Complexity of digital audio

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
I mentioned a little story in another thread regarding swapping out a Toslink digital connection for an asynch USB in front of a rock band members/engineers. They asked me what the improvements would be. I honestly could not describe it to them and had to keep telling them to listen. If I had to give them an answer it would have been, "lots of things can change."

High fidelity digital reproduction is a complex thing. What if I said the timing for a single audio sample changes if the last bit of the previous sample was a "1" and what follows it is a "0." But not in other occasions. Can anyone explain the audible effects here? It won't change frequency response. It won't change THD measurements unless you happen to be feeding the DAC these precise values.

As hobbyist in this field, we have an intuitive sense for a number of concepts in audio. We know about frequency response variations. We know about noise floor. We know about distortion (at least in gross amounts) of say, an amplifier. We know about coloration of speakers. We know about room effects. But what I described above?

The above also calls for appropriate measures to hear and instrument digital system fidelity. And of course, deep, deep understanding of esoteric aspects of this technology. This Saturday I am planning on teach a mini-class to my team at the store on this topic. I wonder how successful I will be in explaining these nuances. The topic an get as deep as one wants it to get!

I am not sure even I have mastered it. The USB to S/PDIF converter I use has a jitter simulator. It changes its clock to a lower fidelity one. Since that is under the control of the system, it also has a specific jitter spec. When I flip that switch, there is a noticeable change in audio. As I started this post, the changes are all over. It is not just "the highs are a bit less bright." More surprising is that the low quality clock is not so low quality. Its maximum jitter is 1.2 nanosecond. I can't explain the level of difference I hear relative to the number I just stated.

What is ironic is that when one thinks of "digital," the complexity of what I just described is the opposite of what comes to mind. Indeed, digital is thought to be perfect, and 100% clean. And here comes explanations of its performance that gives one a serious headache. :)

So what is the purpose of this post? Well, I am not quite sure :). Other than saying that simple answers just don't do it for advanced digital audio reproduction. Saying it is all good because distortion is this low in digital doesn't either. Stuff is hard to explain, hard to characterize, and certainly very difficult to measure.

Comments?
 

garylkoh

WBF Technical Expert (Speakers & Audio Equipment)
Sep 6, 2010
5,599
225
1,190
Seattle, WA
www.genesisloudspeakers.com
Good luck Amir!!

You are getting into the what can I measure that I can't hear, and what can I hear that I can't (or don't know how to) measure. Digital is supposed to be very, very simple - 1 and 0 - how complex can that be? If a real-time 16/44.1 A/D/A loop has been proven to be inaudible by Meyer and Moran with all its jitter and errors, what hope can we have that we can ever explain that a lower jitter clock can be noticeably different in sound. Did you do a DBT on that USB to S/PDIF converter? :D

The comments you get during Saturday's session will be extremely interesting to me. It will show how different people think. I hope you post a summary.
 

Bruce B

WBF Founding Member, Pro Audio Production Member
Apr 25, 2010
7,006
512
1,740
Snohomish, WA
www.pugetsoundstudios.com
The comments you get during Saturday's session will be extremely interesting to me. It will show how different people think. I hope you post a summary.

Saturday's session?

What you have to realize also is that all digital audio workstation programs have different protocols for "drawing" a waveform. There is no standardization. You can have small differences of the same exact file in 2 programs. They are just a representation of the audio, it's not actually the audio. The original audio remains intact and unchanged and whatever you do in the editor is called an EDL (edit decision list). So this part is not exact either.
 

fas42

Addicted To Best
Jan 8, 2011
3,973
3
0
NSW Australia
Other than saying that simple answers just don't do it for advanced digital audio reproduction.
My take on this is that you have to always be very clear about whether you are talking about digital, or whether you are talking about analogue. If when talking of digital to analogue things people confuse or mix together these two concepts then they will only end up in a muddle ..

There is a clear divide, a fence between digital and analogue, which should always be kept there when talking about the concepts.

First of all, digital: this is always perfect, provided error correction is doing its job. No matter how much jitter, interference, even if it's mangled to the n'th degree; so long as the receving device is sensitive or smart enough to untangle the mess then the signal, as digital, is still perfect.

Next, analogue: we all know about this temperamental beast, flakey as hell, sneeze 10 feet away and you have problems! You have to apply massive amounts of TLC to this fellow to keep it happy ...

So, in a reasonable, not perfect(!), world, there comes a point when we want to turn the perfect digital into touchy analogue, within the D/A chip or whatever. This is the crucial place where it all happens, the two worlds collide. So this area has to be handled as well as it possibly can, and in this area the essential workings are considered to be analogue. In other words, TLC is needed, which is fine, people know how to go about this.

The big problem is, that at least two of the signals coming into this area are the digital audio data, and the clock. What are they, digital or analogue? Answer, both!! From the point of view of inside the D/A chip they are analogue, yes, they really are, but from the point of view of the design of the circuitry that sent them there, they are digital! This is where the headaches start ...

