How does the stereo image in digital audio shrink?

[MikeRivers]

 

I listened to my system last night. I put two different signals in two different Pro Tools tracks. I sent each track to a different D/A output. The D/A outputs are connected to a mixing board. I panned one signal to the left speaker and the other to the right. I played both signals simultaneously. I could hear them both out of their assigned speakers. I then muted one channel on my mixing board and could only hear one sound coming from the other speaker. I could not hear any trace of the other sound. I reversed the process and I got the same result.

 

What is this wonderful mixer you have that has no crosstalk? I'll buy one.

 

Actually, I can conduct that same experiment with any of the Mackie mixers I have here and get the same "I could only hear" result. However, if I conduct the experiment scientifically, putting a tone into a channel, and looking for it in an output to which it's not assigned, I can measure its presence. It might be 85 or 90 dB down from the assigned output, but it's there. I can see that you might not be able to hear that in a normal listening environment.

 

You can make a hypothesis and then test it. I tested your hypothesis that I surely must be hearing crosstalk. I don't hear it. I think you're hypothesis is wrong.

 

I didn't directly say that you were hearing crosstalk. What I hypothesized was that you could be hearing the effect of jitter in your stereo mix. I likened it to crosstalk. But unlike crosstalk in an analog system, "crosstalk" resulting from jitter isn't constant, but varies with the jitter spectrum. Hence it's not just a fixed mix of the two channels (however slight), it's a dynamic effect.

 

Now I'm not a psychoacoustician. I don't KNOW that this is the cause of the "shrinking" that some have observed. I do know that I've never heard two different digital systems that I can say sound identical, and this is just one of the little differences that we have to live with. Just like little differences between analog boxes.

 

The problem of not getting exactly the same thing out that you put in never goes away, it just changes. And I say "Who cares?" If you don't like what you get out, change something. If you're happy with it, move on. If you're unhappy simply because you KNOW there's something wrong, then better get another job.

 

If the same signal is being sent to two different speakers and there's some kind of delay, you don't have a shrinking of the stereo image you have a widening of the stereo image.

True, sort of. You can manipulate it to do that. But suppose you send each channel through its own delay (which you do in any digital system) and the amount of the delay between the channels keeps changing. One millisecond the left channel might come out first, the next the right channel might come out first. Can you predict what you'll hear in that case?

 

Actually I think the claim that digital shrinks the stereo image is mistaken. It seems more cogent to reason that digital might have a smaller stereo image in comparison to tape because tape recordings are widening the stereo image with distortions that aren't in the digital recording, both in the recording and in the playback.

 

Analog tape has far more real interchannel crosstalk than a decent digital system. You can predict that the stereo image will be wider with digital than with analog tape and you'd probably be right.

 

I think you're trying to oversimplify what's a very complex problem, but we tend to do that when we don't feel like doing a year's worth of research and then summarizing the result in a 500 word posting on a forum, only to be contradicted.

[Rudolf von Hagenwil]

Artists and producers know, and they wonder what happened to all their hard work, especially in pop music where layering and chorusing are used to create space, including things like artistic interpretations of outer space.

[Ethan Winer]

Perhaps through a phenomenon unique to digital that can

[Han]

Ethan, digital and analog both have their limitations and flaws. What I really hate is those analog/digital debates and it feels like barking to trees. I love the sound of a hi end analog board and a well aligned 2" machine and I love the unlimited possibilities of digital.

The discussion here is about the stereo placement, width and in general the dimension of the analog sound adn I have never had the experience that it stayed 100% the same.

What you say about generaton loss is no subject in the 2008's, a decent studio has an automated board and if the mix is sent to an analog two track, we make a number of analog masters, nothing gets copied anymore these days.

A full blown analog studio still sounds different from a full blown digital studio and what you like better is your own choice. There's no need to bash either analog nor digital. There are more important things in life, right?

