Speaker Cables

[jj]

I wonder if Mr. Randi will see fit to comment on the prices and advertisements for some of the speaker cables on the market.

I work in the field, to say the least, and I use 12-2 AWG for anything but the most extraordinary speaker runs, and various low-inductance (Belden, mostly) equivalents for long runs where inductance or coupling to metal conduit might be an issue.

I don't recall ever having paid anything over one dollar per foot, either, even for the most expensive, complicated stuff that had a really annoying (200') length, inside of metal conduit, closely squeezed in with 9 other speaker cables, to an extent a come-along, lots of cable lubricant, and about 300 lbs of tension were required to get it all in there.

It was a case where the crosstalk was interesting. With no speaker connected, crosstalk was down a large number of dB (meaning I could see the crosstalk at high frequencies, with a scope, differential input, and care). With a speaker connected (i.e. with current in the circuit) I could see crosstalk down 60dB. While this might seem like a problem, the presence of the sound source from the speaker 60dB higher than the unloaded crosstalk was not a real big issue, i.e. even if the second signal could be loaded at full voltage into a loudspeaker the term "inaudible" enters immediately.

Attempting to measure crosstalk with two speakers connected was difficult, the reverse-transduction of the second speaker from the first speaker's output, working through the .1 or .2 ohms worth of cable, was the predominant thing visible.

I didn't have any large non-inductive resistors handy, unfortunately, at the time.

The point, however, is that all things considered, any artifacts resulting from that long run inside a bad environment were down (I made many more measurements than the ones I just mentioned) to the point where I couldn't discern between any transmission-related problems and the amplifier's distortion characteristics, which was down over 100dB with levels running at about 95dB SPL in the room (i.e. d**m loud). Yes, I measured things like dispersion in the cable, or tried to, the total dispersion under 200kHz wasn't worth noticing. I could easily see the length of the cable (it was situated where I could also use a short cable on the low-power side to compare to), and see that it was running at about .5c propagation velocity, but the dispersion and delay due to inductance was quite minimal.

Instruments, in this case and application, have quite a bit more sensitivity than the ear, something I can say in this case because the relationship between electrical level and sound intensity (i.e. SPL) was known to under .25 dB, and the artifacts were, and I mean this quite literally, quite below the noise created at the ear drum by the fact that molecules are discrete objects.

Which brings me to the word "microdetail" seen in some audio commentary. Ok, I need the "steam coming out of my ears red faced frown smiley" here, or perhaps a donkey looking back at its own behind with the caption 'mine!'...

Here's a fact, one that can and has been measured. For an average-sized eardrum, the atmosphere itself, by its very presence, creates a noise level circa 6dB SPL, 20Hz to 20kHz white noise, due to the fact that "pressure" consists of molecules bouncing off the eardrum. This level, taking in critical bandwidths (the ear has a mechanically implemented filter, the exact workings of which are under hot discussion, and a "critical band" or "ERB" is a measure of bandwidth) comes out to a level that just MIGHT, under perfect conditions, BARELY be perceptable at the point in frequency (ear canal resonance) to a person with unimpaired hearing.

The confirmation, however, is admittedly hard, to prove, directly, that this was the case, and that it wasn't internal noise in head (the internals of your head are at some level noisy, blood flow, etc), since in order to prove it, you'd have to remove the air on one side of the eardrum. This is not a test one can ethically consider for obvious reasons, and that wouldn't do anything but serious injure the subject even if somebody was imbicilic enough to try.

Ok, that's a limit to hearing. The narrowest bandwidth in the hearing apparatus is on the order of 80 to 100 Hz, dropping the SPL from the noise by about 24dB from that +6dB around the filters that detect the lowest bandwidth. That puts us at -18dB SPL. The threshold of hearing at that frequency (50Hz) is on the order of +30dB SPL.

'nuff said? I think so. Needless to say, I can find you lots of people who don't agree and lurch on with fantasy after fantasy about "microdetail" and how "digical cuts off at the noise floor" (a fallacy both involving noise and misunderstanding of how sampled, quantized systems are implimented), and more and more and more.

Hissss --- That's the remaining steam leaking out of my ears.

Sorry for the rant, maybe I just ran into too much Bachelor of Science today or something.

[Jyera]

JJ, I'm skeptical whether is it worth it to pay a premium for speaker cable. But I thought also that it ought to make some difference. Otherwise why is people will to pay extra? Otherwise why these speaker cable maker can sustain their sales to justify their business?

I recall moving into a new home and had the chance to do some renovation.

Specifically I had 3 needs:
1. I needed the best cable for my system.
2. I needed to know what is the best way to lay my cables.
3. Needed to put up cornics to hide the cables to ensure it is tidy

On best cable for my system:
- I bought the BOSE sound system. It was considered very expensive investment to me then. So I was trying to get things right, to ensure the sound quality isn't compromised due to poor cable decision, and defeats the purpose of buying a good sound system.

On Cable laying:
- Should I avoid laying power cables along side, signal cable and speaker cable? Will that defeats the purpose of buying good cables?
The vendors provided no answer. I asked the person who sells the the sound system to me what is the best cable and they just their recommended their standard cable. Which to me seems just simple copper cable with white insulation.

