R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: 1 ... 7 8 [9] 10 11 ... 15   Go Down

Author Topic: Proper word clock implementation  (Read 173845 times)

Sam Lord

  • Full Member
  • ***
  • Offline Offline
  • Posts: 139
Re: Proper word clock implementation
« Reply #120 on: November 16, 2004, 09:00:49 PM »

Thanks Dan, your answers were very helpful.  Sorry for the slow reply; no need to reply unless you so wish.

[quote title=danlavry wrote on Sat, 13 November 2004 17:22]

[Accuracy is not jitter, and jitter is what matters.]

Oh that's right, jitter is variation in the clock period, but accuracy refers to the number of clock cycles per time interval, I think.  I remember Stereophile actually listening to music from a modest CD player way back, and just putting an FM amp on the *digital* output, and hearing music because of the correlated jitter.

>>Regarding the “noise from clocking”..... the interference is completely rejected because it falls on 0Hz and multiples of Nyquist, thus can be later filtered without impact on the audio range.<<

Oh, now I see.  Thanks again, Sam
Logged

Andy Peters

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1124
Re: Proper word clock implementation
« Reply #121 on: November 19, 2004, 06:53:11 PM »

Nika Aldrich wrote on Fri, 12 November 2004 07:04

You may have thought this all through and I may just be missing a link, but I still don't understand how your circuit deals with the situation of two clocks that think that 44.1k is different.

Let's start with a scenario where the two clocks are out of synch by one 44100th of a second.  The external clock keeps shoving the info into the buffer and the internal clock keeps reading it from the buffer, but after 1 second there's 1 less sample in the buffer than there was to start with.  After two seconds there're 2 fewer samples in the buffer.  After every second we lose a sample in the buffer.  How are you going to reconcile this situation?  Do you have an infinitely large buffer?  Or do you recalibrate the clocks so that they both see time as roughly the same?  How do you do this?


Ah, sorry for the delay (busy!) but I do see your point!

-a
Logged
"On the Internet, nobody can hear you mix a band."

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Proper word clock implementation
« Reply #122 on: December 15, 2004, 01:08:31 PM »

Nika said:

“I posted a lengthy list of questions to help me iron out the ongoing dispute on the "clocks" issue. I understand Dan was kind enough to reply at length to those questions. I received a phone call when I was on the road today advising me of the response and asking me to read Dan's comments. I came home and the thread is gone.

What happened? Where is it? Why?

Nika”



This is the clock thread!

In the thread “Something I do not Understand” you wrote. “I’m trying to get you to show how. You’re the engineer. It’s all math. Math doesn’t lie, and engineers don’t make mistakes, right?” I took much time to write a good explanation and when you discovered I was right you did not acknowledge my effort or my position.

Instead, you immediately started another Clock thread demanding answers from Lavry and another company, in what you called an on-going dispute, stating that you and Bob need to get to the bottom of things. Your message tried to create a new format of “Dan answer this” and "Apogee answer that”.  As far as I am concerned there is no “on-going dispute”. I already stated technical facts about the clock that no one can or did dispute, and that is all I will do for now.

At the same time, on a thread “Understanding Dan’s 192Khz Arguments” you virtually repeated what I said and asked me to comment on it.

Dear Nika, I have a life and cannot be your personal tutor. Your level of activity here is welcome if you can bring new ideas instead of creating a circuitous rehash. I do not have time for such interactions. Also, it is inappropriate for you to pose as an authority, translator or moderator, while at the same time using direct sales techniques to sell your book.

One to one personal communications are now necessary so that I will not have to embarrass you. Please contact me directly.

Regards,
Dan Lavry


Logged

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: Proper word clock implementation
« Reply #123 on: December 15, 2004, 01:38:02 PM »

danlavry wrote on Wed, 15 December 2004 13:08

This is the clock thread!


Great.  Will you please repost my previous questions.  

Quote:

In the thread ?Something I do not Understand? you wrote. ?I?m trying to get you to show how. You?re the engineer. It?s all math. Math doesn?t lie, and engineers don?t make mistakes, right?? I took much time to write a good explanation and when you discovered I was right you did not acknowledge my effort or my position.


No.  I found your answers to be unsatisfactory.  I did not think that they addressed at all the specific question I asked.  I found my answer another way - in a way that specifically addressed my questions, and explained the answer to me in a way that was helpful.  Rather than hash out the discrepancies between what you posted and what my question was I thought I should let it drop.  I appreciate your time in answering, however.  I do understand your post and certainly agree, but don't find it to be related to what I asked about.  I didn't see the need to belabor the issue.

Quote:

Instead, you immediately started another Clock thread demanding answers from Lavry and another company, in what you called an on-going dispute, stating that you and Bob need to get to the bottom of things.


