitter. What might be happening is that by varying this buffering, you're introducing unrecoverable errors into the setup. If so, that's not jitter. There is a test CD that you can buy from an outfit out in Canada (~$25), that'll allow you to test your player for things like that. If interested, I'll post or PM you the link as you wish. If the changes in that switch somehow introduce level changes in output, you'll have to burn a CD with a test tone and see if that's the case by measuring the voltages at the speaker terminals. In a variety of work that Dunn did, not necessarily related to jitter, test tones were invariably included with musical program material to provide further understanding into what may've been going on. What you do Craig, depends upon to what extent you want to put your knowledge of the scientific method to bear upon what you're hearing. Statistics, blind testing, and all that.
There's no question that the high end crowd is going to spend money, and substantial amounts of it, to achieve goals, embrace design concepts, that to them mean something. Some of these design concepts will be foolish and incompetent such as belt driven CD players with colorful lights. Some will come from escoteric and limited production products that rely more on the beauty of their artistry rather than the competence of the designer. Some will be pure fluff as when companies change a couple of capacitors and resistors for 'audiophile ones', prominently advertise it, but in reality they do nothing.
I see that as things like jitter, as in the Gannon paper, are driven to levels of inaudibility as they seem to be now, that this is a good and absolutely wonderful thing for the more typical consumer who is working on a budget, even a $2000 budget, to obtain good and technically competent as well as transparent CD players.
Now if we could just put this myth of the quality of hand made to rest...

 

[Chu Gai]

Lee, both of the papers which were cited here and elsewhere were concerned about jitter from the playback perspective on a listener's system, not at the recording studio. I'd read elsewhere on HTF where you conducted a listening session but ran only one trial. Now honestly, how does one draw any significance out of one trial? In all fairness, for you to have put that up and mentioned those papers in the same breath is like taking pictures of Hussein at a children's school and saying he's pro-kids. A gross exageration on my part though that seemed to work for CNN However if you want to hold your personal findings up and expect them to receive the same consideration as those papers, you'll have to do more than flip a coin one time.
Have you considered contacting the researchers or Dolby labs and either apprising them of your findings or perhaps offering to send them your disks?

 

[Michael R Price]

Chu,

Thanks for sparking this lively discussion.

Unlike Lee and some others, I have no experience actually listening to, or measuring in any way, systems with different levels of jitter. I have no scientific explanation for why jitter levels in the picosecond range are audible. However, I trust others (especially independent DIYers not trying to sell some silly thing) who do. My blind trust stops at the comments of others who installed LEDs in their players to light up the disk platter and seemed confident that the sound improved.

Elso Kwak has been kind enough to send me a schematic for his latest low-jitter oscillator, which I should soon get around to building and installing in my Denon CD player. While I might form a stronger opinion then (or reverse it entirely), I will never be able to substantiate any difference in sound quality from the upgrade. Isn't that great?

You know, what's funny is that this discussion was originally about something a bit less subjective and unexplained than transport jitter... do we agree the DAC (and its subsequent analog stage) make a much bigger difference?

 

[Chu Gai]

It's actually a sucky proposition Mr. Price and my tendencies would be towards looking for a competently made product with low jitter rather than adding onto an older product. Will it improve your existing product? Maybe but truly how to tell seeing as audio memory regarding small differences is a poor way of doing things.

FWIW, Dunn's general position regarding CD players was that adding outboard DACs or 'low jitter cables' was not monies well spent. Of course, he was likely thinking along the lines of where products are today.