Print

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

[Johnny B]

The following link is to a white paper entitled "A 'Better Than' Nyquist Pulse"

This paper appears to show that Nyquist is not quite the "end-all, be-all" that some would want us to believe...that, in fact,...there just may be significant room for some more digital innovation...like in the area of timing...

I want to thank Eric Bridenbaker, the original poster, for this link...

     http://www.ee.ualberta.ca/~beaulieu/online_pubs/ncb_cct_mod_ clett.pdf

[Gunnar Hellquist]

Hmm, surely a very interesting topic. Only question I have is how often do you yourself work with Nyquist pulse-s ? They are an interesting tool for some measurements surely, but surely has very little relevance to what most of us do?

Gunnar

[Johnny B]

Emmm, if you are doing any digital at all, Nyquist is a big deal and very much part of the process that will be used. Here, this paper suggests that Nyquist can be improved...all improvements are, by definition, good.  

[Gunnar Hellquist]

I believe you are a bit off. What the paper talks about is the Nyquist Pulse, which is something very specific if I remember my theory right. Sinc is a special kind of Nyquist pulse though. The theory area we are talking about is how to transfer digital signals over real-world analog channels, without getting inter-symbol-interference. You might be way ahead of me on signal theory though, so help me out if there is something I am missing.

Gunnar

[Jon Hodgson]

Nyquist was a mathematician working in the area of communications theory who did more than one thing in his life. Just because it says Nyquist on it doesn't mean it has any relevance to digital audio.

The Nyquist pulse is distinct from the Nyquist-Shannon sampling theorem.

In simplistic terms the idea of the Nyquist Pulse that when you have to choose a pulse shape for digital transmission which has the best chance of being correctly read at the other end, you want a pulse which is as strong as possible whilst interfering with adjacent pulses as little as possible. The Nyquist pulse is an attempt to find the best shape for this.

So this paper is potentially good news for digital transmission, but won't make any difference to you in the studio.

[Johnny B]

Oh, I see...so no one uses networks or transfers digital audio signals around from place to place...

Interesting...

The only other observation is that Nyquist's work was done way back in 1928, well before computers, Shannon's work was done in the early 40's, so you've had over 60 years to make improvements. And we *do* see from this white paper that some of Nyquist's work *can be* improved.



 

[Johnny B]

And there are people who feel that digital audio sounds about as bad as telephones, they prefer the sound of analogue. There are quite a few people in this "Analogue Lover" camp.

Therefore, it seems to me that it behooves those working in the digital audio field to leave no stone unturned in order to improve the sound quality of digital. Now, I'll be the first to admit there are are heated arguments about the best way to go about making those improvements, there are arguments about where the focus, the priorities, the funds, and the effort ought to be expended, but I do think there is widespread agreement that there is quite a bit of room for digital sound quality to improve.

 

[Jon Hodgson]

Johnny B wrote on Sun, 02 October 2005 23:11

Oh, I see...so no one uses networks or transfers digital audio signals around from place to place... Interesting... The only other observation is that Nyquist's work was done way back in 1928, well before computers, Shannon's work was done in the early 40's, so you've had over 60 years to make improvements. And we *do* see from this white paper that some of Nyquist's work *can be* improved.

Ok, if you want to be pedantic, I should have specified
"it won't make any difference to the sound in your studio unless of course you happen to be transmitting stuff in digital across huge distances with inadequate error correction"

But I assumed the people reading here would have the intelligence to understand that as given.

[Jon Hodgson]

Johnny B wrote on Mon, 03 October 2005 02:27

And there are people who feel that digital audio sounds about as bad as telephones, they prefer the sound of analogue. There are quite a few people in this "Analogue Lover" camp.

As bad as telephones? My god you must have bad hearing if you think that

Johnny B wrote on Mon, 03 October 2005 02:27

Therefore, it seems to me that it behooves those working in the digital audio field to leave no stone unturned in order to improve the sound quality of digital. Now, I'll be the first to admit there are are heated arguments about the best way to go about making those improvements, there are arguments about where the focus, the priorities, the funds, and the effort ought to be expended, but I do think there is widespread agreement that there is quite a bit of room for digital sound quality to improve.

There's also room for analogue sound to improve, people are working on both. Ironically a huge part of the work in digital audio involves emulating the distortions in analogue because people like them... some have even come to believe that the distorted, bandlimited, compressed and frequency modulated sound they hear coming off their turntable is how the musician actually sounded in the studio.

[Johnny B]

Agreed that analogue could also be improved...that's the nature of all tech...constant improvement....

But in regard to a lot of analogue distortion...it works with the harmononics in a pleasing way...

