Tate DES

[Disclord]

I've talked with Martin Willcocks (via email), inventor of the Tate DES, and I wish we could get him to develop a modern DES chip for SQ decoding. He told me he's continuted working on the DES and designed a precision 16-pole phase shift circuit, ideas for the decoder to adaptively identify the type of SQ encoder used (such as Forward-Oriented, Position, Ghent, etc...) and modify the decoding coefficients for most accurate decoding, and new, optimized B-matrix decoding coefficients. He'd like to do it digitally, all in DSP. He said it's the lack of funding that's holding him back from doing it - that it would cost a lot of money. :-( Someone needs to win Lotto so we can fund it!!!

As a side-note, I've also talked to Lynn Olson, inventor of the Shadow Vector for SQ, a few times too and gotten lots of info about the SV decoder and its development... if anyone is interested, I'll post about it in the next few days.

 

[The Quadfather]

I have always been curious about the shadow vector system, was it ever produced and sold to the public? And how did it compare to the Tate DES?

 

[Disclord]

I have always been curious about the shadow vector system, was it ever produced and sold to the public? And how did it compare to the Tate DES?

No, it was never sold and never made it out of the prototype stage. Audionics decided to abandon the Shadow Vector system and work on the Tate DES because they thought the Tate would make it to the market faster (little did they know!).

 

[JonUrban]

I've moved a few posts from the CD-4 thread to their own thread here.

Carry on! :banana:

 

[Lynn Olson]

Hi all! I finally discovered the forum after coming across a link on AVSForum - tracking down info on the discrete quad mix of Dark Side of the Moon. It's been many years since I've heard the SQ/EMI version of DSOTM that we used to play through the Shadow Vector prototype (only one was made, and it is long gone). As I recall, it was superb mix, far superior to the stereo version of the LP.

As Disclord mentioned at the top, I had a fun conversation about the TATE and Shadow Vector systems. Audionics had both systems on tap, although it was TATE that was reduced to chip form, not my system. Wesley Ruggles had Hollywood backers with deep pockets, while I was 25-year-old employee at Audionics, so TATE got the nod, and I was re-assigned to loudspeaker development. At the time, I was under the impression the two systems were pretty similar, but based on Disclord's conversation, they were more different than I knew.

Shadow Vector was dynamically optimized to deliver 25 dB or better separation at all points in a 360-degree circle, not just the 6 SQ cardinal points - in other words, the phantom images were treated exactly the same as loudspeaker localizations. This was intended to prevent "detenting" towards loudspeaker positions, and create a smoothly distributed reverberant field all through the room - something modern HT systems still seem to have trouble doing.

 

[Lynn Olson]

I should clarify the term "dynamically optimized", which sounds like marketing gibberish. The usual way decoders are set up is to measure separation at the cardinal points using test tones, one at a time. This tends to favor decoders that are optimized for the 6 cardinal points of SQ, or the 4 cardinal points of QS or Dolby Surround. Separation for the intermediate (phantom) localization usually isn't even tested - and unless this is a design priority of the decoder, this figure may be quite different.

The test Cliff Moulton and I used to align the Shadow Vector was a pair of sine-waves driving the LT and RT inputs, generated with a pair of oscillators a few Hz apart. This creates a spinning localization (the speed depending on the frequency offset between the two oscillators). By examining the 3-axis logic control signals and output levels coming from the decoder outputs on an X-Y scope, you can literally see how much separation there is across a wide range of localizations (it looked a bit like early radar signals). If the control logic wave shaping was just a little out, there would be "bumps" associated with separation variations at different localizations. Similarly, by adjusting the speed of the spinning localization, the attack and release times of the logic circuitry could be optimized so it wouldn't "bounce" under dynamic conditions. So the separation optimization, unlike other folks, was a test done under dynamic conditions, not a static single frequency test for a limited number of localizations.

I had assumed that TATE was doing this as well; but according to the conversation from Disclord, this was not so. I was never given much information at the time on the internal workings of the TATE DES system; the information flow was all one-way. I must confess I still don't fully understand how the TATE feedback system works, or even the need for it at all. We could easily achieve 40 dB separation by having very precise 90-degree phase shift matrices; the performance of the decoder was essentially limited by these, not the precision of the variable-gain elements, which were controlled by a multi-element resistor-diode array that shaped the control waveforms.