On Cable Tidiness:
- I had only one chance to hide my cable in the cornics

JJ you seemed knowledge. Any comment ?

[epepke]

Originally posted by Jyera
JJ, I'm skeptical whether is it worth it to pay a premium for speaker cable. But I thought also that it ought to make some difference. Otherwise why is people will to pay extra?

Um, because they're stupid?

This is the same crew that had a number of ardent believers in the idea that marks from a green pen would somehow magically reflect the laser on a CD player and made it sound better.

[jj]

Originally posted by Jyera
Otherwise why these speaker cable maker can sustain their sales to justify their business?



http://www.skepticalcommunity.com/phpbb2/viewtopic.php?t=2117


On Cable laying:
- Should I avoid laying power cables along side, signal cable and speaker cable?

You shouldn't mix power cables and signal-level cables, really. Speaker cables might pick up some hum from the power cable, so I'd avoid that, too.

Unless your low-level cables (signal) are not particularly well shielded, or your speaker cables don't have some twist to the conductors, you shouldn't have TOO much trouble with low-level cables and speaker cables, AS LONG AS YOU DO NOT BUNDLE THEM TOGETHER.

Best is apart, really, but the practice necessary depends on the specifics of your house, the equipment, etc. There is no "one answer".

[Just thinking]

Originally posted by epepke
Um, because they're stupid?

This is the same crew that had a number of ardent believers in the idea that marks from a green pen would somehow magically reflect the laser on a CD player and made it sound better.

Please don't think that just because someone enjoys the benefits of some areas of high-end audio, they must jump in blindly to all the claims that come out of that establishment. I never bought a green pen, Bedini Clarifier, cable lifters or any other number of items that peg the BS meter. Besides, one of my pet peves in the audio community for reproducing music that goes without comment is the introduction of Surround Sound 5.1. Give me two good full range speakers for the same $$$$ of some exotic mutichannel system and I'll blow your doors off.

[epepke]

Originally posted by Just thinking
Please don't think that just because someone enjoys the benefits of some areas of high-end audio, they must jump in blindly to all the claims that come out of that establishment. I never bought a green pen, Bedini Clarifier, cable lifters or any other number of items that peg the BS meter. Besides, one of my pet peves in the audio community for reproducing music that goes without comment is the introduction of Surround Sound 5.1. Give me two good full range speakers for the same $$$$ of some exotic mutichannel system and I'll blow your doors off.

Um, so why exactly are you responding?

[Just thinking]

Originally posted by epepke
Um, so why exactly are you responding?

It just seemed that you were lumping all audiophiles into one big group -- If I'm wrong, I'm sorry.

[epepke]

Originally posted by Just thinking
It just seemed that you were lumping all audiophiles into one big group -- If I'm wrong, I'm sorry.

No; I'm saying that there exists a class of people for whom this is a market, and I think they're stupid.

Personally, I built my own speaker cabinets from one-inch particle board, and they're damn near flat from 20 Hz to 20 KHz. But still, I use 12-gague zip cord to drive them.

[jj]

Originally posted by Just thinking
Give me two good full range speakers for the same $$$$ of some exotic mutichannel system and I'll blow your doors off.

Now there I have to disagree. Give you a straightforward 5 channel system, no exotica, and simple, documented recording techniques, and you'll have better realism than you can possibly get with your 2 channels, and over a much larger listening area.

Fletcher and Snow, 1929, were the first to point the way. Their results remain unimpeached.

Please don't ask me what I think about the present state of multichannel production, they don't allow that kind of language around here. You'll have to tune into Penn and Teller, when Penn is in a real bad mood, to get the flavour.

Some simple things to think about. 1) if you use "panpots" like they use for 2channel systems, you convert the "line" where hte good listening spot is into a "point". This is the opposite of "good".
2) Given 1, and the fact that nobody needs mono compatability any more, what obvious techniques bite you on the rump, now?

[epepke]

Originally posted by jj
Now there I have to disagree. Give you a straightforward 5 channel system, no exotica, and simple, documented recording techniques, and you'll have better realism than you can possibly get with your 2 channels, and over a much larger listening area.

Do you know if any of the old "quadrophonic" recordings have been remasterd for 5 channels? I still have a vinyl copy of I Think We're All Bozos on this Bus in quad. It would probably be really good in 5 channels.

[jj]

Originally posted by epepke
Do you know if any of the old "quadrophonic" recordings have been remasterd for 5 channels? I still have a vinyl copy of I Think We're All Bozos on this Bus in quad. It would probably be really good in 5 channels.

Well, quadraphonic, with the 4 speakers in the corners, is perhaps the most pessimal way that one can remain with a bilaterally symmetric arrangement of speakers in the plane.

In other words, it's the wrong approach. The center front loudspeaker is key to everything involved in good multichannel.

Having said that, I'm sure it would come off well in 5-channel, you'd have to remix from scratch with tools nobody uses at present to make it fly, though.

[Just thinking]

Originally posted by epepke
Personally, I built my own speaker cabinets from one-inch particle board, and they're damn near flat from 20 Hz to 20 KHz. But still, I use 12-gague zip cord to drive them.

