|  IP: Logged

 |  IP: Logged

------------------
John P. Pytlak, Senior Technical Specialist
Worldwide Technical Services, Entertainment Imaging
Research Labs, Building 69, Room 7525A
Rochester, New York, 14650-1922 USA
Tel: +1 585 477 5325 Cell: +1 585 781 4036 Fax: +1 585 722 7243
e-mail: john.pytlak@kodak.com
Web site: here

The main reason when why Quadraphonic Sound faltered and Dolby Surround was so much more successful, even though Dolby actually used the Sansui VarioMatrix in their early prototypes, was because by the time Dolby was perfecting their 4 channels born out of 2, delay had become fairly inexpensive and could easily be integrated into a surround channel.

In the early 70s it was a different story -- digital delay was much to expensive to incorporate into consumer Quad products. And it is the use of time delay in Dolby Surround that is the "still point of the turning world"....that one all-important ingredient that made Dolby Surround actually sound like distinct surround. It's the aural illusion that Quadraphonic Sound simply didn't have. Without time delay, Quad, even under the best situations, was not much more than "full ambiance," a kind of circle surround rather than the ability to create locatalized rear positions (be it illusion or not).

Rather than a monumental leap into new technology, Dolby's 4 channel Cinema system was actually more of an briliant integration of long existing technologies: 1) Sansui's VarioMatrix, Type II from the Quad era 2) the application of time delay to make use of the Haas Precedence Effect and 3) Dolby's own Type A and Type B Noise Reduction that was already well established in recording studios and consumer products for a number of years, 4) the splitting of the bilateral optical stripes that had been carrying a mono channel for years on 35mm film, into dual channels and lastly, 5) third octave room equalization, pionered 8 years before by Altec Lansing in their "Theatre Voicing" system. None of it new or earthshaking, but put it all together with refinements in each, add a few format switches, a slew of CAT numbers and voilà, the optical answer to mag.

Granted, I am oversimplifying and certainly not disparaging the great contribution of the Dolby's bilateral stereo optical recording/playback system for 35mm film -- it gave us real stereo in what had much too long been a mono celuloid world -- in all the other media, stereo sound was well established in FM, consumer records and tapes, and accepted as superior by the public. Dolby finally made it cinema's turn, except that cinema had been doing it since FANTASIA, THIS IS CINERAMA, OKLAHOMA! and THE ROBE....eveyone just forgot how breathtaking it sounded back then with those discrete mag tracks.

For a kid who thought stereophonic sound was the most incredible thing he had ever heard -- magic for the ears that he couldn't get enough of, the death of 4 trk mag was a great, great disappointment. For me, Dolby Stereo never quite measured up to the same excited as much as those old 4 track mag prints.

Frank

God, how that used to drive me insane. I would be running around the theatre like a mad man during testing, yelling to the rest of the crew, "Doesn't anyone HEAR that damn tone?" (this after installing 20db attenuation super expensive notch filters on all 4 channels -- I STILL could hear it!) Those who frequented lots of Van Hallen and Zepplin concerts without ear protection only heard silence -- ah, the blessing of cochlea abuse.

Actually most of the studios other than Fox abandoned the track 4 12kHz tone quite early on -- we played many mag prints that had what we marked down on the projection report as a "clean surround track;" i.e., no trigger tone. The clean trk 4 was so prevelant that I installed a Fairchild "ducking" unit that was made for the recording studio to deal with the prints without a tone. I had to rip out the relay because it would not play the surround track if no tone were present. The Fairchild unit solved that problem.

It was a very small unit that simply used the audio itself on channel 4 to either pass the signal without any attenuation as soon as it "heard" audio of a set amplitude or it would "duck" it down by 20db when the audio dropped below a preset level. It also had a release setting that allowed it to gently reduce the channel when audio stopped. This worked perfectly for surround tracks with both trigger tones or without.

I agree that mag's time has come and gone; with the advent of digital, the issue of the Dolby analog system limited by its lack of discrete channels is gone. And yes, I will even agree that a well aligned SR system can sound very good, but that means almost always having to mix dialgue in the center and always being afraid of sounds that are by accident out of phase in the recording session winding up zipping around in the theatre to somewhere that they are not supposed to be.