 

[The Quadfather]

Dear Lynn:
I am an early adopter of the Audionics Tate system, I bought mine back in 1979. I never got the upgrade either. Mine is an original. It is still working, and I love it. I took out a bank loan to pay for it, and it met my expectations and then some. It blew a chip once, and I fixed it myself. That was back when you could still get the parts. Still, it would be interesting to hear a system based on the principles you describe. But alas, the shadow vector system will never be heard again.

The Quadfather

 

[Lynn Olson]

I never spent much time listening to the Audionics "Space and Time Composer", Audionics' name for their TATE DES decoder. In 1979, I left Audionics for a job as tech writer at Tektronix (also in Beaverton, Oregon), and didn't stay in contact with the Audionics folks much after that. I think the S&TC came out in 1980 or 1981, well after I left, and a time when Audionics' business was falling off - they went from 22 employees down to 5 or 6, from the rumors I heard. When Audionics closed its doors a few years later, Jim Fosgate bought the intellectual property of Audionics, which included the S&TC technology. As for the physical Shadow Vector prototype (an ugly-looking black box), it was still at Audionics when I left in 1979. Yeah, I should have taken it with me for historical reasons alone, but I was kind of burned-out on the hifi industry at the time.

During the time I was Audionics, from 1973 through 1979, I had little contact with the Wesley Ruggles and Martin Willcocks. I don't know why: it might have been them, or it might have been Charlie Woods (president of Audionics) keeping me away from his associates. For reasons of his own, Charlie kept me away from Dave Fletcher of Sumiko, who I had wanted to meet while I was working at Audionics, and I finally did get to meet just last year at the Rocky Mountain Audio Fest. I have no idea what Charlie's or Wesley's motives were, and I only had one brief meeting with Wesley in Los Angeles, where I showed him a set of equations and schematics for converting a Sansui QS decoder to a mysterious new format that encoded Left, Center, Right, and Back. I was paid $500 in consulting fees, and it was very hush-hush at the time. Years later I first heard Dolby Surround in a movie theater, followed by a big "AH-HA" moment just what the conversion matrix was used for.

The whole Wesley Ruggles/Martin Willcocks thing was shrouded in mystery while I was working on Shadow Vector, and afterwards as well. I played no role in the development of the Space & Time Composer, nor the reduction of TATE DES to National Semiconductor chip form. Audionics did use the Shadow Vector two-knob separation optimizer system in the S&TC and the preamp that came after the BT-2. So the information flow was unidirectional from my perspective - TATE DES was some kind of black box that used dynamic decoding for SQ and QS content - that much was obvious from the sales material. But the info on "sum of the K's" and "feedback optimization" of separation was something I never understood, along with the unusual system of separation-enhancing a static SQ decoder with an add-on circuit, instead doing at all at one go, like Shadow Vector. The URL to the Shadow Vector patent is here, if anyone is curious how it works:

http://www.nutshellhifi.com/library/pat4018992.pdf

 

[Lynn Olson]

I should add that it's never been clear to me if the TATE DES was a constant-separation system with identical performance for phantom images as well as the loudspeaker cardinal points. From the limited talks I had with CBS engineering team at the Chicago CES in the mid-Seventies, that was never a goal for the CBS Paramatrix. At EMI Labs in England in 1975, I heard a direct comparison between Shadow Vector and the Paramatrix (for the first time), and it was obvious from the panning effects that it did not maintain full separation for circular pans, with the images on the sidewalls becoming noticeably diffused and decentered during a front-to-back pan. There was also more "detenting", as well as a much dryer sound overall.

The hallmark of the Shadow Vector was a very "wet" sound, thanks to the reverberant energy having an exactly 360-degree diffusion with no hotspotting at the speaker locations, as well as fully symmetric dynamic operation (no favoring of cardinal points with fast pans and quick point-to-point shifts in localization). The wetness was real, though; it was just more apparent when the decoder wasn't trying to force random energy into preferred locations. That was my big takeaway from the Shadow Vector project; random reverberant energy had to have a smooth distribution throughout the room, or the sound would get dry, like a studio jazz recording with not enough reverb.

