Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Tomas Danko]

Nika Aldrich wrote on Wed, 27 April 2005 13:45

Tomas Danko wrote on Sat, 23 April 2005 09:42 I'm thinking about it... Apogee Big Ben, perhaps? I don't think so.  I'm pretty sure that Weiss' box is intended to mask jitter going to the D/A by adding additional jitter of a white noise variety - again, think of it like dither for jitter, though I hate to go down that path. Big Ben (I believe) attempts to reshape the existing jitter to conform to a curve that is more harmonically pleasing.  We could think of Weiss as pure dither and Apogee as noise-shaping.  While the frequency response of jitter does not affect us quite like the frequency response of distortion in audio, there are certain frequencies that we are less prone to hear, and if certain spurious components are present we may hear them more noticeably. The trick of each of these boxes seems to be to treat the jitter.  One does so by masking the spurious components.  The other by reshaping the floor altogether. FWIW, Dan's "Crystallock" technology is much more akin to Apogee's solution - reshaping the existing jitter so that it is less audibly offensive.  Dan uses very lengthy buffers and PLLs to "reshape" the jitter such that it has extremely high amplitude at very low frequencies where it is less likely to cause audible distortion, thereby removing the jitter at the higher frequencies where it will cause more audible distortion. Nika

Hello Mr. Aldrich,

I was talking about Apogee, not Weiss, and we seem to be suspecting the same thing (noise-shaping in the time-domain). I am aware of Mr. Lavry's clocking technology, where the jitter is down close to DC so-to-speak. It's a very elegant solution.

Sincerely,

Tomas Danko

[Tomas Danko]

Nika Aldrich wrote on Wed, 27 April 2005 13:49

Tomas Danko wrote on Tue, 26 April 2005 11:57 To me, the sensible way to approach this on a practical level is to first ask the question 'how bad are the internal clocks of today?'. Can they really be so bad, they need to be 'fixed' by an external clocking device? Now, that would mean they are very bad. I think it is certainly possible that the clocking in internal devices measures better in an overall amplitude measurement.  An overall amplitude measurement, however, does not necessarily indicate how the jitter will manifest itself upon listening.  It is certainly possible to have two boxes - one with an overall jitter of, say, 1ns and another with an overall jitter amplitude of, say, 10ns, but the latter to sound significantly better. If internal clocks have lower amplitude but the frequency distribution of the jitter is more audibly destructive then external boxes may still have a place?  Yes? Nika

This rebounds to my first aber in this matter. Are mid-priced to high-end converters this flawed in the clocking department? Really? If they do, a box may still have a place, although I would much rather buy something that wasn't broken for starters.

Once again, I don't think it's a good investment to get a less expensive converter and then try and improve it using external clocking. Better then, to invest that sum into something that is competently clocked from the get-go.

...I wouldn't hesitate to use a master clock if I had lots of different converters and computers and a need to have them in sync, however.

Sincerely,

Tomas Danko

[Nika Aldrich]

bobkatz wrote on Thu, 21 April 2005 20:01

HI, Daniel! Daniel is one of the good guys... Remember, everyone has to market to survive and get noticed. Personally, if I were in Daniel's shoes I would be afraid to market in the audiophile high end sector, you can get burnt and it is not a guaranteed profit. So I admire Weiss's adventure! The results will only be better Weiss professional equipment for us all. A well-funded company is a healthy company. BK

Bob,

Every once in a while I'm prone to call them like I see them.  I hemmed and hawed on your post throughout the day and I am convinced at this point that your post has dangerous implications.

What you SEEM to say is that the marketing may very well be bogus, but it's OK because Daniel is a good guy and it's better for us all if he has enough money to make good gear, even if it's on the backs of poor and inaccurate marketing.