There is a funny story about exhibitor execs who had watched a roadshow in 70mm and then saw the same title again in 35mm SR. They were all upset and wanted to know why the 35mm sound had a lot more music coming out of the surround channel than the 70mm presentation. It was difficult to explain that what they heard coming out of the surrounds in the 35mm optical version was simply natural phase varients that the matrix interpreted as out of phase and hense surround audio. They would rather have the errant matrixed sound than the truer discrete sound -- the old idea, "Hey, kid, I paid good money for those damn surrounds....I better damn well hear them."

My contention is that mag shouldn't be judged by the limitations of the 50s and 60s, but think of what could have been, had the refinements of more recent advancements in low noise, high output tape oxides, low-noise circuitry and high resolution head construction were applied to the mag concept. However, even if we could timeshift it all, no matter what, the window for mag was relatively short and digital has closed it pretty much for good. Forgive me if no matter, I still lament its passing along with movie palaces and mom & pop neighborhood theatres.

Frank

Matrixing of 4 channels for a two channel media (4-2-4) was invented by Peter Scheiber, a New York musician. He filed a patent (#3,632,886) titled "Quadrasonic Sound System" on Dec 29, 1969. In that patent he discloses a four channel system with a speaker directly in front, one either side, and one behind the listener using L+R and L-R (-3dB) encoding/decoding. He also disclosed a logic steering system to increase separation.

He lodged another patent in Jun 1970 which disclosed an option to swing the speaker array around by 45 degrees to produce a left front / right front /left back / right back system using the same encoding techniques.

In July 1973 he filed a third patent with improvements to the steering circuits for the original system.

Sansui and friends came along after Scheiber with the idea of a 90 degree phase difference between the rear channels. It was an idea that gave very poor separation (even with steering) and was destined for failure. However a US company Tate, did develop an analog chip set (two LSI chips) to improve separation for quad systems in general. It used a combining network, rather than gain riding for the logic steering which was a huge leap forward.

Dolby licensed their original matrix from Scheiber. Sansui had nothing to do with it. The first really successful attempt by Dolby at a matrix decoder came in the form of the Cat150 which used the Tate chip set in the steering circuits. It also used delay for the surrounds but it was NOT digital, but in fact a bucket brigade analog system. The digital version, which was initially made by a third party company (Delta Lab), came much later.

The modern Dolby pro-logic system still uses the Scheiber matrix and (a refined version of) his steering circuits, enhanced by (a refined version of) the Tate combining network. AND Dolby have managed to re-patent it at least twice. Go figure.

One improvement Dolby did make (in addition to adding surround delay) was to phase shift the surround channel by 90 degrees in the encoder, to decorrelate the rear channel from the front channels for front-to-back panning. Without that phase shift, a front to back pan would cancel out in the right channel at mid pan. So the pan would go round through the left speaker rather than down the middle.

------------------
Ray Derrick
President/Chief Engineer
Panalogic Corporation Sydney, Australia
Phone: 61 (0)2 9894 6655 Fax: 61 (0)2 9894 6935

 |  IP: Logged

 |  IP: Logged

The reason it stuck out in my mind was that we had the Sansui encoder and decoder units in our recording studio, so when I saw mention of it in the piece I was reading, I remember thinking, gee, quad makes a comeback. And I also recall that the Dynaco amp people had put out info on how to get a kind of poor man's ambiance surround channel by using the +L & +R terminal on their amps to get a L-R signal for the rear speaker. Seems like everyone had a way to get Quad at the peak of the craze -- Sony's SQ, Sansui's QS (talk about confusing the consumer) and lots of component manufacturers making receivers with decoders built in with four channels out. I recall Lafayette Radio --folks on the East Coast will remember them -- had a flyer that featured nothing but Quad gear. It was quite an exciting blip on the audiophile timeline.

I would also like to say that I respectfully disagree with Ray that the Sansui matrix had bad L/R separation. Perhaps their prototype or first generation might have, but the Type II certainly didn't. I recorded many a master tape with that unit and one of the things I was always most concerned about, being a two channel stereo recording engineer first, was that no matter how any new-fangled matrix system was going to get rear channels, it couldn't compromise my front left and right stereo channels. I did classical recordings for a Conservatory of Music of uncompromising, many times b-busting musicians -- before we moved to any new Quad system, we were all adamant that it had to leave L/R untouched; it had to keep a robust front stereo soundstage. We did lots of listening sessions with conductors, composers and musicians in the studio who were hyper- critical about the evaluations of the new quad gear. We even did a few dual recordings with two decks, one recording standard two channel stereo and one recording with the Sansui matrix. The challenge was to do an A/B test between the two channel recording and the matrixed recording but only playing the Left/Right output from the decoder, to see how left/right alone sounded. Although we did hear some slight differences, mainly in amplitude of the phantom center, Front Left/Right from the decoder maintained a well defined soundfield; it was airy and certainly didn't collapse, even when reproducing difficult program material, like solo instruments at the extreme left or right. My studio also released an LP recorded with the Sansui matrix -- an original cast recording of a musical, "Rubber Nickels." I can send you a copy and you can tell me if there isn't excellent L/R separation.