I called it an ongoing dispute because apparently it is.  I did not demand answers.  I merely asked questions to help the bulk of us reconcile disputed and discontinuous and incomplete information - information that is not reconciling between the various parties involved.

Quote:

At the same time, on a thread ?Understanding Dan?s 192Khz Arguments? you virtually repeated what I said and asked me to comment on it.


People were asking for clarification and more specific examples on a point you had raised (and one that I had raised several times before over the past few years).  I tried to answer their question but provided the courtesy to you of not trying to speak on your behalf erroneously.

Quote:

Your level of activity here is welcome if you can bring new ideas instead of creating a circuitous rehash.


I believe that NONE of the questions I posted last night were a "continuous rehash."  If the answers to those questions can be found somewhere on this thread or in your other writings please direct me to them.  I don't see how it is "rehash" at all.

Quote:

Also, it is inappropriate for you to pose as an authority, translator or moderator, while at the same time using direct sales techniques to sell your book.


RE: posing as an authority, doesn't anyone who answers a question pose as an authority in some capacity?  How can you claim that noone should pose as an authority on an open forum?  As for a translator, you're the boss.  I disagree.  Not everyone understands everything the same way all the time.  Sometimes a different approach is needed.  We do a lot of re-interpretting information on these forums.  As for moderator - I don't pretend to be one.  I gave up that responsibility on this site many months ago.

Yes, I am using direct sales techniques to sell my book.  I don't see the consistency in your statement, however.  

Nika
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Proper word clock implementation
« Reply #124 on: December 16, 2004, 07:40:13 PM »

“Dan,

1. I have heard now from multiple places that you have claimed that the Big Ben "has no crystal clock inside of it," and that it is "synthesized."”


I never stated it, and in fact if you go and read that clock thread, I assumed the best case – that they do have an internal crystal, only at higher frequencies. One of my points was that a fundamental harmonic crystal is less jittery than higher frequency crystal. But that was not my main point. Here you are making me rehash what I already said.

I am not against external clocks when they are needed. I already stated that whenever one can use a reasonable quality internal clock, located inside, near the AD, they will have less jitter than using external clock, sending it through a transmitter, cable, receiver, and phase lock loop circuitry pushing and pulling a VCXO crystal that is more jittery than a fixed one...  Do I need to rehash it again?

I already explained that a “clock box” can not send a signal to reduce jitter in another box, without knowing what is taking place at that other box. The idea of being able to correct something without having knowledge of what it is bogus! I stated that EVEN IF the clock box were able to send ANY signal to correct the jitter inside the AD (AND IT DOES NOT!), it would not know what signal will correct the jitter, because the clock box DOES NOT KNOW what is wrong in the AD box ahead of it.
That is so because it is a one way communication- No feedback. The clock box sends THE SAME signal to any AD box. It does not change the type of signal for different AD’s. I already said all that. Do I need to rehash that?

But let us pretend that there is magic in the air, and the clock box knows exactly what to do, and what is taking place at the end of that cable leading to the AD box (and that is already VOODO ELECTRONICS). Given that the clock attempts to send a perfect square wave, perfect timing, no jitter… where will it put the message that contains the instruction to fix the jitter in the AD? The information to make things better is not sent via amplitude modulation. It is not sent via frequency modulation. It is sent via no modulation. Modulating the clock signal is adding jitter.

So putting it all together, you send a message via no modulation (thus no message) to fix something you have no idea about (that varies from AD to AD) and get less jitter?

Or is it that you send a message via modulation, thus adding jitter, though you do know which message to send, and the message will exactly cancel the jitter due to unknown factors, including random jitter?

Is it worthwhile bringing such possibilities in a technical forum? But then you said:

“3. Is it possible that the Big Ben puts out a clock signal with "tuned" jitter at certain frequencies such that clocking from it MIGHT yield better results than an internal clock in certain situations? I know you have said that if the Big Ben is the clock master then it has no feedback from the downstream devices with which to create an "improved" clock over the device's internal clock. Is it not possible, however, that Apogee knows enough about both the internal crystal and the PLL of a given A/D converter that they can design a clock system which, when coupled with the receiving PLL, produces jitter at more desirable frequencies than the internal clock in the A/D converter?”

No!!!

The path from clock input to AD clock input is VERY INDIRECT. It is not like you can “grab a clock edge” and “move it”. The PLL itself is a very “soft rubber band”. Also, no one has knowladge of random behaviour before it happend. After it happened it is too late to fix!  


This part of my message is not a rehash. I welcome occasional questions from people that do not know the fundamentals of electronics, or wish to get into an area that I can help with. But you have posted nearly one message for each 2 of mine in this forum (where I am the moderator). Most of them are not questions, but answers. You said:

“We do a lot of re-interpretting information on these forums. As for moderator - I don't pretend to be one. I gave up that responsibility on this site many months ago”.