I don't know if Jim Fosgate's modern Dolby Pro Logic II decoder has the same freedom from detenting that Shadow Vector had; the temptation to optimize separation at the cardinal points was hard to resist for almost everyone with a background in SQ decoders - I think I might have been the only exception, at least based on the sonics of the SQ decoders I've heard. By contrast, the Sansui QS team were much more focussed on smooth soundfields, although their decoder always sounded a little lazy to my ears. The BBC/UHJ/Cooper/Gerzon decoder always had issues with a rather diffuse and phasey-sounding matrix to my ears, so it was hard to compare to the various SQ and QS dynamic decoders.

This has affected my speaker-design philosophy (where I was drafted after the shutdown of the Shadow Vector project). I aim for smooth, evenly distributed soundfields with crisp localizations within the soundfield; in practical terms, this means reduction of diffraction, and smooth control of inter-driver phase relations in the crossover region.

 

[The Quadfather]

Dear Lynn:
I'm not what you would call a golden eared audiophile, I am happy if the instruments don't sound as if they are coming out of all speakers at once. The first SQ decoder I heard was a Pioneer built in to a receiver, and it sounded like that. The Tate offered real separation or a damned good illusion of it, and that excited me. It wasn't until many years later I was able to put together a first class CD-4 system, so it was either SQ or 8 track for good quad. CD-4 is very good now that good quality stylii are availlable, but back in the quad era, it was hard to find good cartridges when you don't have the money to just buy cartridges and discard them when they didn't work. A lot of equipment sold for CD-4 use just didn't work, even though it (supposedly) had a shibata stylus.

 

[Lynn Olson]

Yeah, I never got my CD-4 system to work well, which partly inspired my work on Shadow Vector. The Grace F8-F (if I recall the model correctly) was a pretty nice cartridge - Shibata tip - but it didn't play all that well with the Pioneer CD-4 decoder. Extremely sensitive to even the tiniest particles of dust on the record, and I kept my records very clean. The PLL in the decoder would drop phase-lock, with a loud bang, and take a moment to regain it. If the record was absolutely pristine, the sound quality was still just sort of so-so, not what I'd really call high fidelity.

I have no doubt that a good moving-coil, like the Denon or better, would have been an immense improvement, since CD-4 requires flat group delay for low distortion. (FM demodulation translates time errors into amplitude distortion; this is inherent to all FM systems.) That little drawback was something JVC skated over - that the FM difference-signal carrier required a cartridge with ruler-flat group delay through 45 kHz, which is phenomenally difficult in a complex mechanical system. We can do it now with advanced computer modelling of the electromechanical system, but not so easy back in 1972, when it would require a mainframe computer and an in-house FORTRAN programmer.

Moving-coils have far more bandwidth, and the Shibata and more recent line-contact profiles do a really nice job to 50 kHz or beyond. Too bad Compact Discs with their brickwall 20 kHz bandwidth replaced LP's at the same time that LP playback was starting to get really good. It was even more unfortunate the Red Book specification from Philips and Sony was so rigid, with no upgrade path to 20-bit resolution or more channels, permanently locking in the technology to the best that 1982 could do. The visual equivalent would have us all watching 25-year-old VHS-quality television.

Quad also suffered from a premature market introduction; the so-called "full-logic" and "half-logic" decoders (using a static matrix and gain-riding) should never have been put on the market. You can bet the CBS engineers were not thrilled to hear me say that at the CES. The only decent decoders were in the labs (and never heard by the public or most reviewers) - the CBS Paramatrix, TATE DES, and Shadow Vector. All of them were very fast, with 2~3 millisecond attack times, and 10~20 millisecond release times. That allows the decoder to precisely localize 50 to 200 images per second, which is subjectively instantaneous.

The hard part was telling the logic when to give up, such as spaced-microphone recordings of classical music, which do not have consistent and predictable phase relationships. I was still working on that problem when Shadow Vector was abandoned in favor of the TATE DES system. (TATE had outside funding, SV didn't, and Audionics was a small hifi company that had trouble making payroll at the best of times.) As it was, the numerous bugs and implementation errors on the part of National Semiconductor almost sunk the project, and certainly delayed it by three years or more.

The Sansui QS decoder was the only model available to the public that offered a fully variable matrix with no CBS-style gain-riding (I owned one, and knew how it worked), but the matrix steering was very slow and I suspect not very accurate, since it used some kind of opto-coupled VCA, instead of the precision VCA's used by the models in the labs.

 

[Disclord]

Hey, Lynn! I've been out of touch with everything quad related for quite a while - my sister died, then a month later my mom died - life has been a mess for the past year!