[Duddits]

 

I think that is all most of us are trying to say, is that digital is still maturing and improving and has not reached nirvana.

 

That's your position and it's my position. But it's not everyone's position. Let us join forces and convince them! Together, we shall rule the world!!!

[Lee Flier]

I don't mean to sound critical of your point of view, only that I have noticed it is unrelenting. Sometimes people who hold correct views are unrelenting, so I'm not suggesting inflexibility means you're wrong. But in discussions about evolving technology, the person who clings to the past without wiggle room for new information... well, words like "entrenched" and "religion" that have been thrown around might apply.

 

I don't think it's an outrageous mischaracterization to sum that up as "people like Han come off as unrelenting."

 

and I don't put you into that category as well.

 

Again, you have said similar things to me in the past, so I don't think I was out of line in assuming you'd say the same thing when I chimed in.

 

I hear a lot of crappy digital recordings that don't require any A/B to appreciate the depth of their craptasticness. I'm sure you do as well!

 

Sure, and loads of craptastic analog recordings too. But that wasn't the point, as again I've argued many times. The point is one can do a recording that you think is good, and it may well in fact be very good, but that doesn't mean you might not then suddenly hear something else that by comparison sounds much better. And you wouldn't realize it until you did a direct A/B comparison, because even if you've heard the "better" recording a thousand times before, auditory memory is such that you wouldn't remember and would go on happily thinking record B is just as good... until you heard A again.

 

This has happened to me before as I'm sure it's happened to all of us. However since I do hear high end analog and digital compared to each other on a regular basis, and many people don't or have not in a long time, I think I have a good basis for an argument and so does Han.

 

I didn't say that! My response was to Han, about Han, in light of his comment to Ethan, and in light of years of reading his posts. He finds Ethan's argument "entrenched." I agree. I also find his, Han's, argument "entrenched." It is, as I said, "consistent." I find no wiggle room in it, not even enough wiggle room for him to admit that aside from objective preference, he also prefers analog as a matter of taste. As if there's anything wrong with that! There isn't. But by repainting my comment into a polarizing and ridiculous generality, you're making it sound adversarial and indefensible.

 

Well again... because your own comment is "entrenched." I've read it quite a bit before just as you've read Ethan's arguments, and Han's, and mine, a lot before. Every time this argument comes up, you reiterate the point of view that if Han or I or whoever else is advocating analog, it must be because of "nostalgia" or "culture" or some similar reason. You insist that taste is a perfectly legitimate reason to use or not use something (which is certainly true), but then seem to contradict that by calling some people's preferences "religious" and "entrenched" which has a negative, non legitimate connotation. If we are in fact acknowledging that subjective taste is an important factor in one's choice of tools, then let it assume its importance and quit calling it "religious" and "unscientific" in what appears to be a disparaging way.

 

This is, as you like to say, a straw man argument. There is lots of room in my head for legitimate reasons to embrace analog that have nothing to do with voodoo, nostalgia, or entrenchment.

 

Well if that is true, I haven't ever seen you state one. Whereas I've seen the "nostalgia, voodoo, entrenchment" stuff from you a lot.

 

All recording is "a problem" since none fully captures reality. And transfers other than digital to digital are lossy: Live music to analog or digital recording = lossy. Analog recording to analog recording = lossy. Analog recording to digital recording = lossy. Digital recording to analog recording = lossy.

 

Agreed. But folks like Ethan seem to feel that digital can be all things to all people and any opinion to the contrary is the listener's imagination. That's the attitude that Han, Beck, myself and a few others argue against and claim is "entrenched".

 

I didn't say that! And what I said was to Han. Through years of reading his own words, he expresses his taste for analog despite recent protestations that his choice is entirely objective vs. Ethan's "entrenchment" (interesting twist!).

 

Well just because one prefers one sound over another as a matter of taste doesn't mean there aren't also some objective reasons for that. Han made the specific statement that the placement of instruments in the stereo field is more precise with analog in his experience. Taste aside, that's something that should be able to be heard by anyone, no?