I want to clarify that I am not supposing that Dan's marketing IS bogus, but rather am pointing out a disconcerting underlying message in your post.  You, Bob, and I have long been keepers of the flame of upholding truth and full disclosure in this industry DESPITE the poor and inaccurate marketing by the manufacturers.  We have long worked to dispell both industry myths and marketing hype.  It is therefore NOT OK for Daniel to use whatever marketing means necessary to support this product.  The marketing, even if intended for the marketing-myth-prone audiophile industry MUST be accurate and truthful or else it is NOT good for the general atmosphere and full disclosure for which we have long fought.  

I say yee to Daniel and his marketing if it is truthful and accurate (and I do not suppose at this time that it is not) but think we must be cautious about labelling the good guys from the bad guys and writing different rules of conduct for each.  Instead we should write a single rule of conduct and we can then ascertain the good guys from the bad based on who follow it.

Cheers!
Nika

[Terry Demol]

bobkatz wrote on Sat, 23 April 2005 11:29

Daniel Weiss wrote on Thu, 21 April 2005 23:55 Well, the added noise does not "clean up" the wordclock in the usual sense, you may even see higher jitter at your DAC with the "DAC enhancment" on. But the spectrum of that jitter is (ideally) flat. This causes the SNR to rise slightly but does not cause the nasty effects which discrete jitter frequencies do. It is something to the effect like dither, but in the time domain. Daniel www.weiss.ch I've seen approaches like that before, including the way that Meridian manipulated the spectrum of the jitter in one of their processors. It gets down to psychoacoustics and noise masking. But it seems to me you have to know the entire system real well and predict how the particular PLL/DAC system will react to the "pre-emphasized jittery" incoming signal. But this discussion does get down to questions of "is less jitter always better sounding?" And I certainly have experienced audible tradeoffs... relating to the fact that if you reduce the jitter to the lowest you can make it, there will be residual jitter with a particular spectrum that translates to a certain noise floor or distortion floor within the DAC. Which if it hits the ear's sensitivity region could sound worse than if you add a carefully controlled film of white noise to cover that up. Doesn't matter if that white noise was created in the time or frequency domain. Think about it. BK

Bob, yes dither in the time domain, however, remember one
important thing, that the random noise is not upper frequency
restricted by the sample rate, where with dither it is.
IOW you can inject low level random phase noise up to and over many
MHz.

I would like to add that this was done quite a few years ago
by a very clever guy called Charles Altmann in a device he calls
Jisco. Initially it was mostly written off as hi-end foolery by
most pro types but reports point otherwise.

Heres a link:  http://www.jitter.de/

Now on to my second point... all through that god-awful
Apogee bashing Big Ben saga I was thinking -I bet they are doing
something with their DDS clock akin to "Jisco". Since I
objected very much to the general  spirit of that thread I did
not say anything but there could heve been some interesting
ground to cover... but then again I suppose there wasn't a
whole lot of creative thinking going on in the heat of battle

Cheers,

Terry

[danlavry]

bobkatz wrote on Sat, 23 April 2005 11:29

Daniel Weiss wrote on Thu, 21 April 2005 23:55 Well, the added noise does not "clean up" the wordclock in the usual sense, you may even see higher jitter at your DAC with the "DAC enhancment" on. But the spectrum of that jitter is (ideally) flat. This causes the SNR to rise slightly but does not cause the nasty effects which discrete jitter frequencies do. It is something to the effect like dither, but in the time domain. Daniel www.weiss.ch I've seen approaches like that before, including the way that Meridian manipulated the spectrum of the jitter in one of their processors. It gets down to psychoacoustics and noise masking. But it seems to me you have to know the entire system real well and predict how the particular PLL/DAC system will react to the "pre-emphasized jittery" incoming signal. But this discussion does get down to questions of "is less jitter always better sounding?" And I certainly have experienced audible tradeoffs... relating to the fact that if you reduce the jitter to the lowest you can make it, there will be residual jitter with a particular spectrum that translates to a certain noise floor or distortion floor within the DAC. Which if it hits the ear's sensitivity region could sound worse than if you add a carefully controlled film of white noise to cover that up. Doesn't matter if that white noise was created in the time or frequency domain. Think about it. BK

First, a lot of people here are mixing external clocking of AD and Clock issues of a DA.

