Part II

InPart I of Reid Woodbury’s interview with Tom Holman, SB4/90, the master sound technician discussed what is going on atSkywalker Ranch in technological innovations, and also in THX. Inthis issue, Holman [Photo1 ] continues his dialogueon the leading edge of sound theory and technology.

    TH: We’vetalked about the time displacement. We did that for a good practicalrelationship between the woofer and the horns. Hang the screen infront of it. Then the next thing you notice is if you stick your headbetween the screen and the horn, suddenly you realize where al thehighs are. They’re al bouncing back and forth between the screenand wall, causing nasty comb filtering&emdash;many delayedreflections, many hits because it’s very bright, even at thebottom of the wall. It’s obviously done many round trips betweenthe wall and the screen. So the trick is to cover the wall withabsorptive material for high frequencies, where the screen isreflective, which doesn’t absorb lows, so it provides the 2 piboundary. You can see the loudspeakers are built in, flush mounted.The absorption on the face of the wall is rather thin, only about aninch.

    If you made the wallfully absorptive you would be putting the loudspeaker in a 4 pienvironment, and we don’t want that. We want it in a 2 pienvironment. We want the fuzz on the face of the wall not to absorblows. On the other hand, we are launching quite a lot of highfrequency energy at the screen, progressively above 5kHz. Anythingreturned from the screen to the wall we want absorbed. The idea is,the wall works in an acoustical crossover mode where it reinforceslow frequencies and absorbs high frequencies.

    What is the screenmade of?

    TH: PVC or someother plastics. Maybe vinyl.

    It’sperforated?

    TH: Oh, it must beperforated. It’s usually 50 thousand perfs on a little diamondgrid. It’s seven percent open, but the percentage open areaisn’t the most important thing. They’re all about the sameopen area, but it’s the thickness and the hole size. We laterparticipated in an experiment to get an equation for the transmissionloss. One says if you have many finer holes you’d be in bettershape&emdash;smaller for the same open area.

    It does affect thehorn’s radiation pattern somewhat, but it has a tendency tosquare up the pattern. It actually makes the -3, -6, -9 contours comecloser together an interesting thing. In other words, you get lesstransmission loss off axis than on axis. The sheet, when you thinkabout it, is more transparent. You’re looking at more open area,if you look at it at an angle than if you look at it dead on.Transparency goes up at off angles.

FINALTESTING. So, you get all this together and set it up and play it.The first thing you learn is you’re smack on the curve, the Xcurve [Fig.3 from Part 1 ]. You’revery carefully tuned to it, because you employ third-octaveequalization, spatial averaging, and time averaging. A trick commonlydone in this industry, until recently, was using quarter-inch B&Kmicrophones so that you have no diffraction effects to worry aboutwith the microphones. You take a many point average for many secondseach point&emdash;very tedious. This takes 40 hours to tune fivechannels, first time out, because of the equipment andsuch.

    There are a number ofinteresting findings. One is, the room acoustics dominate theuniformity of coverage up to between 300 and 400Hz, in a 70,000 cubicfoot room. The nodal pattern dominates the uniformity ofdistribution. It’s not the loudspeaker. Above there, theloudspeaker’s the most important ingredient in sound fielduniformity (Fig. 1). That’s one finding.

    Also, to prove theloudspeaker system’s constant directivity, you measure tonebursts from the screen, and open up a gate as that tone burst passesthe microphone. You see the early arrival sound, and that earlyarrival tone burst is around l0mS long. So that l0mS includesreflections between the horn and the screen, and includes localreflections off the console, because that’s rather close to themicrophone.

    We’re trying to doit where people are listening. You get comb filtering because ofthese reflections at either end. You also find the average responsefalls off at high frequencies in exactly the same way the pink noisedoes. They track very well. There is no begging the question, whichsound field, direct or reverberant, is more important. With the powerresponse, they are the same. No need to tune it or equalize it.You’ve done it, period.

    The reason for a housecurve in large auditoriums is we’re measuring long-term pinknoise, and what we really should be measuring is the first arrivalsound. And since we’re using non-constant directivity horns, andthe horn directivity is collapsing, plus we’re measuringin a reverberant far field, we re seeing a rolloff. If we measuredthe first arrival, it would be fiat. That’s why we want anoverall rolloff.