Frequency response is not sufficient to determine the realism of a reproduced acoustic event. In fact, it is one of the most abused measuring scales used to exploit bad speakers. Did you check for phase alignment between the drivers in each cabinet? How did you space the elements to get phase-time alignment with your crossover between the drivers? Did you use a Butterworth style filter? And over what power range did the speaker test? What Q-factor did you design around? How did you determine the best crossover design for the drivers used? How did they test with a square wave input? Do they have any hot spots when walking slowly past them at a constant distance?

Trust me, there are a plethora of other design factors, such that wire can become almost an insignificant element if not designed correctly.

PS: I am not saying that your speaker is bad, just that a red flag goes up with me as soon as someone speaks of 20-20k as the be all to end all test.

[Just thinking]

Originally posted by jj
Now there I have to disagree. Give you a straightforward 5 channel system, no exotica, and simple, documented recording techniques, and you'll have better realism than you can possibly get with your 2 channels, and over a much larger listening area.

I have worked in several audio stores (two being high-end), gone to shows, and listened to quite a few home theatres. Maybe it's just me, but I've always found good 2-channel as being the most realistic. As Cary Audio would say, that's what gave me Goosebumps.

[epepke]

Originally posted by jj
Having said that, I'm sure it would come off well in 5-channel, you'd have to remix from scratch with tools nobody uses at present to make it fly, though.

Tomita mixed one of his albums in 5 channels but then reduced to 4 for a quadrophonic release.

I'm sure that Bozos was done in at least 16 channel and then mixed down to 4. There are a lot of spatialization effects that they did which come across pretty well even in stereo.

Also, in these days of iPods, why aren't there more binaural recordings?

[jj]

Originally posted by Just thinking
I have worked in several audio stores (two being high-end), gone to shows, and listened to quite a few home theatres. Maybe it's just me, but I've always found good 2-channel as being the most realistic. As Cary Audio would say, that's what gave me Goosebumps.

Well, I can't argue with your preference. I can, however, point to Fletcher and Snow, who show rather definitively that you can't get good distance localization for center-stage with 2 channel stereo. They didn't know why at the time, but over time we've come to understand that two-speaker systems create interference at the ear that seriously impacts distance cues for central signals.

I suspect that you may not have been subjected to good 5-channel production. I've evaluated quite a few, and due to some very mistaken "rules of thumb" most production methods misuse the center channel.

In fact, some producers are now agressively arguing to GET RID OF THE CENTER CHANNEL after misusing it and annoying their listeners. Great business to be in, eh? The data has been around since 1929, and is annoyingly convincing.

[Just thinking]

Originally posted by jj
Well, I can't argue with your preference. I can, however, point to Fletcher and Snow, who show rather definitively that you can't get good distance localization for center-stage with 2 channel stereo. They didn't know why at the time, but over time we've come to understand that two-speaker systems create interference at the ear that seriously impacts distance cues for central signals.The data has been around since 1929, and is annoyingly convincing.

So has data from those that created 2-channel reproduction. Check out the work done by Alan Dower Blumlein from over 70 years ago.

Here are the problems I see. Whenever you 'mix down' multichannel masters into 2 or whatever number of channels, you are relying on the engineer's preferrences as to how the event should sound. Some recent recording companies have in fact gone back to recording using 2 microphones to eliminate this effect. Also, if you think placing 2 speakers just right for good stereo can be daunting, try 5, or 7, or more. You need more signal processing for the rear channels to get time alignment -- more processing with the signal, more 'noise'. How do you know when you've got it just right? The 'sweet spot' for 5 channels is also much more restricted than for 2 channel -- yes, both can be small. And as for distance localization (I guess you mean front-to-back imaging), that can be accomplished with only 1 speaker, so I'm not so sure I know what you're aftere here.

All the above aside, most aural events occur in front of the listener. Two channel stereo tries to recreate those sounds in front of the listener in their own home. The room acoustics will enhance (or subtract) from that source in its own way, but that does not mean that it is up to the stereo to place you at the original event. It is the job of the stereo to place the event in your home. Trying to do the former is, for all practical reasons, impossible -- there was no set of omni-directional microphones placed in some sweet spot in the audience to capture the sounds coming in from 5 (or 7) directions. But you can place 2 microphones in front of the event which represent the two ears most of use to listen with. No, it's not perfect, but it can be damn good.

Anyway, that's my 2 cents worth.

Back to the music !!

[jj]

Originally posted by Just thinking
So has data from those that created 2-channel reproduction. Check out the work done by Alan Dower Blumlein from over 70 years ago.


You're not listening. I'm well aware of Blumlein's work, and of Gerzon's followup to it, and they both have the simple problem that they eventually radiate out of two speakers, which is where the central depth cues are unavoidably lost. Tom Holman has spoken at some AES meetings about it, but if you have access to any set of HRTF's, just figure out what happens due to the +30 and -30 rays to one ear, when they are combined, and compare that frequency range to the frequency range where floor reflections provide depth cues. Oops.