Nika,
I do not know who “WE” are. The audio police? A special audio elite group?
My personal criteria for re interpreting is:

1. Anyone that wishes to take the role of re-interpretting should have some technical chops in the area of discussion.  

2. re-interpretting  is not the same as changing the signature at the end of a virtually repeated message. Such practice has an element of claiming undue credit.

Also, the reason many people come here is to learn. Yet, you stated “I found your answers to be unsatisfactory” where someone else response to the SAME THREAD was “That is the clearest description of stochastics relating to quantitization I have ever read, great job!”. There are a lot of people that appreciate the forum, and that is good enough for me.

”4. It is my impression that you are wholly opposed to the presence of Big Ben in the industry. From what my clients tell me there are several situations wherein it improves the clock signals in ways that make sense to me.”

First, this last point does not belong here. This forum is not where I come between you and what you said to the client. I would hope that after reading my clock posts one would have enough sense to at least realize that an external clock jitter is inferior to internal operation. I hope that people realize that 192KHz is a crock. I hope they understand that dynamic performance is 6dB/bit and there is no real such thing as real 24 bit performance. I hope they don’t pay 1000$/ foot for some gold cable. That and more may be my main motivation for running the forum, but I am not going to get between a salesman and his clients.

Regarding the statement about “several situations wherein it improves the clock signals in ways that make sense to me.”  This is an opportunity to contribute. Explain what situations and how and why is works, and what makes sense to you. By all means! I promise to be open minded to anything that is not complete bull.

More specifically to your question:
I am not in the practice of categorically opposing specific gear. I am against  hype and misleading the customers with a bunch of bull. So yes, I am pretty down on the Big Ben. What happened to Ardvark? They had a good clock and could not stay alive. Is it because people bought into that low jitter crock marketing bull story about “external clock will make it sound better”? I can not prove that is what did it to Ardvark , but I suspect so.

Dan Lavry  

Logged

Lucas van der Mee

  • Newbie
  • *
  • Offline Offline
  • Posts: 12
Re: Proper word clock implementation
« Reply #125 on: December 16, 2004, 09:08:06 PM »

Dan ,
You think it is ok to down our products without knowing how they work , spec out and/or perform. You use your platform for pure marketing against us, while you claim in the same response that is not on topic.

The reason why people prefer the Big Ben over Aardvark is very simple. We perform better specs wise, we offer more features and our box is better value.
You have yet to design a product that can beat the Big Ben, so I understand your frustration. But I do not understand your abuse of the PSW platform to release this anger. I consider that to be a sign of very low quality.


Lucas van der Mee,
Sr. Design Engineer
Apogee Electronics
Logged
Lucas van der Mee
Sr. Design Engineer
Apogee Electronics

bobkatz

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2926
Re: Proper word clock implementation
« Reply #126 on: December 16, 2004, 09:24:01 PM »

Lucas van der Mee wrote on Thu, 16 December 2004 21:08

Dan ,
You think it is ok to down our products without knowing how they work , spec out and/or perform. You use your platform for pure marketing against us, while you claim in the same response that is not on topic.

The reason why people prefer the Big Ben over Aardvark is very simple. We perform better specs wise, we offer more features and our box is better value.
You have yet to design a product that can beat the Big Ben, so I understand your frustration. But I do not understand your abuse of the PSW platform to release this anger. I consider that to be a sign of very low quality.


Lucas van der Mee,
Sr. Design Engineer
Apogee Electronics



I was supposed to get a sample of the Big Ben to test. What happened?
Logged
There are two kinds of fools,
One says-this is old and therefore good.
The other says-this is new and therefore better."

No trees were killed in the sending of this message. However a large number of
electrons were terribly inconvenienced.

bobkatz

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2926
Re: Proper word clock implementation
« Reply #127 on: December 16, 2004, 09:32:20 PM »

Dan Lavry has been extraordinarily correct technically in all his responses about clocking and how it works, the principles of internal and external clocking, including one recently to Nika today. Since you cannot defeat the laws of physics, and therefore it should be Apogee's onus to demonstrate, via blind listening tests and published technical measurements how an external clock could possibly make a properly-performing converter perform "better".

I personally do not see Dan using this forum to promote one person's products or denigrate anyone else's. If he says he is "down on the Big Ben" it is not an attack on a product but on the unscientific methods by which that product has been promoted.  I do see Dan poking holes in other people's claims to beat the laws of physics, and he does it very well.

Please, let's not have another round of subjective arguments unsupported by technical measurements or facts. Please, let's not have another round of fruitless repeats of the same claims and counterclaims.

I have offered to perform objective tests and measurements and subjective blind tests. Send me a Big Ben for testing and let the tests begin.