Second, I find the analogy between jitter and dither to be very poor. We
dither because we HAVE TO confront reduced accuracy due to word length limitations.
The most common example is when word length reduction from 24 bits to 16 bits for CD. But if the standard allows 24bits, we do not need to dither. In the case of jitter, there is no “standard” forcing us to accept more jitter. Less jitter at the converter is better signal transparency. Period.

Third, a “noise shaped jitter” does not arrive directly at the DA device. It gets altered by the circuitry between clock input and the DA, so the outcome is device dependent. Also, noise shaping and spread spectrum can not eliminate all the non random source of jitter, only jitter due to some limited causes. There have been other schemes to eliminate or help fight against systematic (non random) jitter, but they are all for DA’s.

Fourth, the cleanest way to clock an AD, and to avoid systematic jitter is to use INTERNAL crystal. When that is not possible, the next best thing is to focus on a real good jitter reduction INSIDE the AD (PLL circuits and alike). You can have an internal crystal provide few to a few tenth of psec jitter. By the time you are going for external clock, there is NO POINT in spending tons of money on a few ps (or tenth of ps) because the rest of the damage will be so much larger. Better to put your money on a AD with good clocks and PLL’s, and get a cheap clock, than to spend a lot on a clock that drives a poor AD (in terms of jitter rejection).

For more details regarding my thoughts on clocking, I am going to post a lot of the information about proper wordclock implementation on the forum on my site at:

http://www.lavryengineering.com/forum/phpBB2/index.php

Your comments are appreciated.

Regards
Dan Lavry
www.lavryengineering.com

[Nika Aldrich]

danlavry wrote on Thu, 28 April 2005 11:02

Second, I find the analogy between jitter and dither to be very poor. We dither because we HAVE TO confront reduced accuracy due to word length limitations. The most common example is when word length reduction from 24 bits to 16 bits for CD. But if the standard allows 24bits, we do not need to dither.

Dither is a methodology wherein we add randomness in order to linearize the overall behavior of a system.  Dithering clock signals in the time domain (also referred to as "jither," at least as it is specifically used with 1 bit delta sigma modulators) can help to linearize the system to a certain degree.  As to whether this is a preferred method I am dubious, but it does seem to me that dithering in the time domain is what Daniel Weiss is attempting to do.  I'm not sure the analogy is poor unless we're fixated on a very specific definition of dither having to do with bits.  I often dither the jar of peanuts when I want to get just a couple out.  I also dither the cookie sheet when I'm trying to get a few tater tots off of it.  I dither a cup of sugar to make it flat on top.  I also dither the gas pump to make sure I don't drip gasoline on my shoes.  There are many ways in which dither is used every day.  I think calling this technique "dithering" the time domain jitter is valid?

Quote:

Less jitter at the converter is better signal transparency. Period.

To this I most certainly cannot agree.

Nika

[danlavry]

Nika Aldrich wrote on Thu, 28 April 2005 16:50

Quote: Less jitter at the converter is better signal transparency. Period. To this I most certainly cannot agree. Nika

To a designer (even a beginning designer) this is elementary information. You can disagree all you want, and be wrong forever.

In this thread we are talking about JITTER. Transparency is BY DEFINITION the ability of the converter to produce a PRECISE DUPLICATE of the incoming waveform. If one captures (AD) or reproduces (DA) samples at an inappropriate time – a deviation from constant time interval, one ends up with an ALTERED WAVFORM. That time deviation is what we call jitter.

We are not talking here about number of bits, voltage linearity, clipping… We are talking about the time factor called JITTER.