Blumlein and Gerzon both have the same problem, by the way, they are both co-incident miking systems. This makes for a hard-center listening position that is very, very sensitive, and for a very small, narrow sweet spot for listening. You can see how this arises simply by looking at how the ear responds to second wavefronts vs. first wavefronts. (Haas effect)


Here are the problems I see. Whenever you 'mix down' multichannel masters into 2 or whatever number of channels, you are relying on the engineer's preferrences as to how the event should sound.


Well, perhaps rather than "mix down" for a real classical recording, one might consider simply capturing the event as it is in the hall. There are many ways to do this. Ambisonics (Gerzon's work) does a pretty good job of this, but restricts you to one listening position. There are a variety of other methods, such as Griesinger's work, which I've seen but don't have cites handy for, or "Perceptual Soundfield Reconstruction", Preprint Number: 5202 Convention: 109 (August 2000)
Authors: Johnston, James D.; Lam, Yin Hay (Vicky)
E-lib Location: (CD aes15) /pp9900/pp0009/109188.pdf


For a system I can even point you to a few reviews of.

Also, if you think placing 2 speakers just right for good stereo can be daunting, try 5, or 7, or more.


Well, yes, you do happen to be directly splat center in my area of expertise.


You need more signal processing for the rear channels to get time alignment -- more processing with the signal, more 'noise'.


NO you explicitly, specifically, and unquestionably DO NOT, if you capture information your ears can make sense of!


How do you know when you've got it just right? The 'sweet spot' for 5 channels is also much more restricted than for 2 channel -- yes, both can be small.


Please, please, don't EVER repeat this fallacy ever again. It's wrong, simply wrong. If you use a co-incident technique, then it's true, but that's the wrong thing to do, and that is easily shown. Take some mikes and try it.


And as for distance localization (I guess you mean front-to-back imaging), that can be accomplished with only 1 speaker, so I'm not so sure I know what you're aftere here.


In fact, 1 speaker provides better distance localization than TWO do, for a central image directly between the two. This is not a new result. Before you start arguing about this any more, please go back to Chapter 13 of "Speech and Hearing in Communication" by Harvey Fletcher, Edited by Jont B. Allen. It's available in reprint from the ASA. You'll find, I think, all the proof you need right there to make it clear to you that for distance, ONE speaker is better than two. Interestingly enough, 3 (including center) is as good as one, the physics are very clear if you figure out what kind of interference you get at the ear.

What's more, you do not accomplish that by putting signal ONLY into the center if the sound source is in the center, you put signal from any one source into at least 3 channels, usually 4, all the time, and not with a whole lot of attenuation, either.

Read the paper cited above for ONE way to do it.


Two channel stereo tries to recreate those sounds in front of the listener in their own home.


And, because it takes no control over the complexity of the soundfield (it can't, not enough sources) it sounds like what's behind you is your home, most of the time. A good Ambisonic or Soundfield recording can do otherwise, sometime, for a very good venue and listening room.


Trying to do the former is, for all practical reasons, impossible -- there was no set of omni-directional microphones placed in some sweet spot in the audience to capture the sounds coming in from 5 (or 7) directions.


Your ears aren't omnidirectional. Why ever would you use omni's in such a miking setup?



But you can place 2 microphones in front of the event which represent the two ears most of use to listen with. No, it's not perfect, but it can be damn good.


Of course if you do it wrong, it won't work. Why is this a germane point?


Anyway, that's my 2 cents worth.

Back to the music !!

And now, let me dig up some reviews of the method I cited above. They are a bit "high-end"ish, indeed, but so it goes. They are the people who wanted to come listen.

http://www.onhifi.com/features/20010615.htm

http://www.stereophile.com/asweseeit/313/index1.html

[jimlintott]

JJ

I just read the first review and was wondering, can this PSR technique be delivered easily via Dolby 5.1 or DTS? If you are using seven microphones will 7.1 sound be the way to go? Would that make PSR even more effective?

It sounds pretty cool and as someone with a very good 5.1 speaker system I am very interested. (You'll have to pry my center channel from my cold dead hands ;) )

Thanks

[Just thinking]

Originally posted by jj
I'm well aware of Blumlein's work, and of Gerzon's followup to it, and they both have the simple problem that they eventually radiate out of two speakers, which is where the central depth cues are unavoidably lost.

Well, if that's true, then I guess I'm up for the $1,000,000 prize, as I can certainly hear a distinct center image from my 2 speakers and tell if the source was upfront or wayback from the mikes. I can even make out some front to back distance from a mono recording, using 2 stereo speakes, of course.

There are many ways to do this. Ambisonics (Gerzon's work) does a pretty good job of this, but restricts you to one listening position.

When I listen to music, I'm only in 1 position. If I invite a friend over, I place him/her in that position. Multiple freinds, I put them as close as possible. I've never heard any complaints, and they are usually quite impressed as to how good a system can sound.

NO you explicitly, specifically, and unquestionably DO NOT, if you capture information your ears can make sense of!

How is it then possible to avoid using a time delay? No one can record the rear information with the advanced knowledge of knowing how far back rear speakers will be placed in the person's home. Plus that distance will vary from playback system to playback system.