While I haven't talked with Martin Willcocks for a while, he did tell me a lot about the DES - and clarified some of the misconceptions about it that got spread by people like Bill Sommerwerk, such as the Tate only sensing and decoding the 6 cardinal directions. In the final "Fosgate" version of the DES, it analyzed and enhanced 360 degrees and could enhance 3 directions simultaneously. Here's some excerpts from Martin in his emails to me about his design and performance of the Tate DES:

"The DES was, as you rightly perceived, not a variable matrix decoder in the sense of VarioMatrix, CBS Paramatrix or ShadowVector decoders.... I thought out the idea of making the decision of where the signal was to operate a combination of gain changing and blending by using a superposition of the six "blend" or B matrices with the standard fixed matrix decoding, so that the performance could be similar whatever the source direction was. That was basically my inspiration for the idea of a holistic decoder rather than a piecemeal one. Everything would be considered at once and the best modifying matrix for the particular mix of sounds would result. There were several different approaches that I could have chosen, but it seems that I chose the right option - the six directional matrices would be weighted by the detection of sound levels in each of the six principal directions, and added together, along with the standard fixed decoding matrix to enhance the separation of sounds from any one of these directions. For the SQ system, a linear superposition of three of these weighted matrices onto the fixed matrix works quite well, but it actually does not work so well when applied to QS. As a bonus for SQ, it turned out that the separation of simultaneous sources in up to three axes could be enhanced simultaneously - the front-back, left front-right front, and left back-right back axes were all orthogonal on the energy sphere, so a 3-axis control system was feasible, with six direction control signals and six "blend" matrices. Although I added the center-left/center right (side) axes both for normal and SQ position encoding in the patent, we never really tried out whether the three signals together would work to the best effect. It just happens that they do, allowing separation enhancement of about 9dB for two signals or 6.7dB when three signals are acting together. Since the enhancement is instantaneous (1ms), you never hear that reduced separation. Three axis control comes out of the fact that the pan locus of SQ traverses the energy sphere passing through all three geometric axes, front-back, left-right, and top-bottom, which translate to CF-CB, LF-RF and LB-RB axes in terms of sound locations in the listening area of interest. This is not true of the QS matrix, nor of Ambisonics, which have essentially great circle pan loci, with some distortion of the locations between these axes. So only the Tate SQ DES is capable of three-axis control; the best you can get for other systems is 2-axis control. Three-axis control doesn't work with systems where the pan locus is essentially great circle, which includes QS, UHJ, BBC-H etc. You can use two-axis control, but then the variable matrixing systems do something like that anyway. That's why QS with Variomatrix worked reasonably well, but not as well as SQ/DES. At the 1977 Summer CES, referred to below, Lynn Olson was also present, without the Shadow Vector. I only heard the latter when Charles Wood allowed me to audition it in my Los Angeles home for a few weeks, and to try to capture the schematics from the ptototype circuit boards, which Lynn had built. I think I may still have some of that information in my office, but it's years since I last looked at it, and I may even have prints of the NSC Tate chipset schematics as well.
I have tried a two-band quad setup using two Tate DES's and an active crossover filter to separate the highs from the lows. It's only marginally better than the full band decoder, but I think multiple octave band decoders might be an improvement - just very expensive! There really isn't much need to decode stuff below 200Hz and above 3kHz as the directionality is mainly contained in the central region of the spectrum. Psychoacoustics takes care of much of the rest.
Paramatrix used (axis-crossing) phase detectors rather than amplitude detection as in the Tate. The theory is that the Hilbert Transform converts amplitude to phase and vice versa, which allows zero-crossing detectors to provide information about directional content. I never quite bought into that, as the phase conversion really needs more bandwidth. Still, Dan Gravereaux and Jerry Budelman swore by it and wondered why I had used a complicated agc system and amplitude detection. I think the answer must be that it worked better!
The ADC (Automatic Dimension Control) was only used in the Fosgate DES, but was hinted at by one of the CBS engineers and they had something like it in Paramatrix, but they didn't write a patent on it. Mine being amplitude-based was quite different, anyway. As it turned out (see the ADC patent) the result was that in the side quadrant the two active control signals needed to total exactly 1 unit, while in the front and back quadrants it needed to change upwards slightly, up to about 15% higher. In practice the ADC was about halfway in between these ideal states, so could achieve better than 20dB separation around the entire 360 degrees soundstage for a single panned signal while keeping all other signals in their properly encoded locations. The DES used true analog multipliers in its circuitry. The Fosgate/Scheiber units used FETs as gain control elements, which must have involved some compromises. Im not sure about Paramatrix, as I never saw a schematic. I think the Scheiber unit used FETs as well. If I were to redo the Tate these days it would be a DSP, sub-band decoder using my 16-pole digital phase quadrature networks and incorporating volume, L-R and F-B balance, stereo width and the variable rate and dimension controls, and would have all the other bells and whistles that people like you would want. The patents are all public domain now, long since expired. A project like that would take significant funding and a young keen DSP expert audio engineer.
"