BK
Logged
There are two kinds of fools,
One says-this is old and therefore good.
The other says-this is new and therefore better."

No trees were killed in the sending of this message. However a large number of
electrons were terribly inconvenienced.

Level

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1811
Re: Proper word clock implementation
« Reply #128 on: December 16, 2004, 10:10:39 PM »

That would be a nice one to put to bed Bob. Implementation of the clock, you could give the perspective. I use my internal clock and I just don't get all these clock issues, although I would love to be proven wrong, if my internal is giving problems I do not hear, nor anyone else has noticed to date.
Logged
http://balancedmastering.com

"Listen and Learn"
---Since 1975---

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: Proper word clock implementation
« Reply #129 on: December 16, 2004, 11:12:22 PM »

Dan,

Thank you for taking the time out of your busy schedule to reply.  Your post was a composite response from more than one post on more than one topic.  I am going to try to keep this technical and will reply to the non-technical, non-clock related issues offline or in another thread.  You are welcomed to post them back here if you wish, but I believe that that line of discussion detracts from the issue of clocks and jitter that we are slowly hashing out here.

danlavry wrote on Thu, 16 December 2004 19:40

I never stated it (that big ben has no crystal clock), and in fact if you go and read that clock thread, I assumed the best case ? that they do have an internal crystal, only at higher frequencies. One of my points was that a fundamental harmonic crystal is less jittery than higher frequency crystal. But that was not my main point. Here you are making me rehash what I already said.


OK, sorry, I don't mean to rehash.  I heard a quote attributed to you and it was apparently incorrect.  I do believe that Big Ben has some sort of crystal inside of it.  Moving on.

Quote:

I am not against external clocks when they are needed. I already stated that whenever one can use a reasonable quality internal clock, located inside, near the AD, they will have less jitter than using external clock, sending it through a transmitter, cable, receiver, and phase lock loop circuitry pushing and pulling a VCXO crystal that is more jittery than a fixed one...  Do I need to rehash it again?


No.  I agree with this in principle - I believe I have a chapter in my book dedicated to this concept.  I read Bob's book where he draws the same conclusion.  I have read your papers wherein you draw the same conclusion.  I think we are all pretty clear that, all things being equal, an internal clock running at the oversampled rate is better than the complex transmission path through which an external clock must go.  This is not to say that there aren't exceptions to that rule.  A box could have, for example, a poor quality internal crystal that is highly susceptible to jitter (poor grounding, power supply, whatever) yet a really stable interior PLL.  I would think this situation would be rare, but it is the type of situation that could yield an improved jitter spec from an external box.

Another example would be a box that has no internal PLL and a poor quality internal clock, but which accepts clock signals at the oversampled rate.  The first time I was ever exposed to improvements in signal quality using an external clock it was in this situation and the difference was very noticeable.

These are two situations wherein an external clock would yield an improvement beyond the internal clock in an A/D converter.  There are assuredly others.  But again, in principle, and all things being equal, in theory this should not happen.  The big question right now - especially looking at the market response - is whether or not this situation is a "theoretical" situation or one in practice.  Of course, the difference between theory and practice is that, in theory, they are the same...

Quote:

I already explained that a ?clock box? can not send a signal to reduce jitter in another box, without knowing what is taking place at that other box. The idea of being able to correct something without having knowledge of what it is bogus! I stated that EVEN IF the clock box were able to send ANY signal to correct the jitter inside the AD (AND IT DOES NOT!), it would not know what signal will correct the jitter, because the clock box DOES NOT KNOW what is wrong in the AD box ahead of it.
That is so because it is a one way communication- No feedback. The clock box sends THE SAME signal to any AD box. It does not change the type of signal for different AD?s. I already said all that. Do I need to rehash that?


OK.  I'm following this concept and always have followed this concept.  Allow me to make a dumb analogy:

In your D/A converter design, using Crystal Lock technology (I believe you call it) you incorporate what I think is a rather ingenious scheme.  You essentially decouple the internal clock from the external clock, only occasionally (every 10 seconds?) reconciling the two.  This is aided by a FIFO buffer that makes sure that we never lose a sample in a 10 second period.  When you reconcile the two clocks you make a change to the internal clock to accommodate the external clock's variance.  This yields very high jitter but at extremely low frequencies (.1Hz?) which yields sidebands so close to the fundamental that they are effectively inaudible.  

One might say "But you can't decouple the two clocks from each other - what if the external clock falls so far out of sync that you get gross error between the two - outside the capabilities of the buffer."   The answer is that you, Dan, know enough about clocks in use in the industry that you know exactly how much variance would be commonplace - I believe about 150 ppm?  I forget the specifics.  Knowing that crucial piece of information you can design that circuit such that the external clock and the internal clock will never fall so far out of sync that the buffer will overrun or underrun and you will never lose a sample.  It is that key piece of information - the maximum amount of variance you'll find in a common clock signal - that allows you to design this circuit.  If you didn't have that you'd have to design a much tighter PLL - and you do for purposes of slaving to SMPTE, etc.