A good Ambisonic or Soundfield recording can do otherwise, sometime, for a very good venue and listening room.

This is my point -- for a given amount of $$$$ you will get better sound from a 2 channel system than multichannel. Not only must you buy additional speakers, amps, etc., you must now pay even more attention to the room than ever before. Imagine using all that extra cash on a much better pair of speakers -- clearly a crucial element in the playback chain.

If you use a co-incident technique, then it's true, but that's the wrong thing to do, and that is easily shown.

Then all the recordings I've heard must have done it wrong. As soon as I move about, the balance shifts quite a bit. True, the center speaker keeps dialog and main events focused up front and center, but it becomes too easy to get distracted from the now closer other channel speakers.

Your ears aren't omnidirectional. Why ever would you use omni's in such a miking setup?

In order to capture the sound fields as heard from a given listening position.

[nbenami]

This is my point -- for a given amount of $$$$ you will get better sound from a 2 channel system than multichannel. Not only must you buy additional speakers, amps, etc., you must now pay even more attention to the room than ever before. Imagine using all that extra cash on a much better pair of speakers -- clearly a crucial element in the playback chain.

Perhaps that was true a few years ago, but given modern, cutting edge multichannel systems, it is no longer the case. A properly set up, five to six figure multichannel system, playing the appropriate source material (admitedly hard to find though it may be) is unerguably going to be able to do a better job than a simple 2-channel system.

Remember, modern high-end multichannel systems will include setup microphones and sufficient, and sufficiently high-quality digital processing to account for the acoustical characterisics of the room (which would, obviously, be thoroughly treated to remove first-order reflections.)

An example system would probably include a trio of Wilson Audio Maxx2 speakers for the front and center and a pair of Watt 7s for the rears. Theta or Mark Levinson digital gear would do the processing. The system would be professionally set up.

Is this a system most of us can buy? No, but for that given $$$, and given a high quality multichannel input, a 2-channel system would simply not be able to compete.

[jj]

Originally posted by Just thinking

Well, if that's true, then I guess I'm up for the $1,000,000 prize, as I can certainly hear a distinct center image from my 2 speakers and tell if the source was upfront or wayback from the mikes. I can even make out some front to back distance from a mono recording, using 2 stereo speakes, of course.


Exactly, you can make out "some" front to back distance, and you can tell dry from reverberant. What you can't do is resolve distances well. That's basically what you just admitted, in a very indirect and skewed fashion.

.

When I listen to music, I'm only in 1 position. If I invite a friend over, I place him/her in that position. Multiple freinds, I put them as close as possible. I've never heard any complaints, and they are usually quite impressed as to how good a system can sound.


I won't address the closeness issue, I'll simply point out that you are not only in '1 position'. Head movement is an instinctive (or learned early) part of sampling a soundfield. Learning to listen with the head still (as many people learn to do) is not natural. Go to a concert, you'll notice people's heads moving about a bit (it only takes a bit). This sampling of the soundfield helps people resolve differences between source and hall, characteristics of the hall, distance cues, etc.

When you can't do that, you restrict yourself to a very limited part of the acoustics present in the room. This is what you must do with a stereo recording. It's also the problem with a binaural recording, which is otherwise a good thing. When you move your head, the image moves, too, with headphones and a binaural recording. If you try to use interaural cancellation, you have to sit facing directly to the front, and then we're back to square 1.

If you have a proper 5-channel recording, you can move your head, and get a much more realistic result, containing a substantial number of the cues in the hall.


How is it then possible to avoid using a time delay? No one can record the rear information with the advanced knowledge of knowing how far back rear speakers will be placed in the person's home. Plus that distance will vary from playback system to playback system.


You need no time delay, nor anything else, if your recording captures the cues that the auditory system can resolve, as opposed to the cues (quite different ones) that a least-mean-squares system can resolve. All you need to know is that the speakers are all identical and the listener is more or less in the middle.

Please read that technical paper I pointed you at. It specifically addresses the results of a test involving the measurement of "sweet spot".

The problem with the test is that it is, of course, impossible to make the test a blind test. The 5 channel result is so incredibly, entirely superior that it's obvious which is which, and the comparison as a result is very difficult. In running the test, one subject refused to grant "stereo" any realism at all, even though the stereo results were very good, after the comparison with 5-channel.

This is my point -- for a given amount of $$$$ you will get better sound from a 2 channel system than multichannel.


I'm sorry, but you're still relying on proof by assertion. Did you read those reviews I pointed you to?


Not only must you buy additional speakers, amps, etc., you must now pay even more attention to the room than ever before.


False, with more sources, it's much easier to swamp out the characteristics of the playback room. You need be less careful with the room, not more careful.

Consider, when the reveration from behind you comes from the original room, you're not nearly as dependent on the room you're sitting in. LEDE and such designs are not necessary. Of course, if a room is truly bad, nothing is going to sound good. One of the first rules of practical accoustics is "if you don't want the energy there, don't put it in in the first place, because it's harder to get it back out".


Imagine using all that extra cash on a much better pair of speakers -- clearly a crucial element in the playback chain.