Dan Lavry
Lavry Engineering, Inc.
http://www.lavryengineering.com/forum/phpBB2/index.php

[Nika Aldrich]

danlavry wrote on Thu, 28 April 2005 18:27

Nika Aldrich wrote on Thu, 28 April 2005 16:50 Quote: Less jitter at the converter is better signal transparency. Period. To this I most certainly cannot agree. Nika To a designer (even a beginning designer) this is elementary information. You can disagree all you want, and be wrong forever. In this thread we are talking about JITTER. Transparency is BY DEFINITION the ability of the converter to produce a PRECISE DUPLICATE of the incoming waveform. If one captures (AD) or reproduces (DA) samples at an inappropriate time ? a deviation from constant time interval, one ends up with an ALTERED WAVFORM. That time deviation is what we call jitter. We are not talking here about number of bits, voltage linearity, clipping? We are talking about the time factor called JITTER. Dan Lavry Lavry Engineering, Inc. http://www.lavryengineering.com/forum/phpBB2/index.php

Dan,

We cannot deny that these waveforms are being captured and reproduced as a form of transmission - of communication.  As in any communication model there is a transmitter and receiver.  In this specific case the ultimate receiver is the human ear - a very complicated and non-linear device that does not inherently work like a more linear data-analysis system.

As such, it is certainly possible that the more "TRANSPARENT" transmission of the signal might have MORE jitter than an alternate, but that the jitter manifests itself in ways that are less impacting to the ultimate receiver.  For example, a device that imparts 10ns of jitter at .1Hz will sound more transparent to our ears than a device that imparts 1ns of jitter at 1kHz.

It is simply NOT correct that lower jitter is always better - with respect to the human ear.

Nika

[Bob Olhsson]

The problem is that you can never treat jitter as something isolated from additional jitter with widely varying spectra. It will always combine giving you sum and difference frequency modulation. While it might sound great in the studio, it could easily be a disaster out in the real world decoded by other less that perfect clocking systems. Making sound dependent on a particular flavor of jitter in order to be properly reproduced is at best a slippery slope.

I'll say it again, a converter that changes its sound for better or worse when it is externally clocked is a broken product design. The fact that ARE lots of broken designs doesn't excuse this fact or alter the goal of achieving a truly jitter-proof converter design.

A lens having a built in smear of petrolium jelly may make the model look more appealing but it doesn't change the fact that it isn't an accurate representation of what's in front of the lens. Dither, on the other hand, is equivalent to an optical coating that exchanges ultimate light transmission for a sharper image that passes more information because it's free from reflections introduced by light sources falling on the glass elements.

[Johnny B]

Sounds like some people are crossing into the realm of the "art" part of the category of the "Recording Arts and Sciences."

Sometimes properly applied "art" yields fantastic results, sometimes not. The same can be said for the "science" part. Striking the proper balance between the two is indeed a wonderful thing.

YMMV.

[PP]

[George_]

well, you should write this sentence down to a book...

“You can disagree all you want, and be wrong forever.” Dan Lavry



I love it man..

I have no idea of jitter.. because I accept the small performance of my AD/DA card.. and try to get the best out of my smallpocketcard(s)..

till the day I build a better converter than you(and your crew) does (this is probably in my next life;).. so I want), you are my personal hero on converting something..

can you convert my little money to big money.. please with lil jitter..

     

thanx a lot for taking the time to answer all our questions!!

[PP]

[George_]

so, and it's called??

no, I see the problems in all forum, people have their own opinion wich depend on their own experience or "I've heard that... blah"..

espacially in the PRO forum, and with people thinking "hei dude, my way is the best way" a big problem is the most of 'em are not frank with other users, people, or they have a such small focus they cant take a look a lil bit to the right or to the left side..

this is now by far, very very offtopic..

but I think, I can learn from a 17 year old SAE or Musicstudent and also from the old weasels in audiobuisness..

just keep your mind open, hear what they have to say, then look critically on their statement and answer friendly. If you animadvert his statement/solution/opinion than do that with a complementary statement/solution/opinion..

there is a lot of envy out there.. don't let it take control over you!!

this is my personal opinion.. and I'm open for critics..

[PP]