A mind boggling question.

N

NuTT98

New member
I'm not sure if this is the right forum, but I couldn't find one more suitable. Anyways, this crossed my mind...

You have a stereo recording, left and right channels. Between the left and right channels, you effectively can position a sound in an infinite amount of locations simply by using pan, correct?

In a single channel, you can pretty much fit an infinite amount of sounds, yes? Say you have drums, cymbals, whatever. You put all these in both channels, but vary the amplitude to pan the sounds across the stereo field.

Now using this theory, you would effectively be able to localize an infinite amount of sounds in an infinite amount of locations in the stereo field. You would be able to take all the recordings that ever existed and place them all in their individual spots in the stereo image. But something doesn't add up because that would essentially mean you would be able to create an infinite amount of data on for example a CD with a finite amount of storage space.

Now that would indicate I'm missing a piece of the equation, but what that is doesn't come into mind, a little help? Thx :)
 
What you're saying makes no sense..... you're equating apples and oranges.

In a 2-track mix, the data storage required for that mix depends on the Sample Rate and the Bit Depth... however many tracks you used to create that stereo mix is completely inconsequential to the data storage required.
 
So that a digital recording of one minute's silence would contain the same amount of data as one minute of the New York Philharmonic at full blast - calculated as sample rate x bit depth x length of recording???

(Is there such a thing as the New York Phil? I made that bit up?)
 
however many tracks you used to create that stereo mix is completely inconsequential to the data storage required.
Click to expand...
That's exactly what my point. So how do you disclassify the potential ability of being able to store or 'express' an infinite amount of data?
 
NuTT98 said:
That's exactly what my point. So how do you disclassify the potential ability of being able to store or 'express' an infinite amount of data?
Click to expand...
Sorry, friend..... I have NO IDEA what you're talking about...
 
Garry Sharp said:
So that a digital recording of one minute's silence would contain the same amount of data as one minute of the New York Philharmonic at full blast - calculated as sample rate x bit depth x length of recording???
Click to expand...
Exactly.....

Much like recording a tape.... if you record 1 minute of silence or 1 minute of a rockin' tune, the same amount of tape is still being used.
 
Because once you mix down the "infinite" # of tracks to a stereo mix, you can't retrieve any of the original tracks in their original form. I think I can see where you were going with this, but the part of the equation you missed is the ability to "unmix" the resulting finite mix into it's infinite # of original tracks.

(I'll have what he's having!)

NuTT98 said:
That's exactly what my point. So how do you disclassify the potential ability of being able to store or 'express' an infinite amount of data?
Click to expand...
 
NuTT98 said:
Now using this theory, you would effectively be able to localize an infinite amount of sounds in an infinite amount of locations in the stereo field. You would be able to take all the recordings that ever existed and place them all in their individual spots in the stereo image.
Click to expand...


Yes.

But something doesn't add up because that would essentially mean you would be able to create an infinite amount of data on for example a CD with a finite amount of storage space.
Click to expand...

No. Because in your pondering above, you are assuming that the storage space is infite. A CD isn't. You can't do this with a CD. You can only do it if you have a storage medium with infinite fidelity. That is, a storage with infinite space.
 
Because once you mix down the "infinite" # of tracks to a stereo mix, you can't retrieve any of the original tracks in their original form. I think I can see where you were going with this, but the part of the equation you missed is the ability to "unmix" the resulting finite mix into it's infinite # of original tracks
Click to expand...
I see what you're saying, but the reason I brought stereo localization into the picture rather than simply mixing all sounds into one channel is to show that retrieving these multiple tracks would theoretically be possible given their distinguishing characteristics. Whether this would play out in real life is irrelevant because the 'infinite' information will be there regardless. In otherwords, if you place these tracks in their own place in a stereo image, they are retrievable in themselves.

You can only do it if you have a storage medium with infinite fidelity. That is, a storage with infinite space.
Click to expand...
I realize that no doubt, but my question is, if the above two points are indeed possible (infinite sounds in multiple places of the image), then what prevents you from effectively placing an infinite amount of data in a finite storage medium? Otherwise if that's not possible, what is it that prevents one from doing it.