I can easily imagine what happens. As the reviews and the paper show (did you bother to read the paper or the reviews?) you don't have to have the 'best' speakers.

One thing that fancy, high-end speakers (as well as mundane, low-end speakers) do is add a particular choice (particular to the speaker in the room) of direct/reverberant sound in any given room. (and that varies by room) When the combination of room and speaker matches your preference of direct/reverberant sound, then you're happy. This is why one speaker does not fit all. It's why a speaker you like in one room you may dislike in another. Of course, there are more variables than this that are addressed, but we'll stop there, for now. Speakers can also compensate, exagerate, or otherwise modify other attributes of room, equipment, etc.

And amplifiers are a dime a dozen now, sorry.


Then all the recordings I've heard must have done it wrong.


That is, unfortunately, entirely possible. The desire for "mono compatibility" and good multichannel production are just about diametrically opposed. Present 2-channel techniques that are being extended to multichannel are nearly all effectively coincident techniques.

If you can mono the recording and it doesn't sound like you're swimming in water, then it is probably a co-incident kind of mix of some sort.


In order to capture the sound fields as heard from a given listening position.

Bzzt. Sorry, At any one point, in order to capture the air movement, you need at least 4 channels, x, y, z, and w. Look up "ambisonics" on the net for a very good dissertation on this issue.

A set of omnis don't get you anything like what you imagine that they do.

A widely spaced set of omnis would be something Greisinger would like, I suspect, although I would never try to speak for David, and in my opinion, you'd find that while the imaging was a bit vague, you'd not have nearly as much trouble with image shift when you move off-center.

Before you continue in the same vein you're headed in now, I would suggest that you dig up the references I've cited, read them, and also perhaps read the reviews of recordings that I've pointed you to, journalistic though they are. The fact that you propose omni's in and of itself suggests that you aren't considering some very important factors about room acoustics.

Oh, for more information also check out
http://www.ece.rochester.edu/~gsharma/SPS_Rochester/presentations/audio2003.pdf

Thanks to Dr. Sharma for putting up my slides, btw.

[jj]

Originally posted by jimlintott
JJ

I just read the first review and was wondering, can this PSR technique be delivered easily via Dolby 5.1 or DTS? If you are using seven microphones will 7.1 sound be the way to go? Would that make PSR even more effective?

It sounds pretty cool and as someone with a very good 5.1 speaker system I am very interested. (You'll have to pry my center channel from my cold dead hands ;) )

Thanks

It mixes down to standard 5.0 mix. So, it can go via DTS, Dolby, WMA-PRO, whatever. It's 5 channels of standard PCM when it's ready to consume.

The .1 (subwoofer) is something that is generally not good, although direction is not something you can hear under 90Hz, you can hear differences in phase between ears as spatial effects down to at least 40Hz, which means that you can't address the original venue's soundfield structure with 1 sub. Minimum of 3, and I'd suggest just use full range speakers, myself. You can use a sub for the 'effects' in a movie, but that's not my thing.

check out
http://www.ece.rochester.edu/~gsharma/SPS_Rochester/presentations/audio2003.pdf

[jimlintott]

Thanks

In the perfect setup for this type of recording would all five speakers be the same? Are they all expected to carry similar loads or would a good surround speaker system do the trick (if they are all good full range speakers)? What about the HTIB (home theater in a box), as these typically rely heavily on their sub to reproduce anything below 100 or 150hz, will they provide reasonable performance or suck?

Is it also safe to asume that the five speakers should be timbre matched?

[jj]

Originally posted by jimlintott
Thanks

In the perfect setup for this type of recording would all five speakers be the same? Are they all expected to carry similar loads or would a good surround speaker system do the trick (if they are all good full range speakers)? What about the HTIB (home theater in a box), as these typically rely heavily on their sub to reproduce anything below 100 or 150hz, will they provide reasonable performance or suck?

Is it also safe to asume that the five speakers should be timbre matched?

Best: 5 identical speakers.
Next best: 5 timbre matched speakers.
Next best: 5.1 sub setup - Won't suck if mains are pretty good.
Next best; not mached mains.

MATCH ALL 5 MAINS. At least in performance if not in appearance. I know how hard it is to put a TV in the middle of a center speaker.

[Just thinking]

Originally posted by jj
Best: 5 identical speakers.
Next best: 5 timbre matched speakers.
Next best: 5.1 sub setup - Won't suck if mains are pretty good.
Next best; not mached mains.

MATCH ALL 5 MAINS. At least in performance if not in appearance. I know how hard it is to put a TV in the middle of a center speaker.