OK, I might have a few details messed up in there.  Please forgive them.  What I am driving at is this:  Is it possible that Apogee (or Aardvark, or dCS, or some other company that makes an external clock) might know a crucial piece of information about circuit design of clocks or PLLs in typical devices (or specific devices) in this industry that would allow them to modify their external circuitry accordingly to take advantage of a "weakness" in the internal design?  This isn't an example of some sort of "feedback" that indicates something to the external box.  This is an example of knowing ahead of time what the weaknesses are so that they can be accommodated for certain, specific boxes that may be the target of the manufacturer's market?

Allow me to give a hokey example.  Perhaps there is some A/D converter on the market that, for whatever reason I don't know, performs better with a higher amplitude clock signal than typically comes off of the average crystal circuit.  However, I happen to know that certain PLLs generally put out a hotter signal.  As a designer, knowing this information in advance, and knowing that those A/D chips and that PLL design are prevalent in boxes in the specific market that I'd like to address, I could devise an external clock that would cause this specific circuit to sound better than on its internal clock.  

Now that's a totally made up example as you know, but I hope you understand my point.  Is there NO advance knowledge about clocks, PLLs, circuitry and the conventional box in this industry that a manufacturer could use in order to trump the conventional wisdom that internal is better?

I removed the next substantial portion of the message and will reply elsewhere so as to keep this on topic regarding clocks

Quote:

?4. It is my impression that you are wholly opposed to the presence of Big Ben in the industry. From what my clients tell me there are several situations wherein it improves the clock signals in ways that make sense to me.?


Quote:

First, this last point does not belong here. This forum is not where I come between you and what you said to the client.


Not what I said to the client.  What the client said to me.

Quote:

I would hope that after reading my clock posts one would have enough sense to at least realize that an external clock jitter is inferior to internal operation. I hope that people realize that 192KHz is a crock. I hope they understand that dynamic performance is 6dB/bit and there is no real such thing as real 24 bit performance. I hope they don?t pay 1000$/ foot for some gold cable. That and more may be my main motivation for running the forum, but I am not going to get between a salesman and his clients.


I am not a salesman.  I left Sweetwater some time ago.

Quote:

Regarding the statement about ?several situations wherein it improves the clock signals in ways that make sense to me.?  This is an opportunity to contribute. Explain what situations and how and why is works, and what makes sense to you. By all means! I promise to be open minded to anything that is not complete bull.


Big Ben has a lot of features.  One is that it is a clock distro - and a good one at that.  It therefore improves clock signals over daisy chaining through many boxes and in situations wherein T'ing off is inappropriate.

There are some boxes in this industry (the 888/24 from Digidesign, for example) that have very poor internal clock circuits and can be dramatically improved from an external device running superclock.

Big Ben will soon be able to generate wordclock locked to black.  A higher quality box than others on the market that can do that would be welcomed.  

These situations all make sense to me - especially the first at this point.  I am hearing a lot of positive feedback from people using Big Ben for its many features.  Some of these people are touting it for its as-of-yet-unreconcileable better performance than an internal clock, but not all.  

Quote:

More specifically to your question:
I am not in the practice of categorically opposing specific gear. I am against  hype and misleading the customers with a bunch of bull. So yes, I am pretty down on the Big Ben.


From what I have seen (and I don't see everything) Apogee has not been marketing Big Ben as a "fix your internal clock" solution.  It is their users that are touting it that way.  I could be wrong.

Quote:

What happened to Ardvark? They had a good clock and could not stay alive. Is it because people bought into that low jitter crock marketing bull story about ?external clock will make it sound better?? I can not prove that is what did it to Ardvark , but I suspect so.


I believe Aardvark's problems started long ago.  Perhaps that's fodder for a new thread, but their PC computer solutions did not do as expected.  Further, competition on the market and improved designs that obfuscated the need for Aardvark's solution eroded their market.  Those are just my suspicions.

Again, Dan, thank you for your time.  I think you can see that I mean well, am not trying to waste your time, and am trying to keep this technical, per your wishes.

Nika
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: Proper word clock implementation
« Reply #130 on: December 16, 2004, 11:38:55 PM »

bobkatz wrote on Thu, 16 December 2004 21:32

Dan Lavry has been extraordinarily correct technically in all his responses about clocking and how it works, the principles of internal and external clocking, including one recently to Nika today. Since you cannot defeat the laws of physics, and therefore it should be Apogee's onus to demonstrate, via blind listening tests and published technical measurements how an external clock could possibly make a properly-performing converter perform "better".