A guy in another forum brought up an interesting idea. He (from what I could make out at least) said that the same frequency cannot be positioned in multiple places of the stereo spectrum, obviously this is due to phase overwrite limitations. That to me seems to be a potential solution to my mystery, but it will take a little thinking to clear it up.
 
Nutt98 - they are not really there, all these sounds in infinite number of places. Just the way stereo imaging works fools your ears into thinking they are.

All a digital recorder is doing is taking measurements of sound at (say) 16 points of reference, and doing it (say) 44,100 times a second. If it's done right it can fool your ears into thinking they are hearing an orchestra or whatever, but it's just a fixed amount of data being turned back into signals which power loudspeakers.
 
Something else to ponder -
At least with Digital Audio, you cannot represent EVERY possible sound in a specific interval of digital audio (say a 1 second clip of 16-bit 44.1khz). The possible # of sounds which COULD be stored in that audio clip are quite finite. Since that clip would contain exactly 44,100 samples, with each sample being 16-bits long/wide/big/whatever. There are only a finite # of possible 16-bit words (I need to brush up on my combinatorial mathematics) but I believe it's 2^16 or some big number (but not inifinite)

In undergrad, I did some research on REALLY BIG numbers in the fields of cryptography and Computational Complexity. Now I only deal with how quickly (or slowly) does that Java class take to initialize, so this discussion takes me back to my good old academic days as a Computer Science student!
 
This is incorrect if you're talking about Digital audio. The summing/mixing bus in the audio software has finite precision, so it may appear that there are MANY MANY different positions to pan a signal, this MANY MANY is finite and limited by the resolution of the mixing engine.

You have a stereo recording, left and right channels. Between the left and right channels, you effectively can position a sound in an infinite amount of locations simply by using pan, correct?
Click to expand...
 
I'd like to point out that you cannot in fact mix an "infinite" amount of multiple tracks down to two sterio tracks or even one mono track and not lose a hell of alot. Say you had 65,000 or so 16 bit recordings, and you digitally mixed them to a single 16 bit track. You'd only have about 1 "unit" of amplitude to describe each of those individual recordings (since 16 bit describes amplitude to about 65,000 positions or so). That's not even enough to even be anything more than noise.

Even if you just mix 4 recordings down to 1 track, you still lose a significant amount of the amplitude detail. You have to. The track is quite finite.
 
NuTT98 said:
I realize that no doubt, but my question is, if the above two points are indeed possible (infinite sounds in multiple places of the image)
Click to expand...
Then you have an infinite storage medium.

then what prevents you from effectively placing an infinite amount of data in a finite storage medium?
Click to expand...
The question is still false. You are still assuming that the storage medium is infinite, and then asking how come you can store infinite amounts of data in a finite storage.

Well: You can't.
 
Here's another way of putting it:

The "finite-ness" of an audio track is evident in the headroom you have while mixing. Ever have a mix going, and then you add that one last big fat keyboard solo and your master buss starts clipping even though none of the individual tracks are clipping? That's evidence right there that storage is limited. You couldn't fit that last track in there.

So, what do you do to get that big fat keyboard solo in there? You turn down the volume of its track or the volume of several other tracks. Now it fits. Take this further an imagine you are adding tens of thousands of other tracks. You have to keep bringing the faders down a little bit with each additional one to prevent the master buss from clipping. Pretty soon, every single track is mixed down way below the noise floor, and you don't hear anything at all but really, really loud static. That is how a track is finite. This aplies for analoge as well as digital.
 
Two tracks. That's all it's mixed down to, guys. You could start out with a million tracks . . . but ultimately when you combine them in a stereo recording, they become two.

Now, you don't have infinite tracks anymore. Just two extremely busy and infinitely noisy ones. :D
 
Nutt- Your entire idea would depend on being able to pull out individual tracks from the mix. So far that is impossible to do in any pure sense.

You can store information as a single frequency and combine multiple frequencies at the same time to be able to communicate in parallel which is essentially how analog communications works.

Your idea works only if each discreet piece of data is confined to a specific frequency band. As soon as they overlap it is screwed up. So yes you could mix together 100 different albums and extract them later but you could only use 20-25hz from one and 25-30hz from the next and so on.