OK, JJ, let's agree that it is possible to improve on 2-channel, but I still hold onto my premise, as you seem to imply as well. I have 2 B&W 801's - you argue that the best arrangement is to now have 5 of them. Ahhh, that's now huge $$$$. Even going down to just using other 800 series speakers is $$$$. I have an Audio Research SP14, just how much is a 5 channel ARC unit going to run? I'm using 2 Cary SLAM-100's (purchased used) -- although they are no longer made, 5 new ones would have run around $12,500. And I haven't even touched my source components. Then after reading the articles (and other comments) one quickly realizes that the majority of properly recorded mulitichannel source information is almost non-existant. I also have over 3000 LP's and 500 CD's all in 2 channel. I'm not much of a home Theatre fan as I don't find most modern movies to be very good -- too dumbed down in many cases, and they play action over substance. Therefore, I simply don't see the cost benefit. For all that extra needed $$$$ I can purchase a lot of software (read - music). I noticed in the articles that you mention the recordings not always being the best renditions, probably due to such a limited selction, no? Also, you seem to have a setup unlike most others, which would explain why I (and many others) may not have ever heard multichannel done right -- but I will not be won over just from the "Wow" effect. I must have something good to play on it. Can you obtain a properly recorded recording of Beethoven's 4th as done by Karajan from the 1963 Beethoven Cycle? (Even his later digital versions were poor performances compared to this one.) And that's just the tip of a huge iceberg of recordings. Without ample good source material it's like owning a Ferrari on an island with no paved roads.

PS: Please do not think that just because I have a good amount of 2-channel software that I am locked from ever going into mutichannel. I have over 800 VHS tapes (and continue to buy them when necessary), but have also gone into DVD. Why? The software is there.

[jj]

Originally posted by Just thinking
OK, JJ, let's agree that it is possible to improve on 2-channel, but I still hold onto my premise, as you seem to imply as well. I have 2 B&W 801's - you argue that the best arrangement is to now have 5 of them.

Ahhh, no, you don't even need 2, is the point.

Since you're no longer depending on the speaker to provide you with what the original soundfield capture didn't capture, you may find that you don't need 801's, or even anything near that, at all. You may find that "lesser" speakers suffice perfectly well.

If you already have 801's, yes, I agree, that's a problem.

As I said, part of what loudspeakers, especially expensive ones, to, is control direct/reverberant ratio differently than each other.

When you conciously provide the reverberant sound, or at least the "direct" component of it, this becomes much less of an issue.

[Jyera]

Any idea how this might turn out?

If I had a pair of speakers playing simple mono music/sound.
I pull a cable of length X to the right speaker.
and pull a cable of length Y to the left speaker.

What X and Y will form a perceptible difference?

[ktesibios]

Originally posted by Jyera
Any idea how this might turn out?

If I had a pair of speakers playing simple mono music/sound.
I pull a cable of length X to the right speaker.
and pull a cable of length Y to the left speaker.

What X and Y will form a perceptible difference?

IMHO, that will depend on the gauge of the wire (which controls resistance per unit length) and the impedance versus frequency characteristics of the speakers (which form a voltage divider with the source impedance of the amp and cable).

It seems unlikely that any practical length of cable will introduce enough dfference in propagation delay to be audible, but as for the effects of too long and too thin- I've measured amplitude versus frequency irregularities of +&- 2 dB or so at the terminals of an NS-10 that was hooked up with thirty feet or so of 22 ga. shielded cable- irregularities that definitely were not present at the output terminals of the amplifier. (Don't blame me- I didn't wire that room, I just corrected the problem.)

BTW, if you want to listen to mono program material, you're better off with a single speaker than with two laterally separated speakers being fed the same program. The comb filtering produced if you're off-center between the two speakers won't do your listening experience any good.

The next time you happen to be at an outdoor event where the sound man has done the customary thing of setting up a speaker system at each side of the stage, try this fun experiment:

Stand at a point equidistant from each speaker and about as far back as the separation between speakers or a little farther.

Wait until someone is talking over the sound system (the effect is really really obvious on the spoken word).

Rock from side to side a little- a few inches either way should do it.

Listen to the flanging.

[ktesibios]

Originally posted by jj
Blumlein and Gerzon both have the same problem, by the way, they are both co-incident miking systems. This makes for a hard-center listening position that is very, very sensitive, and for a very small, narrow sweet spot for listening. You can see how this arises simply by looking at how the ear responds to second wavefronts vs. first wavefronts. (Haas effect)


Just wanted to add that having used coincident (Blumlein and X-Y) miking in the studio, my experience agrees with jj.

In the control room, sitting smack dab at the center and listening on coaxial monitors, the left-right imaging was phenomenal. At home, slouching in the living room, it pretty much collapsed. On headphones, it wasn't quite realistic either.

OTOH, what did translate outside the studio was the impression of a mob of ignorant peasant scum waving pitchforks and lighted torches, so on the whole I'd classify it as successful. :D

[Jyera]

Originally posted by ktesibios
IMHO, ... snipe...

BTW, if you want to listen to mono program material, you're better off with a single speaker than with two laterally separated speakers being fed the same program. The comb filtering produced if you're off-center between the two speakers won't do your listening experience any good.

The next time you happen to be at an outdoor event where the sound man has done the customary thing of setting up a speaker system at each side of the stage, try this fun experiment:

Stand at a point equidistant from each speaker and about as far back as the separation between speakers or a little farther.

Wait until someone is talking over the sound system (the effect is really really obvious on the spoken word).

Rock from side to side a little- a few inches either way should do it.

Listen to the flanging.

1. Can you describe the effect on spoken words, as described above? Echo?
2. What is flanging ? Can you describe it?
3. What is comb filtering?