    Well, fortunately thatdidn’t turn out to be the real reason. In the last few years,the real reason for the house curve has more to do with diffractionabout the head, and all these things acoustically going on.They’re actually there, but the fact is that we’reevaluating ears instead of a microphone.

    If you play CDs on thissystem it’s a great shock&emdash;it sounds likehi-fi.

    And you say, thisdoesn’t sound like a movie at all. What’s wrong with thissound system?

    It doesn’t honkat you?

    TH: Itdoesn’t honk! Well, that was a bit difficult for people to getused to. The combination of all these things we’ve talked abouthas the effect of lowering the bass corner by about an octave, and ofraising the treble corner by more than an octave. So suddenly thereare two octaves you never had before. It’s much more uniformaround the room.

SUBWOOFER. So that happens. Thecorner of the system is 40Hz, and since we really do want to go downto 25, or so, we use a subwoofer just for that octave. The really lowstuff.

    Is there any thoughtof going lower?

   TH: Not really. The trouble with lower ones would be thatyou’re down at frequencies where the room nodes are so strongyou can’t get it covered uniformly. Small changes in level havea lot more change in the loudness, and the room acoustics are againstyou. There are a couple of things wrong with theater subwooferdesign. First they were one-note bass design&emdash;everythingsounded exactly the same. Second, at high drive levels they wouldclatter. We bottom out the drivers. The design we picked does neitherof those things. It’s smooth and it can’t clatter becauseof special electronic limiters. In different Dolby systems there is asub-woofer output for its lowest octave.

    The low end of THX isdown 1dB at 40, and it’s two poles below that, so there’s arather rapid rolloff below 40. The subwoofer goes down to about 25.The system really stays on the X curve, which we’ve talked aboutbefore, from 25Hz to 16kHz. We’re within ±1dB of the Xcurve here. The screening room subwoofer at the ranch has two 15s: apowered Kintek subwoofer with a 500W amplifier. You can set itdifferently for two purposes. One is a bandwidth extension down to25Hz from the main channel’s low end of 40Hz. The other issimulated baby boom channels. The dubbing stage doesn’t have thebaby boom channel, no left extra or right extra so the subwoofer isset at 250Hz to simulate them.

MIXING AND THEATERS. My colleagues inHollywood say you only want a dubbing stage to be like an averagetheater. If so, I believe you never know what your program materialis. So what you want, I think, is a dubbing stage which is alaboratory. You need a lot of electronic degrading facilities to makeit more nearly like the theaters, so the mixers can tell what thecompromises are. We measured all sorts of theaters for frequencyresponse. At first, we put in an average bass filter. Later on, wegot more sophisticated and modeled specific systems, like with andwithout EQ. with and without Dolby.

    Dolby was an amazingimprovement.

    TH: Oh, yeah,because they did it without equalization. They Band-Aided a lot ofthe system problems, in effect. It’s nice they did it, but itwas a Band-Aid.

    They could go into anexisting theater with a few pieces of equipment.

    TH: They dideverything they could with a piece of electronics in the booth,that’s right, and THX tries to do everything else. We brought inpeople, first from Hollywood, then the exhibitors, to see what thissounded like, and they were amazed. The sound was so different fromwhat they were used to in a movie. Some of the Hollywood dubbingstages and some of the theaters wanted to put it in the theaters intime for Return of the Jedi. We installed four of them. Thenwe had this battle of what the noise should be, what should thereverb time be, what should this be, what should that be. I had towrite the manual and establish what the standards were going to befor the system. That’s why it’s very different from Dolbyselling a piece of equipment, because it’s licensed to meet acertain level of performance.

    Dolby has alicense?

    TH: Well, they mayhave one with respect to the use of their logos, but it’s not onthe theater. It’s no performance standard, unfortunately.They’re the first ones to say that.

    I have been in so manymisaligned theaters. The one in which I saw Amadeus had thesurrounds so loud I couldn’t hear the music over thereverberation. At another theater, you were hit as hard with the bassas with the characters on the screen.

    TH: I teach thatvery strongly in my classes. That has to do with the differencebetween “production” and “reproduction.” Peoplein the theater business are supposed to be in the reproductionbusiness. They are not supposed to be the producers of the movie.They’re remixing the movie, and causing lots ofproblems.