Stereo has nothing to with it.
 
Nutt98 - they are not really there, all these sounds in infinite number of places. Just the way stereo imaging works fools your ears into thinking they are.

All a digital recorder is doing is taking measurements of sound at (say) 16 points of reference, and doing it (say) 44,100 times a second. If it's done right it can fool your ears into thinking they are hearing an orchestra or whatever, but it's just a fixed amount of data being turned back into signals which power loudspeakers.
Click to expand...
Well duh :). But again, that's not the point. Whether it's a physically discrete channel is not the issue, obviously it's not, but the sound itself will have it's own location in the audible 3d space.

The possible # of sounds which COULD be stored in that audio clip are quite finite. Since that clip would contain exactly 44,100 samples, with each sample being 16-bits long/wide/big/whatever. There are only a finite # of possible 16-bit words (I need to brush up on my combinatorial mathematics) but I believe it's 2^16 or some big number (but not inifinite)
Click to expand...
When you look at it that way, yes ofcourse. But I already implied the assumption that any digital medium has a finite amount of data storage, doesn't quite answer the question.

No. Because in your pondering above, you are assuming that the storage space is infite. A CD isn't. You can't do this with a CD. You can only do it if you have a storage medium with infinite fidelity. That is, a storage with infinite space.
Click to expand...
Again, missing the point :). I'm not assuming the storage is infinite, I'm assuming the storage is finite but the sound that is represented on this finite storage is not. You don't have to explain that a CD has a finite amount of data on it and that what I'm trying to explain is physically impossible, I KNOW IT'S IMPOSSIBLE, but I want to get down to exactly why that is. I'm not asking what is the limitation of a storage medium, I'm asking what is the limitation of how many individual sounds can be represented in a stereo recording and how specifically is that limited.

Here's another way of putting it:

The "finite-ness" of an audio track is evident in the headroom you have while mixing. Ever have a mix going, and then you add that one last big fat keyboard solo and your master buss starts clipping even though none of the individual tracks are clipping? That's evidence right there that storage is limited. You couldn't fit that last track in there.

So, what do you do to get that big fat keyboard solo in there? You turn down the volume of its track or the volume of several other tracks. Now it fits. Take this further an imagine you are adding tens of thousands of other tracks. You have to keep bringing the faders down a little bit with each additional one to prevent the master buss from clipping. Pretty soon, every single track is mixed down way below the noise floor, and you don't hear anything at all but really, really loud static. That is how a track is finite. This aplies for analoge as well as digital.
Click to expand...
Thx for that, that's some good reasoning. You might have uncovered another piece of the puzzle :)

Your entire idea would depend on being able to pull out individual tracks from the mix. So far that is impossible to do in any pure sense.
Click to expand...
Not so, if they have their own location in a stereo image, you can effectively pull out this sound in itself simply by comparing the difference between the two channels. Kind of like pulling out mono vocals in a stereo recording, they have programs that do that, just not as extensive as the one in question.

Anyhow, thx for the replies guys, but it appears most of you are missing my point :). Seems like the picture is coming together, not crystal clear but I'm getting the idea from here and dudes at audio asylum. A very trippy topic when you think about it.
 
NuTT98 said:
Not so, if they have their own location in a stereo image, you can effectively pull out this sound in itself simply by comparing the difference between the two channels. Kind of like pulling out mono vocals in a stereo recording, they have programs that do that, just not as extensive as the one in question.

Anyhow, thx for the replies guys, but it appears most of you are missing my point :). Seems like the picture is coming together, not crystal clear but I'm getting the idea from here and dudes at audio asylum. A very trippy topic when you think about it.
Click to expand...

You cannot pull out individual tracks from a stereo recording unless they are mono. Even then it pulls out all the mono tracks together. There are no "programs that do that".

We get your point but you are not getting the reality of it. There is no 3D space. There is only two channels of info and the 3D effect is a trick being played on your brain. You should spend more time studying physics books and less time staring at the ceiling.
 
You must log in or register to reply here.
Back
Top