 

Han hears a degradation in sound quality when transferring analog to digital in his studio, and he is also nostalgic for analog. Why are those mutually exclusive?

 

First of all he's not "nostalgic" for analog - he still uses it every day and it is still a vital medium for him. And while I agree that his taste for analog is not mutually exclusive to his hearing a degradation in sound quality, as I've just pointed out above, it's hard to deny that if you put those two statements together it's very easy for a reader to infer some causality - i.e. Han hears the degradation because he has a personal taste for analog, and someone without his bias wouldn't hear it. Ethan seems to feel that is the case, and you yourself asked whether Massenburg would hear the same thing in his studio, whether Han's experience with digital was a factor... etc.

 

Whether or not you intended it, I think the implication is sufficient that a lot of readers would infer that, and I think that does the reader a disservice.

 

I believe the way he expressed it was "I am an old fart." So, he loves analog! I don't see that as bad. Do you?

 

No of course not, but again, that doesn't seem to be all there is to your position or Ethan's. The implication when taken in context is "anybody who prefers analog is an old fart who's merely entrenched in their own (past) way of working. Any truly objective person would prefer digital and/or would not hear the same artifacts in digital recordings that Han does." Again, perhaps that's not the intention in your case, but it certainly can easily be inferred from the context of your remarks.

 

That's your position and it's my position. But it's not everyone's position.

 

Who has said anything different? Has anybody said digital will never improve and will never be as good as analog?

[MrKnobs]

The OP was "How does the stereo image in digital audio shrink."

 

So far, most or all of the responses can be put into one of two categories:

 

(1) It doesn't,

 

(2) I'm sure it does but I don't know what the mechanism is.

 

I don't have a dog in this hunt except this one: this is something that can be measured. It's not a matter of opinion, of taste, of who has the better hearing, of double blind testing, of audiophiles, or anything else except measurement.

 

Therefore it's completely useless to argue it, though ignorance certainly does make for some more interesting arguments.

 

Now, before you get mad and hit that Post Quick Reply button, realize that I'm not saying that YOU are ignorant, I'm saying that *I* am ignorant and I really don't like being ignorant about things I have the interest, time, expertise, and gear to resolve.

 

So I submit that those of you who DO hear a reduction in the "soundstage" (what a pretentious term!) please explain to the rest of us exactly what you did to create this perception so that I may attempt to recreate it and measure it in my lab. Keep in mind we don't have a 2" tape machine at the university, at least not one I have access to.

 

My apologies if I missed someone's post that did explain the procedure in detail.

 

Thanks. Let's test this thing and find out what the reality is.

 

Terry D.

[John Sayers]

So I submit that those of you who DO hear a reduction in the "soundstage" (what a pretentious term!) please explain to the rest of us exactly what you did to create this perception so that I may attempt to recreate it and measure it in my lab.

 

 

Terry, the recording I referred to was a stereo hammond organ. I can distinctly remember how the sound would rotate within the soundstage when I recorded it, it also transcribed to the vinyl which is interesting as a record pickup as I recall has around 35db crosstalk yet the record rotated on speakers and in headphones. The digital 16/44.1 doesn't rotate to the same degree yet the digital was remasterd by Don Bartley, one of our legendary mastering engineers, from the original 1/4" Dolby A master.

 

I'd post the digital but unfortunately my site is down so I'll have to wait till it's up again.

[MikeRivers]

The OP was "How does the stereo image in digital audio shrink." I don't have a dog in this hunt except this one: this is something that can be measured .

What units do you measure "shrink" in? You can test the perception if you can convince enough people that your test was valid, but how do you quantify your results?

 

I suppose you could test the hypothesis that the shrinking effect is a a result of jitter by adding jitter to see at what threshold (if any) your validated test subjects perceive a shrinkage of the stereo image. But first you have to train them as to just what a shrunken image is. And who are you going to let choose that example? Someone who thinks digital recording shrinks the stereo image, obviously.