Bob,

I hope you do not mind me posting a few segments from this archive here.  The following is a segment from a dialogue that took place from November 1 through November 7 of 2001 between David Josephson, Julian Dunn, yourself, and many others on this very topic.  The evidence that it would be possible to best the performance of an internal clock with an external clock was compelling enough that you changed something in your book:  Responding to Ben Gilsdorf's reason that an internal clock might be subject to more problems than it's accompanying PLL you write: "...And so do I... I'll have to revise my book....

BK"

This is the dialogue that led up to that:

Bob Katz:
>  I've started a running argument in the Digital Performer maillist. My
>  basic contention is that converters will inevitably perform better
>  (jitter-wise) on internal crystal sync unless the designer of the
>  converter was brain-dead or made some severe design compromises. No
>  one denies the contention that a phased lock loop can never exceed
>  the performance of a perfect crystal oscillator, the best it can do
>  is equal that performance. And typically, be worse, at least
>  measurably.
>
>  However, many critical listeners report better sonic performance from
>  converters such as the Digidesign 888 under stable external WC sync.
>  I find it very hard to believe that the internal crystal oscillator
>  of the 888 is so bad that it can't do better on internal! Perhaps
>  this is just a matter of "the lesser of two evils".

David Josephson:
Bob, you of all people should be careful about ignoring critical
listener reports in favor of supposed theory!  A crystal oscillator is
no magic device; the crystal is simply a very high Q mechanically
tuned circuit that passes signal predominantly at one frequency. Put
it in a feedback loop around an amplifier and it oscillates, same as
any other tuned circuit. But the bandwidth is not infinitely narrow,
the Q is not infintely high. Design of a low phase noise crystal
oscillator is not terribly difficult but it isn't as simple as connecting
it around some CMOS inverters like you see in most digital implementations.
Different crystal cuts have different bandwidths too; sometimes at the
expense of frequency stability over temperature. I can very easily
believe that a good phase locked loop would perform better in terms of
phase noise in the audio band than an ordinary crystal oscillator
would.

Bob Katz:
>  "In general, a typical converter will likely perform better on
>  internal sync than external".

David Josephson:
In general I think that's probably so. But the reader should be
cautioned that if it seems that it sounds better on external
sync, believe your ears.

Bob Katz:
>  It just has always stood to reason.. it makes so much engineering sense.

David Josephson:
As someone who has designed and built many crystal oscillators over
the years, mostly for radio applications, I will refrain from commenting
on whether it makes so much engineering sense.


later:

Bob Katz:
>  Then let's talk practical engineering at the price/performance levels
>  of a Pro Tools 888.
>  Now I don't know shinola about what's inside an 888, so I could be
>off base...
>
>  If, as opposed to the crystal-around-inverter-trick, Digidesign were
>  to buy an off-the shelf DIP-socket type rectangular metal can master
>  crystal oscillator, feed it from a clean, bypassed power supply,
>  grounding, etc. This is "engineering by Xerox" or at least buying
>  someone else's module, but it is doable at this level of dollars. But

David Josephson:
A quick look at off-the-shelf crystal oscillator modules in the Digikey
catalog reveals only one with a phase noise or jitter spec, and it's 30
psec typical, 50 maximum. It would be safe to assume that the more
generic parts are worse than that.

Bob Katz:
>  to my knowledge, you can't buy a good PLL off-shelf... it has to be
>  weaned and wined and dined and tweaked and double-PLLed until it's
>  blue in the face.

David Josephson:
If you are in the business of producing such equipment, it is customary
to have on your staff at least one competent analog designer with some
background in PLL design. Techniques for combining off-the-shelf VCOs,
phase detectors and reference oscillators to produce a predictable level
of PLL performance have been well documented for at least 30 years.

Bob Katz:
>  Anyway, the question, based on the above is: Which is more likely to
>  perform better, an ordinary PLL designed with a few hours of
>  engineering, or the off-the-shelf crystal oscillator module... which,
>  internal, or external, is likely to perform better?

David Josephson:
While I agree that in my experience, it is more likely that internal
clock will produce lower jitter at a given converter, there are enough
variables to make a blanket statement that it "will sound better" unwise.
There is also the question of the spectrum of the jitter; while the
PLL will almost certainly have higher jitter at low frequencies, a
crystal oscillator may well have significant phase noise (which becomes
jitter) at high frequencies in the audio band, which may mean accentuated
jitter artifacts at certain frequencies of program material (this is
possible with PLLs too).

later contribution by Julian Dunn:

>Another explanation is that the phase alignment under external WC sync
>reduces some crosstalk effect.  If the comparison of "sonic" qualities
>is made with the WC connected and a switch selecting between internal
>and WC sync then on internal there would be the potential for crosstalk
>from the WC input.