With digital distortion...it seems to be in the ugly odd harmonics which people do not like at all...

Big difference...

Analogue distortion often sounds more musical...whereas, digital distortion is just ugly noise...and that's a huge difference  

[PookyNMR]

Interesting paper.  I went to the University of Alberta.  The Engineering dept was tops.

Maybe I'm dense here, but I fail to see how this applies to digital audio.  Nice paper though.

[Johnny B]

Well you have the digital audio communication thing...so that's relevant.

The second point is that Nyquist was improved on in one area...I know it's a bit of a leap for some to get their heads around...but it could mean other areas of Nyquist are worth a re-exam to see if similar improvements can be made...so it's relevant on that account...

I suppose there are those who might say everything he did is never subject to any improvement whatsoever...but I suspect those individuals may be in the Luddite camp.

Everything in the universe can be improved...at least everything mere mortal men have done can be improved upon...Man is not perfect, all his schemes, ideas, therories, and inventions can all be improved...

[Jon Hodgson]

Johnny B wrote on Mon, 03 October 2005 08:18

Well you have the digital audio communication thing...so that's relevant. The second point is that Nyquist was improved on in one area...I know it's a bit of a leap for some to get their heads around...but it could mean other areas of Nyquist are worth a re-exam to see if similar improvements can be made...so it's relevant on that account... I suppose there are those who might say everything he did is never subject to any improvement whatsoever...but I suspect those individuals may be in the Luddite camp. Everything in the universe can be improved...at least everything mere mortal men have done can be improved upon...Man is not perfect, all his schemes, ideas, therories, and inventions can all be improved...

Tell me....

How can you improve on 100%???

Sampling at frequency more than twice your signal bandwidth preserves 100% of the information in your original signal, proven mathematically, experimentally, and practically.

After you have 100% of the information, sampling any more is just adding redundancy, it gives you nothing extra.

There are a number of places in which digital systems can be made more accurate, or more pleasing (the two are not neccessarily the same thing),

Luddite? Well if understanding and accepting the Nyquist-Shannon sampling theorem (and incidentally Nyquist admitted that he was only formalizing what was already known for many years) makes me a Luddite, then I'll accept it. You can also call me a car luddite, because I believe that as long as we use wheels, they should be round.

[Ronny]

Johnny B wrote on Mon, 03 October 2005 03:18

Well you have the digital audio communication thing...so that's relevant. The second point is that Nyquist was improved on in one area...I know it's a bit of a leap for some to get their heads around...but it could mean other areas of Nyquist are worth a re-exam to see if similar improvements can be made...so it's relevant on that account... I suppose there are those who might say everything he did is never subject to any improvement whatsoever...but I suspect those individuals may be in the Luddite camp. Everything in the universe can be improved...at least everything mere mortal men have done can be improved upon...Man is not perfect, all his schemes, ideas, therories, and inventions can all be improved...

Nyquist merely stated that to capture the full wavelength of a frequency, you need to sample at twice that frequency. It was a theory until 1948, when Claude Shannon proved it. Improving what Nyquist stated is moot, as it's a physical law. He never said not to use higher sample rates and I've heard you inscinuate that Nyquist is solely responsible for 44.1k. If you used 384k sample rate, that you were saying would be so much better, Nyquists theorum would still apply, you could only capture freq's up to 192k effectively, as the sample rate would need to be twice the highest freq recorded, proven without a doubt.

[PookyNMR]

Johnny B wrote on Mon, 03 October 2005 01:18

Well you have the digital audio communication thing...so that's relevant.

That application is quite different, with a different set of problems.  

Johnny B wrote on Mon, 03 October 2005 01:18

The second point is that Nyquist was improved on in one area...I know it's a bit of a leap for some to get their heads around...but it could mean other areas of Nyquist are worth a re-exam to see if similar improvements can be made...so it's relevant on that account...

"a bit of a leap to get their heads around" - tell me, what education do you have in Math?  Any post secondary level?  Masters level?  Doctoral level?  We already have had brilliant PhD's look over these things and prove the theorm.  Do you not think that many gropus of highly educated and experience individuals have not already combed over it a hundred times?  Your statement is quite silly...

Johnny B wrote on Mon, 03 October 2005 01:18

I suppose there are those who might say everything he did is never subject to any improvement whatsoever...but I suspect those individuals may be in the Luddite camp.

I'd be extremely cautious if I were you using the word "Luddite" when you so obviously do not understand Math.  

Johnny B wrote on Mon, 03 October 2005 01:18

Everything in the universe can be improved...at least everything mere mortal men have done can be improved upon...

Then why don't we start with your knowledge of math, physics, and digital audio!