[GVDub]

Here's the experiment that I would propose: Take a recording like this stereo Leslie recording and isolate each side of the stereo image (obviously, if the master tracks on the multi-track are next to each other, there's no way to completely eliminate possible crosstalk from that, but let's try), feed them to independent amps doing due diligence as to level matching and other such esoterica. Rig a system to inject a little of each signal into the other so that you can create and precisely control crosstalk. Do the same thing with the tracks that have been transferred to a digital format.

 

What will this tell us? It will establish if there's some element of crosstalk (perhaps some phase relationship) that maybe spreads a stereo field a little bit. It wouldn't be the first time that it's the imperfections in a system that add what turns out to something desirable. After all, analog recording and playback systems are inherently pretty chaotic in nature, and never really the same twice, unlike the (reasonably) consistent nature of the digital process.

 

Of course, I could be just thinking out my butt from a position of ignorance, but if we proceed by isolating the difference between the two processes and injecting (or removing) from one what's there (or not) in the other, maybe we can come to some conclusions that have practical applications.

[Lee Flier]

I don't have a dog in this hunt except this one: this is something that can be measured . It's not a matter of opinion, of taste, of who has the better hearing, of double blind testing, of audiophiles, or anything else except measurement .

 

Yeah I agree, it can probably be measured... damned if I know how, though.

 

Therefore it's completely useless to argue it, though ignorance certainly does make for some more interesting arguments. Now, before you get mad and hit that Post Quick Reply button , realize that I'm not saying that YOU are ignorant,

 

I have no problem admitting I'm ignorant about how one can quantify stereo field perception and placement. I am. I have no idea how one would do it, so please feel free to explain that part.

 

The only thing I have a problem with is being told I'm not really hearing what I'm hearing!

 

So I submit that those of you who DO hear a reduction in the "soundstage" (what a pretentious term!) please explain to the rest of us exactly what you did to create this perception so that I may attempt to recreate it and measure it in my lab. Keep in mind we don't have a 2" tape machine at the university, at least not one I have access to.

 

Hmm well that's going to make it difficult then, although I've noticed the same phenomenon when mixing in a computer or a digital console vs. a decent analog console. Got a good analog console at work?

[Lee Flier]

 

What will this tell us? It will establish if there's some element of crosstalk (perhaps some phase relationship) that maybe spreads a stereo field a little bit. It wouldn't be the first time that it's the imperfections in a system that add what turns out to something desirable.

 

 

But again, if that's really what's going on then transferring an analog recording to digital would also transfer the "desirable" artifacts and the "shrinking" wouldn't happen. But it still does happen.

[rlindsey0]

What units do you measure "shrink" in? You can test the perception if you can convince enough people that your test was valid, but how do you quantify your results? I suppose you could test the hypothesis that the shrinking effect is a a result of jitter by adding jitter to see at what threshold (if any) your validated test subjects perceive a shrinkage of the stereo image. But first you have to train them as to just what a shrunken image is. And who are you going to let choose that example? Someone who thinks digital recording shrinks the stereo image, obviously.

 

Why not with a reasonably rigorous blinded listening test? If the question to be answered is whether human ears can hear a phenomenon, surely human ears are the most suitable measuring instruments to answer that question--provided that potential confounding variables are ruled out, which is something that a lot of audiophile tests have been quite bad at AFAICS.

 

There's nothing wrong with using ears as measuring instruments if they're appropriate to the task and the usual controls are applied.

 

If someone can reliably and reproducibly hear soundstage differences, to a degree significantly greater than can be attributed to chance, in a properly controlled test, that ought to suggest that there is something there. If not, then not. I don't know if this has ever been tested with any rigor, but it would be interesting whatever the results might turn out to be.