   We’ve started something else at Lucasfilm called TAP, theTheatre Alignment Program, to tune up theaters before the release offilms. That’s less known than THX, but it’s a veryimportant ingredient. It tunes up THX and many other kinds oftheaters to the best level possible, and gets the levels set right.Sets the subwoofer level right and gets the screen brightness right,and all these factors. That’s an important part of it,too.

    So THX became acommercial enterprise. After that first four and our dubbing stagewe’ve built, oh, gosh, how many, I don’t know. There’sabout 12 rooms at Lucasfilm with THX sound systems and projection inthem.

    What’s thesmallest room size?

    TH: The smallestpractical room size in which the whole system fits is about 20 by 30feet or so. Any smaller than that the source size gets to be animportant factor. You don’t want a discrepancy between trebleand bass, a vertical discrepancy. Then you go to smallersystems.

    Then it’s not[considered] THX?

    TH: No. We triedto match them as best we can. One of the main things I’vestudied the last few years, is how to make the loudness and thetambour match across the different room volumes.

    People wanted to be in onthe program and we had to certify products, so we were doing producttesting. The first surround loudspeaker was done by taking a BostonA70 and hanging it up in the dubbing stage and saying, OK, itmeasures like this₄ and we want it to measure like this.Can you change the network to do that? And they did. They made a newmodel, the A70-T which has the X curve built in for the averagetheater. The trouble is, the Dolby equalizers are prohibitivelyexpensive for surrounds because you have to buy an accessory rack andthe equalizer card. So surrounds are not routinely equalized. Wedesigned a loudspeaker that, at least on average, will be on thecurve.

SYSTEMLOUDNESS. We had to do studies on the loudness range of movies tosee how loud it should play without distortion [Fig2 ]. (Opens ownunpublished book) Here’s a quiet movie, TheVerdict. This is a statistical sound analysis, 16 minutes long,where the L50 in the 31.5Hz band is 50dB. That means that half thetime it’s greater than that, half the time it’s less than50dB. The average is 51. One’s the median and theother’s the mean. In Raiders of the Lost Ark, the rollingboulder scene, it goes to a hundred and five at 63Hz[maximum]. For at least one percent of the time it’shitting 105dB. You find this sort of thing, hitting 105, 106 at lowfrequencies, 1% of the time, and very low background noise levels.You see this is NC 15,whichisavery quiet studio. The levels areactually below there. I had to use an exceedingly quiet room to dothis, to capture this whole dynamic range. It rolls off at highfrequencies, but this is because this is a one-eighth second timeconstant. So transients are not very well handled by this system.Every eighth second this is valid. Basically you’d likethe rooms quite quiet.

    I know I said before thatwe got the rooms down to very low background noise level, NC 12, andwhat happened then is interesting, because we took the movie, playedit at the Northpoint Theatre about a month before Jedi cameout. There’s a scene, at the beginning of reel five, where Lukeand Leia are on the bridge and the nature of their relationship comesout, the fact that they’re brother and sister, and it’s avery quiet dialogue scene. It’s very intense. There’s alittle string, cello, music in underneath that scene. And we took itto this theater and it was gone. Totally, utterly, gone. Well, wedidn’t punch in. We must have been playing the track but wedidn’t record it. OK, we go back and listen to it. No, it ison the track. The Dolbys must be mis-tracking. Let’s gomeasure the Dolbys. We measure the Dolbys. Dolbys arefine.

    I got the bright idea tomeasure the background noise level of the theater. It was NC 45! Werecorded it in an NC 12 room. They had put that cello right in thatregion between NC 12 and NC 45, and it was thoroughly masked in thetheater. The cello overtones didn’t go up very far, and weregone.

    So we came back. I’dalready done these filters, because we knew we had to limit thefrequency range. Since then we’ve added this background noise,like an average theater, and a clipper to show optic clash like anaverage theater. So you give them X amount of power, but you punch inbuttons that show them what it’s going to do when you finallyget it to an average movie theater.

    Does the backgroundnoise include people, air conditioning and traffic?