So what do we do about these two key signals? They must be as clean as possible going into the D/A chip, need to be looked at purely as analogue signals, so the answer is at some point we have to start cleaning them up, starting to forget they are digital, and beginning to think of them as analogue, purely analogue. The big discussion, argument, carry on then becomes, where, when and exactly how we go through this process of losing the digital concept and taking on the analogue concept.

Which is where the industry, as a whole, is now ....

Frank
 

Bruce B

WBF Founding Member, Pro Audio Production Member
Apr 25, 2010
7,006
512
1,740
Snohomish, WA
www.pugetsoundstudios.com
Which is where another fuzzy gray area can arise. You will need to divide digital into 2 different camps, PCM and DSD. DSD is more akin to analog in that it's either on or off.
 

Phelonious Ponk

New Member
Jun 30, 2010
8,677
23
0
I mentioned a little story in another thread regarding swapping out a Toslink digital connection for an asynch USB in front of a rock band members/engineers. They asked me what the improvements would be. I honestly could not describe it to them and had to keep telling them to listen. If I had to give them an answer it would have been, "lots of things can change."

So what did they hear?

Tim
 

Robh3606

Well-Known Member
Aug 24, 2010
1,471
463
1,155
Destiny
Hello Amir

Well for openers how do you know which type of connection will give you the best sound?? If so under what circumstances?? I have a universal player. I can use the Optical Toslink, digital Coax and HDMI plus of course analog. Playing a CD does it make a difference what I use and if so why?? I know I don't really understand the differences and limitaions of the 3 listed digital connections. Might not hurt to explain what if any differences there are under real world conditions?

Rob:)
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
Had to look up DSD... I can see why the controversy (pure 1-bit conversion has issues, though they can be overcome, or at least significantly reduced, with technology).

Amir, I suspect, as always, deterministic jitter (or sinusoidal jitter) is the culprit, thus 1.2 ns RJ (if that's the spec) is fairly meaningless. We listen through noise fairly well, but patterns, those we find.
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland
Am I too late for your Saturday mini-class on digital, Amir? How did it go?

Here's something which concurs with my listening & answers the question often asked "but what does jitter sound like?"

This is a quotation from Bob Katz, well known recording & audio mastering engineer posted here

After an engineer learns to identify the sound of signal-correlated jitter, then you can move on to recognizing the more subtle forms of jitter and finally, can be more prepared to subjectively judge whether one source sounds better than another.

Here are some audible symptoms of jitter that allow us to determine that one source sounds "better" than another with a reasonable degree of scientific backing:

It is well known that jitter degrades stereo image, separation, depth, ambience, dynamic range.

Therefore, when during a listening comparison, comparing source A versus source B (and both have already been proved to be identical bitwise):

The source which exhibits greater stereo ambience and depth is the "better" one.

The source which exhibits more apparent dynamic range is the "better" one.

The source which is less edgy on the high end (most obvious sonic signature of signal correlated jitter) is the "better" one.

And a reply:
The better one, and it is better, is also easier to listen to. . . less fatiguing. I would also add to this that the low end just "feels" bigger and more solid. This is perhaps a psychoacoustic affect more than a measurable one. It may be that the combination of a less edgy high end and greater depth and width makes the bass seem better.

All of this makes sense if thought of in terms of timing (that is what we're talking about isn't it ;-]). With minimal jitter nothing is smeared, a note and all its harmonics line up, the sound is more liquid (a term probably from the "audiophile" crowd but one which accurately describes the sound none the less), and images within the soundstage are clearly defined.
 

mep

Member Sponsor & WBF Founding Member
Apr 20, 2010
9,481
17
0
This thread makes my head hurt.
 

jriver

New Member
Feb 10, 2012
37
0
0
My take on this is that you have to always be very clear about whether you are talking about digital, or whether you are talking about analogue. If when talking of digital to analogue things people confuse or mix together these two concepts then they will only end up in a muddle ..

There is a clear divide, a fence between digital and analogue, which should always be kept there when talking about the concepts.

First of all, digital: this is always perfect, provided error correction is doing its job. No matter how much jitter, interference, even if it's mangled to the n'th degree; so long as the receving device is sensitive or smart enough to untangle the mess then the signal, as digital, is still perfect.

Next, analogue: we all know about this temperamental beast, flakey as hell, sneeze 10 feet away and you have problems! You have to apply massive amounts of TLC to this fellow to keep it happy ...

So, in a reasonable, not perfect(!), world, there comes a point when we want to turn the perfect digital into touchy analogue, within the D/A chip or whatever. This is the crucial place where it all happens, the two worlds collide. So this area has to be handled as well as it possibly can, and in this area the essential workings are considered to be analogue. In other words, TLC is needed, which is fine, people know how to go about this.

The big problem is, that at least two of the signals coming into this area are the digital audio data, and the clock. What are they, digital or analogue? Answer, both!! From the point of view of inside the D/A chip they are analogue, yes, they really are, but from the point of view of the design of the circuitry that sent them there, they are digital! This is where the headaches start ...