 

Personally, I dunno whether this shrinking happens or not. I don't recall ever noting it, but I've never tried to test for it either. The fact that many people believe it does, in itself, proves nothing. Of course, it doesn't prove those folks are necessarily wrong either. I'd love to see this potential phenomenon really tested.

[GVDub]

But again, if that's really what's going on then transferring an analog recording to digital would also transfer the "desirable" artifacts and the "shrinking" wouldn't happen. But it still does happen.

 

 

I'd say that depends on how you transfer it. Just as an example, I know some guys who transfer multi-track from analog to digital a single track at a time rather than all 24 (or 48 or more) at the same time, for bandwidth considerations.

[rlindsey0]

The only thing I have a problem with is being told I'm not really hearing what I'm hearing!

 

 

But surely you can admit that that's at least a theoretical possibility? We're human beings; we're quite capable of seeing things that aren't there, hearing things that aren't real, and believing things we can't possibly know. It would seem to make sense to at least leave open the possibility that just maybe you aren't really hearing something you think you are. I'm not saying you're wrong even; but to rule out the possibility that you could be would seem unwise.

[rlindsey0]

 

I'd say that depends on how you transfer it. Just as an example, I know some guys who transfer multi-track from analog to digital a single track at a time rather than all 24 (or 48 or more) at the same time, for bandwidth considerations.

 

 

But do you have any data to suggest that that would make any difference at all? Why should it matter, and why should we believe it would?

[Lee Flier]

 

But surely you can admit that that's at least a theoretical possibility? We're human beings; we're quite capable of seeing things that aren't there, hearing things that aren't real, and believing things we can't possibly know. It would seem to make sense to at least leave open the possibility that just maybe you aren't really hearing something you think you are.

 

 

As I think I've mentioned elsewhere in this thread, the "placebo effect" does indeed happen to everyone when it comes to sound, and any audio engineer is well aware of this phenomenon. We all know that we hear things differently from one day or hour to the next and from one listening environment to the next, and that our perception of how things sound can be influenced by any number of purely psychological factors. This is something that gets hammered into any engineer as an intern, and demonstrated over and over.

 

Therefore we tend to be very careful about forming general opinions. I would never make a blanket statement such as "digital shrinks the stereo field" without having heard the phenomenon MANY times, in many different circumstances, on many people's work, systems, etc. Including several times being able to consistently identify the same mix on analog vs. digital in blind tests.

 

That's why I find it somewhat patronizing to be told I'm not really hearing what I'm hearing. I wouldn't say that I hear anything if I had even the slightest suspicion that I was hearing placebo effect, because I certainly don't think I'm immune from that.

 

I'm quite open to the possibility that there are circumstances under which my "blanket" statement isn't true. But if so I've heard it rarely if it all. If there is a way to mitigate the shrinking stereo field, I haven't found it and I don't think I've heard anyone else who has either, including bigtime Grammy winning engineers.

[John Sayers]

It appears to me that our ears are smarter than our technology. If our ears can hear something our technology can't explain are our ears wrong? I think not.

As I said earlier when I, and another every experienced engineer, compared the vinyl of Dire Straits with the CD side by side in realtime in a master control room with top of the line gear they sounded different. The difference was in the reverb tails and the determination of the positioning of instruments within the sound stage.

One day someone will probably discover why it occurs, I await that outcome.

[GVDub]

But do you have any data to suggest that that would make any difference at all? Why should it matter, and why should we believe it would?

 

 

I have no data pointing either way, so, I suggest a way of collecting some data points. Without properly established experimental methodology, we're all just shooting off our mouths. Everybody can sit here and cite subjective experiential instances for the next 10 years and nobody's going to change anybody else's minds. I want to see some empirical evidence, and if we find that there's stuff that proper testing shows is happening that we haven't figured out how to measure, we need to suss that out, too.

 

I don't really have a dog in this fight. As I've stated elsewhere, the single most important thing to me in any music is how it feels

[Rudolf von Hagenwil]

I wouldn't say that I hear anything if I had even the slightest suspicion that I was hearing placebo effect, because I certainly don't think I'm immune from that.