    TH: No. That wasall air conditioning noise. I consider air conditioning noise to bethe insidious one, because it’s the one that’s constant.People say that audiences sitting still make NC 25. And I say true,on the average, but they can al hold their breath for that dramaticmoment. It’s not a good reason to make an NC 25 concert hall,just because the audience can make NC 25 noise on theaverage.

SURROUNDS. What is the Dolbyspecified frequency response of the surrounds?

    TH: The Dolbymatrix encoder has a 70Hz high-pass and a 7kHz low-pass. And thetheater system has a 7kHz low-pass. I don’t think there is ahigh-pass in the theater playback system.

    The rule is the matrixrequires as much power in the sum of the surround channels as in onescreen channel, so smaller speakers can be used. Or you can splitsurrounds, and you can say each of them should handle as muchfrequency response.

    I’d readsomewhere that the lowest frequency in the surrounds is 300 to500Hz.

    TH: No. A splitsurround format was used on very few films, on Apocalypse Now, TopGun, a few others. You see you’ve got the first stripes(he picks up clear plastic block imbedded with one 70 mm frame ofReturn of the Jedi), and the outside stripes are wider so theycontain two tracks and these two contain one track each, so there aresix tracks. In the old format they went: left, left-center, center,right-center, right, surround. When Steve Katz worked on this forStar Wars in 1977 the format was L, C, R, S. The other twochannels L-E and R-E [left-extra and right-extra],intermediate channels, were used only below 200Hz for extra basspower handling, called the “baby boom” format. Most filmsmade in baby boom format have a level stager between channels so thatthe flux on the film is lower on the LE-RE channels. It’s turnedback up in the theater. It helps the headroom both on the film and inthe theater because you’ve got more speakers going.

    Now for these few filmsthe split surround format was invented and that may be where the300Hz crossover is. In split surround you get the bass for thesurrounds off the surround channel, and you get left and right highfrequencies off the LE and RE channels. There’s a crossover inthe L-E and R-E and that’s probably at 200 or so, because thebass information still goes to Le and Re. The high frequencyinformation becomes left surround and right surround. So it’sprobably that split surround format you’re thinkingof.

   When I toured Twentieth Century’s sound department, they hadonly three surround speakers in their mixing/screening theater. Othertheaters seem to have many.

    TH: Oh,definitely.

    Is there a THXstandard for that?

    TH: Yes, there is,with THX. The theaters that have two or three are based on the 1950sidea of the “effects channel,” as it used to be called.When I saw Ben Hur at the McVickers Theatre in Chicago, about1957 or so, they had two A7s in the back corners to play the effectschannel.

    Now, there are severalproblems with that. First, it’s a point source&emdash;everybodycan localize it. Second, people who are close to it get creamed inlevel. People who are farther away don’t. So there’s no wayto set the level. Third, an artistic use of surrounds problem existsin those early films that went like this:

    In Ben Hurthere’s a kind of subjective camera shot with the cameraoverhead on a moving dolly walking behind him as he walks through abazaar. The bazaar is al across the screen and comes aroundyou into the effects channel. Well, at one point a cow moos in theeffects channel. And me, at age seven, turned around and looked atthe loudspeaker and said, “What is that cow doing loose in theauditorium?”

    This has to do with thewhole effects nature of the movie. Today, it’s much more of asurround nature. Surrounds have two kinds of artistic use. One isthe low level ambiences without specific hits to put you in thesame space. And one of the best of those is in Apocalypse Now,where they get off the boat and go into the jungle to getmangoes. The sound of the jungle is sneakily crept up behind you soyou don’t even know you’ve been pulled into the scene. Andthen a tiger jumps out! It’s like you’ve been made a partof the action. That use of surround is a very, very good one. Theother current one is the fly-over, when something “fliesoff” the screen or onto the screen.

    You feel it goover.

    TH: That’sright. Those pans really work. So those are more discrete effects,but they usually lack transients.

    They can come from allsides.

    TH: Yes. Usuallythey won’t be ticky, talky things, they’ll be whooshes.They come from an indeterminate place and they wind up on the screen.The whole point is that the screen contains the sound of the images,whereas the auditorium contains the enveloping, sounds as though youwere there. It has to do both with the sound system and artisticuse.