There was more.  I picked what I thought was probably critical to your revision on this topic in your book.  Of course, I made the same point in my own book, but I think you and I both allowed for the notion that in some conditions boxes like Big Ben COULD provide better results - or at least one should be open to it and do their own listening tests.  This notion is directly contradictory to the point made so emphatically by Dan when he writes, regarding whether or not Big Ben could improve on an internal clock, "No."

Cheers!
Nika
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

bobkatz

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2926
Re: Proper word clock implementation
« Reply #131 on: December 17, 2004, 01:59:17 PM »

I am perfectly aware of all the nuances, Nika. Which is why more than once I have said, "A poorly-designed converter (one with a poor internal clock) can conceivably sound better on external sync".

Then you may recall, in this thread, the guys from Apogee replied something to the effect that "many respectable engineers using what are recognized as excellent converters feel that the Big Ben improves their sound".

This sounds exactly like a marketing attempt to look a the tail side of the penny instead of the head.

Bob (and any properly-schooled engineer) says:

"By definition, if a converter sounds better on external clock, then it must be defective. At the very least its internal clock must be poorly designed, and its PLL may also be poorly designed".

I believe that the marketing geniuses at Apogee have turned this fact all around into marketing hype. "Put our expensive bandaid on your bad converters to make them sound better".

I will attempt to settle these issues and many more with definitive measurements and blind listening tests. Apogee is supposed to send me a Big Ben. This was at least a month ago and I haven't seen it. Of course I can see their reluctance to send a unit to a "biased" tester, but if I'm biased, it's thoroughly on the side of science. Fortunately, I have ears, too!

BK
Logged
There are two kinds of fools,
One says-this is old and therefore good.
The other says-this is new and therefore better."

No trees were killed in the sending of this message. However a large number of
electrons were terribly inconvenienced.

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: Proper word clock implementation
« Reply #132 on: December 17, 2004, 02:07:25 PM »

Bob,

I understand.  I guess I don't see their reports of what other people say as a marketing department gone amuck.  They aren't claiming that the box does one thing or another.  They are claiming that other people have stated that it does one thing or another.  I remember Aardvark claiming "plug this in and it will make your stuff sound better."  That's a bold marketing claim, in my opinion.  Claiming that some reputed engineers found it improved their signal chain is perhaps incomplete, but not overly hyped from my perspective.  Perhaps that's a nuanced position?  Perhaps it's a fine line?  I'm not sure.  Again I am led to believe that it's Apogee's users that are marketing its sound quality so well - not Apogee.

I hope Apogee sends you one.  To your credit, your tests have always been scientifically conducted and complete with thorough results and whatever caveats were necessary.

Cheers!
Nika
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Proper word clock implementation
« Reply #133 on: December 17, 2004, 03:51:15 PM »

Nika Aldrich wrote on Fri, 17 December 2004 04:38

bobkatz wrote on Thu, 16 December 2004 21:32

Dan Lavry has been extraordinarily correct technically in all his responses about clocking and how it works....


Bob,

I hope you do not mind me posting a few segments from this archive here....Cheers!
Nika


This is the dialogue that led up to that:

David Josephson:
Bob, you of all people should be careful about ignoring critical
listener reports in favor of supposed theory! A crystal oscillator is no magic device; the crystal is simply a very high Q mechanically tuned circuit that passes signal predominantly at one frequency. Put it in a feedback loop around an amplifier and it oscillates, same as any other tuned circuit. But the bandwidth is not infinitely narrow,
the Q is not infintely high….


I do not have a strong disagreement with Mr. Josephson. I find the way he states matters to be a lot “softer” than mine. I respect his method of leaving a lot of room to insure that nothing wrong is being said.

However, my view of crystals is a very definite.  So I will explain.  David J. said that a crystal a  simply a very high Q mechanically tuned circuit, which may seem to the casual observer as no more then some “piece of  metal” like a tuning fork. In fact, the whole concept of the crystal is its very orderly arrangement of the molecules themselves. It is tuned on a molecular level, and it has 3 axis’s: mechanical, electrical and optical. Think of a 3 dimensions “bee hive” type of molecular organization.

Indeed, it does not have infinite Q, but I would go beyond saying “is simply a very high Q” device.  I would say it is an INCREDIBLY HIGH Q. To quantify it:
A “standard tuned circuit” with Q=100 is very high Q in my book. A crystal Q ranges from 10^4 to 10^7 (from 10000 to 10000000). See the book “Design of Crystal and Other Harmonic Oscillators” by Benjamin Parzen. My point about the fundamental mode oscillators providing less phase noise can be found at the book “Crystal Oscillator Circuits” by Robert Matthys.

The complete equivalent circuit of a crystal is complex, but the simplified version of an inductor, series resistor, series cap and a parallel cap yields a good first approximation.