 

 

???

 

So, do you hear the placebo effect or do you hear I not?

[Anderton]

If there is a way to mitigate the shrinking stereo field, I haven't found it.

 

 

Lee, I've always respected your opinion, but I just can't believe you don't know how to avoid stereo shrinkage: Wash it in cold water, not hot, and use Woolite instead of detergent. And all this time I thought you were a professional...clearly, there are some SERIOUS gaps in your knowledge. Harumph.

[the stranger]

I admit I have not used a professional analog recorder in many years, so I don't know how good the newer tape formulations are. (Can you even still buy 2" analog tape?) But I'm sure they're still inferior to digital in every way. If you believe otherwise, tell me in what way newer formulations of analog tape are better than modern digital. Again, please be very specific.

 

 

How about tape just sounds better? Inferior in every way? Bullocks.

 

The only thing that is inferior is believing one is better than the other. Analog and digital both have their pros and cons. Anything else is just subjective opinion.

[MikeRivers]

 

Why not with a reasonably rigorous blinded listening test? If the question to be answered is whether human ears can hear a phenomenon, surely human ears are the most suitable measuring instruments to answer that question--provided that potential confounding variables are ruled out, which is something that a lot of audiophile tests have been quite bad at AFAICS.

 

This is exactly the problem. Different humans hear things differently, and the same human hears things differently every day. If you tested one person five times a day for a month with a double blind test of a very subtle audio effect, I suspect that only a few would do any better than a guess. Anyone who wants to believe in the effect will be convinced by such a test, and anyone who doesn't want to believe can always poke a hole in the test based on the fact that the measuring instrument was neither calibrated against a standard nor stable.

 

If someone can reliably and reproducibly hear soundstage differences, to a degree significantly greater than can be attributed to chance, in a properly controlled test, that ought to suggest that there is something there.

 

According to some people here, all you have to do is make a digital copy. But not everyone will hear what they say they're hearing.

 

Personally, I dunno whether this shrinking happens or not. I don't recall ever noting it, but I've never tried to test for it either. The fact that many people believe it does, in itself, proves nothing.

 

How would it affect you if it was conclusively proven that the "shrinking" effect does inded occur? Would you seek some other means of recording and reproduction? Would you throw away your CD player and just go to live concerts? Would you make some compensation in your mixing to attempt to undo something that you can't hear? Would you go out and buy the latest plug-in that does that for you?

[rlindsey0]

 

How would it affect you if it was conclusively proven that the "shrinking" effect does inded occur? Would you seek some other means of recording and reproduction? Would you throw away your CD player and just go to live concerts? Would you make some compensation in your mixing to attempt to undo something that you can't hear? Would you go out and buy the latest plug-in that does that for you?

 

 

I personally wouldn't do any of those things. I'm not yet convinced that the phenomenon is real, and even if it were, it wouldn't matter enough to me to outweigh all the other reasons I record digitally. I hear enough imaging in digital recordings to satisfy me. But then, I'm not a pro. If I were, it might matter more. I guess my interest in this relates more to intellectual curiosity than to practical need. I like to know stuff, insofar as I'm able, and I think that in general, believing stuff that's not true is unlikely to be useful (though it isn't always particularly harmful either). I just would enjoy knowing what the actual scoop is on this. If the phenomenon's real, that would be nice to know. If not, that would be equally nice to know.

[John Sayers]

Ok - my site is back up - well done Boden slackmaster!!

here's the wav file of the organ intro.

http://johnlsayers.com/mp3/organ.wav 10 meg file. I posted the wav in case people think it's an mp3 problem.

BTW - the distortion is intentional - it's the effect you get when you remove one of the output bottles on the leslie.

here it is in context.

http://johnlsayers.com/mp3/organ_1.mp3

you may have trouble streaming it as the site isn't up to full steam yet.