I copied/pasted from Martin Willcocks email so spelling errors and such are his - don't blame me!

I'm interested in comments on what he told me.

BTW, Lynn, until your email of a week or so ago about DD+, I hadn't realized the phase problem with DD - I'll fill you in soon on what Roger Dressler told me back in the AOL days (when Dolby was only on AOL) about DD's phase performance when coupling channels.

My best to all of you,
Ty Chamberlain

 

[surroundophile]

Moving-coils have far more bandwidth, and the Shibata and more recent line-contact profiles do a really nice job to 50 kHz or beyond. Too bad Compact Discs with their brickwall 20 kHz bandwidth replaced LP's at the same time that LP playback was starting to get really good. It was even more unfortunate the Red Book specification from Philips and Sony was so rigid, with no upgrade path to 20-bit resolution or more channels, permanently locking in the technology to the best that 1982 could do. The visual equivalent would have us all watching 25-year-old VHS-quality television.

CD Redbook provided for 4 channels (Quad) form day one. It's just that nobody implemented it (as Quad had just recently died)

 

[Disclord]

Moving-coils have far more bandwidth, and the Shibata and more recent line-contact profiles do a really nice job to 50 kHz or beyond. Too bad Compact Discs with their brickwall 20 kHz bandwidth replaced LP's at the same time that LP playback was starting to get really good. It was even more unfortunate the Red Book specification from Philips and Sony was so rigid, with no upgrade path to 20-bit resolution or more channels, permanently locking in the technology to the best that 1982 could do. The visual equivalent would have us all watching 25-year-old VHS-quality television.

CD Redbook provided for 4 channels (Quad) form day one. It's just that nobody implemented it (as Quad had just recently died)

Yep, both CD and JVC's AHD (which was also selected, along with CD, as the 'standard' for digital sound by the DAD Group) had 4 channel modes that could be used either for quad or for 3 channel stereo with the remaining channel used for graphics and such. The only problem with the CD 4 channel spec is that it was lower sample rate (32kHz, if I remember correctly) and spun at double speed with 1/2 the play time. R-DAT also had a quad mode. It's too bad no one ever implemented Michael Gerzon's system of using the CD's subcode to increase the sample rate & bits or number of channels.

 

[Lynn Olson]

It's really too bad that corporate politics prevented Martin and I from meeting and collaborating - especially considering we were both connected to Audionics. I mean, how many people were working with variable-parameter SQ decoders, anyway? Add up everyone on the CBS and TATE development teams, and myself, and it couldn't have been more than 10 people, worldwide. Did we work together? Nope.

The net result of all this petty rivalry in the SQ world was a disastrous 5 to 7-year delay in the introduction of a really good-sounding decoder - my feeling is this delay played a major role in sinking quadraphonic. The public was stuck with the artifact-prone "Full-Logic" CBS/Sony decoders, and almost nobody even knew Paramatrix, Shadow Vector, and TATE existed, much less heard them. The recordings were glorious - but only a handful of people ever heard what was really on the disc. My personal copies of Bridge Over Troubled Waters, DSOTM, and Santana got plenty of rotation at the Chicago CES demos.

The Shadow Vector used a fast AGC circuit with a 40 dB range and an array of high-gain precision rectifiers to do the 3-axis logic sensing. I call it 3-axis because each axis was bidirectional: center front/back, left front/right front, and left back/right back, and each decoding (output) channel was independently steered in 2 directions, along with a constant-gain correction as it moved around the surface of the Scheiber sphere. Modifying it for QS or any other matrix, or going from 4 to 6 channels, was just a matter of signal switching, although this was not done in the prototype.