So what do we do about these two key signals? They must be as clean as possible going into the D/A chip, need to be looked at purely as analogue signals, so the answer is at some point we have to start cleaning them up, starting to forget they are digital, and beginning to think of them as analogue, purely analogue. The big discussion, argument, carry on then becomes, where, when and exactly how we go through this process of losing the digital concept and taking on the analogue concept.

Which is where the industry, as a whole, is now ....

Frank

Hi Frank,
I would like to quote this on our site and on computeraudiophile.com. Woud that be OK?

Admins? Any objections?

I think it's important to try to educate our customers and other users, and this is very well said.

Thanks,

Jim Hillegass, JRiver
JimH on our forum
 

Ronm1

Member Sponsor
Feb 21, 2011
1,745
4
0
wtOMitMutb NH
Which is where another fuzzy gray area can arise. You will need to divide digital into 2 different camps, PCM and DSD. DSD is more akin to analog in that it's either on or off.

Ok, now I'm really confused since when did analog just become on/off. I worked on a few analog computers in my early h/w days and they were hardly two states.
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland

FrantzM

Member Sponsor & WBF Founding Member
Apr 20, 2010
6,455
29
405
Which is where another fuzzy gray area can arise. You will need to divide digital into 2 different camps, PCM and DSD. DSD is more akin to analog in that it's either on or off.

:confused::confused:
 

mep

Member Sponsor & WBF Founding Member
Apr 20, 2010
9,481
17
0
Just when you thought digital really was perfect, along comes Amir...
 

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
Since Frank and the forum have joint rights to these posts, I can give you permission to do that. Thanks for asking. Please provide a link back to the source here.

And oh, welcome to the forum :). Your company's product is a topic of frequent discussion and interest here. So I hope you hang around.
Hi Frank,
I would like to quote this on our site and on computeraudiophile.com. Woud that be OK?

Admins? Any objections?

I think it's important to try to educate our customers and other users, and this is very well said.

Thanks,

Jim Hillegass, JRiver
JimH on our forum
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland
Ok, now I'm really confused since when did analog just become on/off. I worked on a few analog computers in my early h/w days and they were hardly two states.

yes, I have to agree - this is confusing :confused:

Jitter is confusing enough & often acts as a catch-all term for other distortions which may be the result of RFI or ground noise & can give rise to the same audible issues in the analogue domain i.e harshness & sibilance! But the DSD reference is confusing!
 

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
Am I too late for your Saturday mini-class on digital, Amir? How did it go?
I got a lot of glossy eyes :). That's the reason I wrote an article on digital audio jitter.

Here's something which concurs with my listening & answers the question often asked "but what does jitter sound like?"
Jitter can be all of those things. The problem is that they can also have other causes. I find two identifying clues most useful, one of which was mentioned:

1. Increase in high-frequency content/emphasis which he called edginess. It is easy to explain this mathematically. Jitter distortion products are predictable. If you have a 5 Khz signal and jitter is varying the time of those signals by 2 Khz, then you will have distortion at 2K Khz and 7 Khz. The 7 KHz is obviously higher in frequency and new component that did not exist in our source. Now multiple that by countless waveforms that describe your music and you get countless new high frequencies which can make the sound be harsher.

2. Lost resolution. Distortion can be thought of raising your noise floor. The higher it gets, the harder it gets to hear low level detail. So when a note decays into black, it will not sound as well toward the end when jitter is higher.

The above two can be perceived using headphones. This is what I did to learn how to hear such artifacts. The down side is that you don't hear any of the impact on spatial aspects of the sound that Bob Katz talks about. Research I have read from likes of professor Hawksford does talk about how subtle changes brought about by jitter can impact imaging and hence, the threshold of jitter should be lower than we might think when we are considering one channel at a time.
 

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
Ok, now I'm really confused since when did analog just become on/off. I worked on a few analog computers in my early h/w days and they were hardly two states.
Analog becomes just 1 or 0 if you sample it fast enough. Let's say an input voltage goes from 0 to 100 volts in a straight line. The difference then between those samples is 100 and needs a number of bits to represent it (7 bits). Now let's sample 100 times faster. Now the sample after 0 will see a value of "1" volt. The sample after that will be 2 volts and the one after 3. Therefore, the value is only changing at a maximum rate of 1 volt. If the input changed lower, then we could represent it with 0. Assuming that is sufficient resolution for our system (i.e. we are achieving a noise floor below analog system we are sampling), then we have represented than analog signal faithfully with just one bit. The analog signal became a digital signal of only one bit.

I way oversimplified the topic above :). But that is the basic concept and hence the line used to talk about DSD. At lower detail though, we still have a digital system and our noise is not where we want it to be. So a technique called noise-shaping is used in DSD to push the distortion up in frequency. So under a microscope, the stated goal is not achieved that way.
 

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing