Princeton invention delivers 3D sound from ordinary laptop speakers

[Jay481985]

[Mallette]

Well, I've listened several times and I will need to try it elsewhere. My office computer has two Frazier Monte Carlos up on top of bookshelfs in the corners of the room and sounds quite good, if a bit off access unless you are standing.

Neither the flies nor the conference room provided any localization I could get a handle on even after several playings. As this is an inherently simple system using an external M-Audio Audiophile USB and a T amp, it's hard to imagine any setup issues.

I'll keep an open mind, but the buzzing fly was no where near as successful as Hafler's DynaQuad demo LP I have that features the best localization of surround I've ever heard. The "Flight of the Bumblebee" on that disk literally and umabigiously flies around the head.

This one, not so much...at least yet. I'll try it elsewhere.

Dave

[djk]

Gee, he re-invented the Carver Sonic Hologram Generator (US 4,218,585), the Polk SDA (US 4,489,432), or the NAC by Alan Hulsebus (US 5,677,957).

[Mallette]

Gee, he re-invented the Carver Sonic Hologram Generator (US 4,218,585), the Polk SDA, or the NAC by Alan Hulsebus (US 5,677,957).

Hard to tell. I followed his explanation...which all of us already know...as to why binaural doesn't work on speakers, but he didn't really explain what he was doing to "cancel" the spurious cues from L and R to the ear. While something of a "holy grail," it also has always struck me as being as likely as the philosopher's stone in reality.

I still think 4 mikes and 4 discrete channels is the only way to provide localization info to the brain. It's the only thing I've heard that works reliably and organically...

I do not dismiss the idea of circuits that might handle this, and DPLII can do so with very good engineering, but it's complex and as we all also know is a "rare medium well done" in the few instances where it works.

Dave

[Danzo]

Like video 3D, 'Polarizing ear filters' is the future. Preferably pink and fluffy.

[djk]

I figure from his description, and the sound demo of the processed signal, that he was doing what the above three do (although enough different to justify a patent for each).

I am, of course biased, I built the prototype of US 5,677,957 that was shown at the fall COMDEX 1997. I designed a much simplified version that was built into a stereo TV made and distributed in China (for communal living areas).

Alan is Lead Engineer and Designer for Orca Design & Manufacturing (Focal, Raven, PHL) , and I currently work for Rockwell Collins (avionics).

[Mallette]

djk, what are the "above three?" Your qualifications have my attention.

My impression of the systems that I have heard, including sonic holography, the computer (I think it's SDS or similar...can't recall but delivered with many soundcards for a long time), et al, is that they offer an illusion but are never either entirely accurate or predictable.

The only convincing and reasonably stable complete images I've heard are from my own "virtual presence" experiments of a few years back and from Hafler's demo record. As my own system requires 4 identical speakers precisely placed it is not really practical, though certainly is the most simple. Hafler, I don't know. I never read or saw a clear mike or mix plan for his LP so I don't really know what he did to get such a clean and clear sound field but he either guarded it very carefully or just couldn't get the attention of the "big boys."

Certainly, our current systems are descended from his work, but all are overly process intensive and the engineers still can't seem to deliver a clean, clear soundfield reliably.

Even though 2 channel is very much an acquired taste that cannot deliver anything like reality, aesthically pleasing stereo is far easier to produce and harder to seriously screw up than any real image.

Personally, as much as I think it overkill, I feel that digitally steered multichannels (and I mean 20 or more) is the most likely path to real 3D imaging. I've described this elsewhere, but I believe that a 4 mike array can be feed to a computer program that could digitally construct and steer as many channels as one wants and then deliver them as accurately as the number and positioning of the playback speakers would allow. That is, two channel stereo (or mono, for that matter) if only two speakers were present or 20 channel surround if that many speakers were present and the playback system knew both how many and how they were arrayed.

Even letting the imagination go wild it's hard to figure out a way that would work within the limits of reproducer technology as we now have it. Perhaps some day we might have "transducer walls" that could be energized at any specific x/y coordinate to produce sound from a specific point in space, sort of an acoustic holodeck...but that it is just a dream at the moment.

Dave

[BluBitRates]

Well, I've listened several times and I will need to try it elsewhere. My office computer has two Frazier Monte Carlos up on top of bookshelfs in the corners of the room and sounds quite good, if a bit off access unless you are standing.

Neither the flies nor the conference room provided any localization I could get a handle on even after several playings. As this is an inherently simple system using an external M-Audio Audiophile USB and a T amp, it's hard to imagine any setup issues.

I'll keep an open mind, but the buzzing fly was no where near as successful as Hafler's DynaQuad demo LP I have that features the best localization of surround I've ever heard. The "Flight of the Bumblebee" on that disk literally and umabigiously flies around the head.

This one, not so much...at least yet. I'll try it elsewhere.

Dave

Mallette as i am sure you knew he said it was designed for laptop speakers. The effect was pretty good on my end. I agree about multichannel being the future of 3d i am sure there is going to be a dsp in the short future that boasts it or even a dsp combined with specific source material.

[Mallette]

The speakers shown in the video are smallish, but a bit more than laptop. However, I did note the title and have considered that I should try it at work monday. All my home systems are on "real" speakers.

In any event, seems to be a development with a severely limited application. The SDS (I'm not looking that up, but I think that's the ubiqituous free app I mentioned) does a good job enveloping the listener with laptop speakers.

Problem is, who cares? I don't want a cathedral in a shoebox...

Dave

[djk]

"what are the "above three?" "

Carver Sonic Hologram Generator (US 4,218,585), the
Polk SDA (US 4,489,432), or the NAC by Alan Hulsebus
(US 5,677,957).

"they offer an illusion but are never either entirely accurate or
predictable."

I would agree, but that is mainly a function of the recording, not the hardware (in specific, the Carver).

The biggest problem with the Carver (or any similar device) is the recording. The cross-talk cancellation is done with an all-pass delay circuit, and after this is done it is easy to hear that too much delay/echo/reverb has been added to all commecrial recordings (to make them sound more spacious).

The Princeton demo with him pointing at the sources is childs play to any of the three I mentioned. All three will produce stable images that you can point to sources that are further apart than the speakers, or even behind you. I used to demo the Carver with the Billy Cobham CD Warning!, there are some incredible directional cues in that recording.

Some recordings (Jeff Beck Guitar Shop) sound like they have been run through the NAC processor that Alan came up with, and sound best with the NAC turned off (as they are already processed).

These type devices sound their best with two speakers that are close together, like a computer desktop system, a boombox, TV, or a rack/tabletop stereo (with the speakers virtually touching the rack).

Alan's patent expires next year, and I'm sure he wouldn't mind anyone making one for their own use. The schematic (with the preferred part values) is/are in the patent, only a quad opamp and a handful of resistors and caps.

In terms of a 'real' stereo, the Klipsch 3-channel method sounds most like real music, especially for people off-center, or when you move your head. With the above three, any head movement causes the soundfield to virtually collapse. It can feel like your head has been in a vice after several hours of listening while trying to hold your body motionless. The NAC has less of this effect than the other two, and sounds fine even if you're not in the 'sweet spot' (unlike the Carver).

With the Klipsch 3-channel method, the music sounds like you are there, and works the same way. If you are seated off to one side, it sounds like you are seated off to one side. If you turn your head to face a soloist, the image smoothly shifts just like it would do if you were in the same room as the musicians, and the sound you turn to face gets louder. Again, the quaility of the recording is paramount (listen to The King James Version. It was recorded with a single stereo mic, and then they decided to add a close mic for the drums. The music sounds just like the stereo mic position would have, mid hall, except it sounds like you are sitting at the foot of the drum kit for the drum sounds). The more natural the recording, the more real the Klipsch 3-channels method sounds.

Do we really need multi-channel playback?

The industry says yes, as they expect to sell more hardware and more new recordings for said format.

A number of years ago a car design I did was doing very well at IASCA. It had L, Center (L+R), R, Surrounds (L-R), and Subs; all from 3-channels of amplification, less than 100W! I later changed the design to a 4-channel amplifier with delay to the surrounds, you would be amazed what sound can be had from a 20Wx4 amplifier.This basic method also works well for use at home, but requires the speakers have a very flat impedance curve for best results, or another amplifier channel. Here is the basic 2-channel amplifier set-up for driving L, Center (L+R), R, Surrounds (L-R):

http://i56.photobucket.com/albums/g196/dkleitsch/passivesurroundandcenter.gif

[JJkizak]

In 1969 I set up two Shure omnidirectional mics (50-15000hz) with the K-horns and Cornwall powered center speaker, the mics 3 ft from the K-horns even with the tweeters. Then recorded to an Ampex tape deck the evenings christmas party. When played back somewhere around 90 dbm SPL the spatial location of sounds was essentially and totally perfect. The phone rang and three people got up from different locations in the room to answer it only to find out the phone wasn't ringing. It didn't matter where you were sitting or standing during the playback as everything was perfect. I expect in the future someone will set up a digital recording device to program your amplifier to modify all incoming musical data to dial in your room acoustics. Of course this would require another digital unit or modification of the present amplifier strategies to impose the recorded data on same.As long as you do not move your speakers the dial-in will remain accurate.