As for Dolby's modification to the matrix, well, all I can say is that in the early days of Dolby A and the first 150 cards, techies had a name for the Dolby system -- Front/Rear Stereo. Mind you, I didn't make that up -- lots of people felt that Dolby sacrificed front Left/Right separation for getting good Front/Rear separation. I never did any kind of tests to prove or disprove this perception, but it certainly was out there.

In fact, at one of the SMPTE NY Section meetings at the dawn of SR-D, Dolby demonstrated what the matrix does with difficult material. They played four different sound....a flute, a piano, a man speaking, and car motor. First each sound was played separately, each coming successively from one of the five locations -- Left, Center, Right and Rear. The matrix was able to place each in approximately the right place as long as they were sounding solo. Next they started with the left channel sound, then added the next in the center, and so one until each sound was playing in its channel simultaniously. As the sounds were added, the soundfield collapsed to what can only be described as a mono sound coming from the center of the room. Next the same sounds were played in Dolby SR-D to demonstrate that with discrete digital, the sounds stayed right where they were supposed to be. No surprise. The point is, although Dolby was all about demonstrating how impressive their discrete digital system was, it certainly made very clear the limitations of the matrix system as well -- kind of like sacrificing SR for the sake of the new Big D.

Forgive me if I can't resist saying that, hey, if you did the same demo with the 50's 4 trk mag technology, the sounds would also have stayed rock hard in each of the channels too!

Frank

quote:

---

one of the very first technical reports that I saw about the then new Dolby optical stereo system mentioned that they did use the Sansui configuration,

---

I'd always heard Sansui, too. Which is not to say they invented it.

(Hiya, Frank!)

------------------
Better Projection Pays!

For the cinema matrix, the theoretical left/right separation is the same as for a normal stereo recording and can be up to 60dB or more. Also the separation between centre and surround is of a similar magnitude provided there are no level or phase errors in the signal chain. However the separation betwen left and centre, right and centre, left and surround and right and suround is only 3dB, which is why we need a steering system.

The SQ and QS systems suffered from similar limitations. The separation between either front channel and either rear channel was 3db (or less). At least with the cinema matrix you can get great separation in two axes, whereas with SQ and QS quad it was only possible in one.

Now Gordy my friend, which Sansui phase shifting chips are you talking about? What was their part number?

quote:

---

Beginning with A Star Is Born and ending the Spring of 1979, Dolby SVA prints had surround information encoded using the Sansui QS Matrix, in addition to techniques designed to reduce crosstalk from the front speakers. The center channel had been derived by a logic circuit which analyzed the differences between the left and right track signals. Separation between channels was of the order of 6 to 8 db.

Hair and Hurricane were the first films to be mixed with a new "center channel and surround derivation card" which allows much greater separation between channels.

---

Certainly I have never come across any evidence that any QS DEcoders were used. Some people have confused the Tate chips on the Cat146 with Sansui because these chips were originally designed for use in quad decoders (for Fosgate I think). I am curious about the Cat116 card that Gordy mentioned. I have never seen one. Was this something that popped up between the Cat110 and the Cat146? (Dolby have more cat numbers than a big city alley-way).

I guess when you think about it you could say that the cinema matrix contains elements of both Scheiber's matrix and Sansui's matrix, in the sense that the Scheiber matrix has a centre channel but no use of 90 degree phase shifting for the rear, and Sansui's matrix has no centre channel but does make use of opposite 90 degree phase shifts for each of the rear channels, resulting in a nett 180 degree phase difference for a centre rear encoded signal.

------------------
Ray Derrick
President/Chief Engineer
Panalogic Corporation Sydney, Australia
Phone: 61 (0)2 9894 6655 Fax: 61 (0)2 9894 6935

The Cat. 146 was one half of the Cat. 150 system. The reticon delay board was the other half (on the first generation).

Steve

------------------
"Old projectionists never die, they just changeover!"