The equivalent inductor is HUGE. It can be as high as Henries(!). For a reference, 1mH (mili Henry) is a big inductor, lots of core material, and many many windings. 1 Henry is 1000 time bigger. High inductance at our MHz frequencies is a problem for many reasons.
   
The equivalent crystal series capacitor is as low as a few pF. For a reference, if I recall, 2 squares plates 1 square mil each (1/1000 of an inch), separated by 1 mil yields about 1pF.

Can you imagine trying to build a tune circuit with such values from such huge inductor and tiny cap?  The crystal provides us just that.

The Q goes up with increased inductance and decreased capacitance (resistance is the third factor). As an analog designer, I do not see any way to get near such a high Q using anything other than a crystal using inductors and caps. A crystal device really stands out there when compared to all other tuned circuits.

“There was more. I picked what I thought was probably critical to your revision on this topic in your book. Of course, I made the same point in my own book, but I think you and I both allowed for the notion that in some conditions boxes like Big Ben COULD provide better results - or at least one should be open to it and do their own listening tests. This notion is directly contradictory to the point made so emphatically by Dan when he writes, regarding whether or not Big Ben could improve on an internal clock, "No."

You seem to be looking for that a possible scenario where a PLL driven by external clock is better than internal. You will have to find a design with a terrible internal clock and a fantastic PLL. But after you find it you could still not say that the external clock improved the jitter. All you can say is that that particular internal clock sucks.

The jitter you end up with when using the external connection is the INHERIT PLL circuit jitter PLUS more jitter due to the external clock, PLUS the interconnects, PLUS the PLL and the rest. What I call the “INHERIT PLL circuit jitter” is the jitter of the PLL circuit itself with no additional contributions. As if it is driven by zero jitter clock theoreticaly perfect clock source. The best way that I know to test it is to drive it with as low as possible jitter fixed crystal, connected directly to the PLL input...

Using an external clock ADDS to the inherit jitter. It does not reduce the inherit jitter. Driving a PLL circuit with a clock does not reduce it’s jitter. The question is how much does it adds to it. Lets all be clear about it. Many people got the impression from advertisements that Big Ben will do just that, reduce the jitter.

Mr. David J’s suggestion is that there is a possibility, though not likely, that a PLL based circuit will have less jitter than a poorly designed fixed crystal. My question to him is: A PLL requires an oscillator. What kind of oscillator will yield the lowest jitter? An RC and a comparator? An LC with a Q of 100? Or is it a pullable crystal, which is almost as good as a fixed one? (A fixed internal crystal is of course the proposed best way to go). Also, as you must know, it is important to have a low jitter oscillator as a building block for the PLL circuit. The PLL loop filter itself cannot remove all the oscillator jitter.

Dan Lavry


Logged

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Proper word clock implementation
« Reply #134 on: December 17, 2004, 08:31:36 PM »

Nika Aldrich wrote on Fri, 17 December 2004 19:07

Bob,

I understand.  I guess I don't see their reports of what other people say as a marketing department gone amuck.  They aren't claiming that the box does one thing or another.  They are claiming that other people have stated that it does one thing or another....
Cheers!
Nika



Do you really believe what you said? Are you saying that they have nothing to do with statements such as:

"Happily a solution is possible so long as the device in question can synchronize to an external clock." ???

Also guess who said the following on this thread? In was Lucas

“In other words, when you HAVE to clock to an external , jittery, source, you will find that your results will be much better if you have the Big Ben do the clocking first and clock your device to the Big Ben. This is objective, measurable and verifiable.”

First, why not clock to a clean source such as a crystal? If you do, you will not need to buy a box to clean the jitter, and a good crystal oscillator will beat a cleaned jittery clock. Big Ben cannot claim to do better than provide say a word clock 0101 pattern, hopefully with very low jitter.  A lot of words about cleaning jitter, sounding better, adaptive filtering and high end technology do confuse the customer. Heck, they tried it on me and everyone else on this web. And what do they say when it does not make sense?

Lucas also said:
“Then there is the much more difficult to explain phenomenon of converters starting to sound “better” when clocked to the Big Ben as a master clock only; the C777 running on internal. That is something we found empirically and like I said in my previous posting, we researched and optimized.”

What can I say? That statement from an engineer? For some reason it keeps bringing the joke into my head… the punch line was “cross the field where the cows graze and don’t step in the bull….”.

Apogee employees did come here trying to tell people that what I said is false.(ie: about internal clock being better than external clock.) Was it not the reason they came here?

Dan Lavry

Logged
Pages: 1 ... 7 8 [9] 10 11 ... 15   Go Up
 

Site Hosted By Ashdown Technologies, Inc.

Page created in 0.111 seconds with 20 queries.