The part I didn't like was Cliff Moulton's (Audionics chief engineer) time-division VCA's, which ruined the public debut in upstate New York with loud amounts of AM-radio RFI noise coming into every channel. I still don't know why Cliff insisted on wasting months on designing his own time-division VCA's, when we could have bought off-the-shelf VCA bricks. I suspect the real answer was that Audionics was too cheap to buy a set of studio-grade VCA's. Opto-couplers (as used by Sansui and the Sony half-logic SQ decoder) were out because they were far too slow and did not offer the 1% precision we needed.

Audionics was also too cheap to buy a pair of BBC LS 3/5A's when we went to the UK in 1975, and then wasted 8 months waiting for the BBC to grant us a license to make a US version (not surprisingly, they refused). While we waited, several of our competitors introduced their own (unauthorized) copies of the LS 3/5A's - while we sat on the sidelines, twiddling our thumbs, waiting for a letter that never came. Missed opportunities, ah well, that's the story of Quad, SACD, and DVD-A, isn't it?

P.S. Just got the latest VT25 3D Panasonic plasma TV and matching 3D Blu-Ray player. Guess what? The Blu-Ray player does not play DVD-A or SACD. Never mind decoding, transmission over HDMI, or anything like that; these discs won't even play. Yup, you got it right, after only ten years, DVD-A and SACD now considered obsolete "legacy" formats on Panasonic's flagship products. So I guess I'll be buying an Oppo so I can enjoy my DVD-A, SACD, and HDCD discs.

 

[MartinW]

As I haven't posted here since 2005, I guess it's about time I put my own material out here instead of secondhand via Disclord who quoted one of my emails above. A few minor spelling errors, I admit; "ptototype" should be "prototype", of course. In the paragraph "It just happens that they do, allowing separation enhancement of about 9dB for two signals or 6.7dB when three signals are acting together." the actual numbers were 6 dB or 4.7dB respectively.

It's good to see Lynn Olson aboard here - Hi Lynn! I absolutely agree with you that corporate shenanigans prevented us from meeting and cooperating on the optimum design for an SQ quad decoder. The only time I met you was that 1977 Summer CES, if I recall correctly, when we had the face-off between the DES, Peter Scheiber's new decoder, Paramatrix from CBS and the L-400 CBS decoder that used the Motorola/Fairchild chip set. You certainly had some good ideas with Shadow Vector and by the time I got to hear it, presumably the VCA issue had been solved, because the unit certainly didn't interfere with other equipment.

I have been corresponding privately with Disclord for over a year, off and on, and this particular quote (posted earlier with my permission) was from a recent email.

In the context of Lynn's post of August 31, the way my direction detector worked produced something very close to what he claims, for decoding between the cardinal points where only one logic signal is at a maximum value. Between these points, the automatic dimension control (ADC) system (see US Pat. No. 4,063,032) was responsible for weighting and limiting the two or three active logic signals in a manner that almost perfectly allocated the two or three associated B (for blend) matrices correctly to obtain the maximum separation. In the front and rear quadrants, the sum of the two active control signals should be equal to the maximum values of any of these signals for cardinal directions, and in the side quadrants the optimum value of the sum increases to about 15% higher than this level. In practice, the ADC allowed the sum of these active signals to increase by about 7% in both front and side quadrants, which resulted in a separation of about 20dB, so for a signal like pink noise panned around the listener, almost perfect separation would be heard. When two separate signal sources were at different positions but with equal levels, the separation would be close to 6dB and with three signals, nominally 4.77dB.
The manual dimension control acted to reduce this maximum separation progressively, thus reducing the apparent "size" of the space towards that of a simple non-logic decoder, but without affecting the localization of a given sound source.

I coined the term "Total Quadraphonic Separation" or TQS to describe the ratio between the total power in the one or two channels reproducing a particular sound to the power in the remaining speakers from that sound source. The dimension control works to reduce the TQS from its absolute maximum down to a minimum value, that of the basic SQ decoder without the logic capability.

The "Total Quadraphonic Power" or TQP relates the power in all loudspeakers for a fixed matrix decoder to the total power in all these channels in a decoder using the steering logic. The basic SQ decoding matrix provides a constant TQP for all directions panned around the room, without any logic being active, and the powers in all the speakers should remain constant if the logic signals controlling the variable matrixing are correct, while TQS should increase to the value set by the dimension control. Typically a good decoder should have constant TQP and a high TQS value as the source signal moved around the sound stage.