[Jyera]

Originally posted by ktesibios
IMHO, ... snipe...

BTW, if you want to listen to mono program material, you're better off with a single speaker than with two laterally separated speakers being fed the same program. The comb filtering produced if you're off-center between the two speakers won't do your listening experience any good.

The next time you happen to be at an outdoor event where the sound man has done the customary thing of setting up a speaker system at each side of the stage, try this fun experiment:

Stand at a point equidistant from each speaker and about as far back as the separation between speakers or a little farther.

Wait until someone is talking over the sound system (the effect is really really obvious on the spoken word).

Rock from side to side a little- a few inches either way should do it.

Listen to the flanging.

1. Can you describe the effect on spoken words, as described above? Echo?
2. What is flanging ? Can you describe it?
3. What is comb filtering?

[ktesibios]

Originally posted by Jyera
1. Can you describe the effect on spoken words, as described above? Echo?
2. What is flanging ? Can you describe it?
3. What is comb filtering?

Sorry about the jargon.

1. The best that I can describe flanging in words is a sort of whooshing, coupled with the sort of hollowness you can get by speaking into a cardboard tube from a roll of paper towels. I can actually do a pretty good impression of a flangey sound in person, but that don't help much. Besides studio effects devices, there are flanger pedals for electric insturments, so if you know someone who plays electric guitar, they might have one and be able to demonstrate the sound for you.

3. the recipe for a comb filter is to take the signal, pass it through something providing a constant delay time (a digital delay line, an analog bucket-brigade delay line, or the time it takes sound to propagate a given distance through the air). Then sum the delayed signal with the undelayed signal in equal proportions.

At frequencies where the delay time is equal to an odd multiple of the duration of one-half cycle, the straight and delayed signal will be 180 degrees out of phase and will cancel. At frequencies where the delay time is equal to an integer multiple of the duration of one cycle, the straight and delayed signals will be in phase and will sum to double the amplitude of either (a 6 dB increase).

If you plot the resulting amplitude versus frequency response on a graph with a linear frequency scale, the peaks and dips will be evenly spaced, like the teeth of a comb, hence the name.

2. Now, if you can alter the delay time of your delay element so that it varies as a slow, periodic funtion of time, the peaks and dips will move on the frequency axis. That changing frequency response is what produces the "whoodhing" sound. You can make the effect more intense by feeding some of the flanged output back to the input of the delay line.

I mentioned spoken word because I personally find the flanging effect much more obvious on that than on something busier, like a live band.

Comb filtering when using two separated speakers to reproduce mono program is due to the time it takes for the sound from each speaker to reach you. If you're slightly off-center, the sound from each speaker will be at about the same amplitude at your ears, but the time of arrival will differ according to the difference in distance to each speaker. As an example, suppose that you have two speakers ten feet apart and you're sitting ten feet from the line connecting the two, one foot to one side of the centerline.

The distance from you to the near speaker will be 10.77 feet, to the farther speaker 11.66 feet. Assuming a velocity of sound of about 1130 fps, the time it takes sound from the near speaker to reach you will be 9.53 ms, for the far speaker 10.32 ms, so the time delay between the two will be 790 us. The difference in level due to the difference in distance will be about 0.69 dB.

The result will be that there will be dips about 22 dB deep at 633 Hz, 1899 Hz, 3165z and so on, and peaks of about +5.68 dB at 1266 Hz, 2532 Hz, 3798 Hz and so on.

If you move from side to side, the difference in path lengths will change, and consequently the frequencies at which there are peaks and dips will change.

In most indoor listening environments, there are enough reflections from surfaces less than a few ms away from each speaker to cover up the combfiltering of the direct sound, but outdoors, which approximates free-field conditions (save for reflections from the ground), or in a pretty "dead" room, the effect will be audible flanging.

It's also known that comb filtering due to reflections from the folds in our outer ears plays some role in localizing high frequency sounds, and that comb filtering due to reflections from surfaces near a speaker interferes with that process.

[ChrisH]

Originally posted by ktesibios
Sorry about the jargon.

1. The best that I can describe flanging in words is a sort of whooshing, coupled with the sort of hollowness you can get by speaking into a cardboard tube from a roll of paper towels.

Is that the same as what we used to call "phasing" or "skying"?

It could be produced by running two copies of the same recording, mixed, and then slightly varying the speed of one of them.

[ktesibios]

Originally posted by ChrisH
Is that the same as what we used to call "phasing" or "skying"?

It could be produced by running two copies of the same recording, mixed, and then slightly varying the speed of one of them.

It's a very similar principle to what's usually called "phasing". The difference is that a "phase shifter" effect uses a variable all-pass filter rather than a constant time-delay element.

The trick of mixing two copies of the same recording is exactly how flanging was first done in the studio. In fact, the term "flanging" is said to come from the practice of applying pressure to the supply reel flange of one recorder to vary the speed, and hence the delay between the two signals, slightly.

edited to add:

Incidentally, if you want a couple of really sterling examples of each effect, about the best flanger I know is the old MXR Auto Flanger, and the best phase shifter is the roughly contemporary Mutron Bi-Phase.