How would you create this kind of digital unit? Don't know.

JJK

[Mallette]

Thanks for the excellent review!

Do we really need multi-channel playback?

I've mulled that over and the answer depends upon just who you mean by "we." It's been a "duh" question for me as I've always favored music where the space is integral: choral, pipe organ, orchestra. I've always dreamed of what the great "Glory of Gabrieli" would sound like with true fidelity...that is to source of the sound and return as well as the sound itself.

However, the audiophile community has never seemed to much care. I've finally realized that many listen to things were the space is irrelvant, mixed and compiled rock, studio recorded jazz, and the like and have acquired a taste and preference for that "spaceless" form. That's cool.

While my preference for true fidelity to all aspects of an acoustic space/time event was always there, it was the Firesign Theatre where I first heard it using a DynaQuad back in the 70's. They had so much out of phase material due to the overdubbing that voices and sounds came from all over the room, sometimes right in your ear. Truly extraordinary audio theatre! Then I found that many LPs benefitted with an added sense of depth and space.

Sense then I've always hoped for a simple and reliable system to provide space fidelity but it's never happened. It seems that DPLII or DTS should be able to deliver but it never does. I find it hard to believe there are simply no engineers out there who understand that all you have to do is use four coincedent identicle mikes placed where you want your ears to be and then get the tracks to the amp as virgin as possible.

Maybe yet...

I was just reading that TAS recommends 5 surround recordings, one of which is a SACD reissue of the Biggs Freiburg Cathedral recording of the four antiphonal organs. I've an original quad LP and it sounds nice, but the directionality and space aren't really convincing. Gotta check this one out...

Hope springs eternal!

Dave

[DrWho]

What about all the research from Bell Labs where they concluded 7 channel surround was the minimum for accurate 360 degree spatial positioning?

[Mallette]

What about all the research from Bell Labs where they concluded 7 channel surround was the minimum for accurate 360 degree spatial positioning?

Well, Mike, that would depend upon the system being used. I certainly proved to my own satisfaction and that of a few who heard my experiments that one could accurately reproduce downright spooky reality with with four mikes, place coincedent, and replayed on four descrete channels to four identical speakers. I didn't need Bell Labs to tell me sound comes from 360 degrees and recording that information and feeding it to the brain for decoding should replicate that situation.

Granted, the sweet spot was small but the effect was extraordinary.

I don't know what will finally work. I know it has to be simple and cheap at both ends, like my own experiments. My gut says that only a lot of drivers all independently targeted will work for a wide sweet spot and accuracy suitable to the masses...but my gut also says they don't really care.

Not sure how or when this will work out. Certainly technology will eventually provide a means. When is the question.

Dave

[DrWho]

Maybe someone else that knows of these experiments could find the writeup, but the point of the investigation was to determine the fewest number of speakers required....and the simple 4 channel approach you mention was definitely part of the experiment - which isn't to say that 4 channel sounds bad, but there are some situations where the 4 channel falls short (and I believe it was you who said he wanted the utmost in fidelity in that regard). The biggest issue comes into differentiating sounds centered in front versus centered in back. You can get some really weird effects that I've personally found make it real easy to lose focus of the soundstage.

Btw, I find that I really enjoy exploring the soundscapes of recordings that weren't trying to capture the ambience of a real performance hall. If you can let yourself zone out and get into the music, some of the "over-processed" stuff can really make the listening a lot more exciting - I dunno if I'm alone in that regard or not, but everytime I listen I'll hear something totally crazy different and run down a different rabbit hole of audio nirvana. I think a lot of Loreena McKennitt's work exhibits a lot of that behavior. Her earlier stuff is a bit sloppy, but later on I think it's gotten much better polished.

I dunno, I just hate to think someone is missing out on good music simply on the premise that it's not striving for the sound of a particular hall...

[JJkizak]

My experiment did not have any sweet spot. Didn't need one. Didn't matter where your location was. It was perfect.

JJK

[Mallette]

My experiment did not have any sweet spot. Didn't need one. Didn't matter where your location was. It was perfect.

JJK

If you were getting 360 degree audio from two mikes replayed through three speakers you've done something no one has done before. I assumed from what you wrote you were talking stereo. I've re-read it and still don't get it.

Dave

[JJkizak]

Mallette:

It only works in the room you record in and only if you record in the room. Otherwise any other medium does not work.

JJK