 

[Disclord]

Martin, I'm glad you're posting here! There is always great info from you.

Ty C.

 

[Lynn Olson]

Hi Martin!!! Good to hear from you! There's aren't many of us dynamic-matrix types left, in the contemporary era of discrete (but compressed) digital. Still trying to decide which of the modes of my new Marantz AV8003/MM8003 provides the best conversion of 2-channel sources - DPL II Music, DTS, CircleSurround, or whatever.

Still surprised that surround music has such a small audience - surround seems 95~99% dominated by the slam-bang Home Theater types, who are into dinosaur thumps, phaser blasts, car chases, explosions, and other synthetic movie effects. Trying to track down a HDMI 1.3A pre/pro or receiver that actually sounded decent on music turned into a major project - the market has really polarized into very loud movie soundtracks with 7 to 9 THX channels vs musically oriented 2-channel, with not much to choose from if you like surround music.

 

[Lynn Olson]

In the context of Lynn's post of August 31, the way my direction detector worked produced something very close to what he claims, for decoding between the cardinal points where only one logic signal is at a maximum value. Between these points, the automatic dimension control (ADC) system (see US Pat. No. 4,063,032) was responsible for weighting and limiting the two or three active logic signals in a manner that almost perfectly allocated the two or three associated B (for blend) matrices correctly to obtain the maximum separation. In the front and rear quadrants, the sum of the two active control signals should be equal to the maximum values of any of these signals for cardinal directions, and in the side quadrants the optimum value of the sum increases to about 15% higher than this level. In practice, the ADC allowed the sum of these active signals to increase by about 7% in both front and side quadrants, which resulted in a separation of about 20dB, so for a signal like pink noise panned around the listener, almost perfect separation would be heard. When two separate signal sources were at different positions but with equal levels, the separation would be close to 6dB and with three signals, nominally 4.77dB.
The manual dimension control acted to reduce this maximum separation progressively, thus reducing the apparent "size" of the space towards that of a simple non-logic decoder, but without affecting the localization of a given sound source.

I coined the term "Total Quadraphonic Separation" or TQS to describe the ratio between the total power in the one or two channels reproducing a particular sound to the power in the remaining speakers from that sound source. The dimension control works to reduce the TQS from its absolute maximum down to a minimum value, that of the basic SQ decoder without the logic capability.

The "Total Quadraphonic Power" or TQP relates the power in all loudspeakers for a fixed matrix decoder to the total power in all these channels in a decoder using the steering logic. The basic SQ decoding matrix provides a constant TQP for all directions panned around the room, without any logic being active, and the powers in all the speakers should remain constant if the logic signals controlling the variable matrixing are correct, while TQS should increase to the value set by the dimension control. Typically a good decoder should have constant TQP and a high TQS value as the source signal moved around the sound stage.

Thanks for the detailed and very comprehensive reply. It's a big mental adjustment for me - I'm used to each decoder channel independently moving across the surface of the Scheiber sphere with fairly small +/- gain adjustments to retain constant power output, while your system adds dynamic signals to the output of a static matrix. It comes out the same as far as I can tell, the visualization, math, and signal paths are just different.

I'm still amazed that CBS actually had to hire a mathematician to get the Paramatrix working. I kind of wonder if they understood their own matrix all that well - particularly the peculiar asymmetries of SQ, like the in-phase and out-of-phase diagonal splits (with standard phasing of the SQ decoder), and the inability to encode a four-channel sum without resorting to delays for the rear channels.

My overall impression is our systems were fairly similar in overall performance, but yours had some rather elegant ways of detecting nasty sources like spaced-microphone classical music, which had fiendishly difficult phase relationships in certain passages (phase spins when two instruments with separate mikes played in harmony with each other).

One advantage of modern digital decoding is "look-ahead" - the logic portion has advance knowledge of what's coming next, which can be used to effectively "pre-steer" the matrix before the musical content itself changes. The look-ahead also allows the rate of change to be just the right ramp for the content that follows, with no overshoots that could be so troublesome in real-time, all-analog systems.

 

[Doug G.]

All I know is that it is incredible for us die-hard quad fans to be reading this stuff from you guys after all these years!

Doug