    We completed the THXsystem in 1982, and fired it up and it sounded wonderful. Pretty soonpeople are playing more old films and they’re saying, “Wow,this is really what movies sound like!” For me, the problemearly on was that we made these auditoriums dead, so we had a lot ofdirect sound. Yet, that’s exactly the opposite of what we wouldwant for an enveloping sound field. You would want it to benon-directional, nonlocalizable. So what do you do aboutthat?

    Add many smallloudspeakers as opposed to two or three largeloudspeakers.

    TH: So the reasonbehind using many small loudspeakers was so that you would notlocalize as much. Now you sit in the center of the dubbing stage, onthe center line, and what you get is this left, center, right, verydiscrete impression, plus this middle of the head, earphone like,impression of the surrounds, because they’re not spatialized. Ifyou move two feet to the left side of the console the surrounds areall left, because the Haas effect is so predominant. You move to theright, it’s all right. I don’t know how many times over theyears we answered the problem, “Man, the surrounds are leftheavy. Well, move to the middle.”

   We need many surround channels to promote spatialization. Or we needelectronic surround scrambling, which is something I’ve beenworking on for a while (and trying to get patented), to yieldnon-correlated signals having the same program material. A number ofways have been tried in the past: comb filtering, phase shiftnetworks, what have you. Those al cause problems, and I thinkwe’ve found a better way that doesn’t.

    Is SR actually gettinginto the consumer theaters?

    TH: Oh, yes. Theyexpect to be into, oh, I don’t know how many, a hundred maybefor big releases.

    Is the level set foreach movie?

    TH: No, al themovies are mixed for the same standardized sound pressure level. Sonow movies are much better than recording or broadcasting otherthings, because at least we’ve all agreed on the target: acertain sound pressure level. It’s 85dBC for 50% modulation ofthe optical with one channel playing 50% pink noise. The C curveis practically flat. It’s just ignoring the veryends.

    I’ve read howsome films have their dynamic range packaged to go into the home.This was before hi-fl video came out.

    TH: There is adifference between the dynamic range of a 70 mix and a 35 mix. The 70mixes are too wide for home, I think. The 35mm Dolby A mix works justfine. It goes straight across. Absolutely perfect copy if youcan get it. Clone copy it on the digital laser disk, then you canhave a crack at decoding it as well as the theater can. That’swhat I’ve been working on the last couple of years, but Ican’t talk about it yet.

    I understand thatDolby reworked the Sansui QS.

    TH: There’sbeen a lot of work done since the early QS. Initially they actuallyused QS boards. But then Dolby did their own proprietary boards forthe second generation. Now they’ve done a third. The lateststuff is actually easier to understand than the older stuff, becausethere are left-right and front-back detectors. And those detectors gointo the VCAs and the VCAs add up these voltages that come Lt and Rtto make L, C, R, and S.

    That’s one of theissues on THX. Two years ago one of the first ones was going over toCBS and hearing people overdo these crazy over-matrixed,over-dematrixed signals; where they try to take left and rightand ship it off to left and right surrounds, to make it biggerthan ever. But the picture, if anything, is smaller thanit ever was! It makes no sense at all to have the footsteps coming atyou from the left surround when you see the guy on the left side ofthe screen. It makes no sense at all.

AMP/DRIVER INTERFACE. SpeakerBuilder is just starting to address the issue of amplifierinteraction with the speaker. You’ve touched on it with thecrossover. What have you done with that?

    TH: Well, I didquite a bit of work on it. I published a paper in the AESJournal about it when I designed the Apt pre-amp and poweramp. I went out and got a whole pile of loudspeakers to find out whatthey did at different frequencies. I published the photographs of theexcursions caused by various loudspeakers.

   I showed that some loudspeakers have very strange behaviors like,pushing the voice coil out of the gap, then at the end of the travelthe impedance changes dramatically. Or an input autoformer, to changeimpedances, saturating, so you’d go to five amperes verysmoothly, but when you add one more dB of drive, you’re suddenlyat 30 amperes. We jerk from five amps to 30 amps, because the ampwent from seeing a six ohm load to seeing a tenth of an ohm load,because the transformer saturated.

    It makes designing apower amplifier a very difficult chore, because every power amplifieris designed with some kind of economic performance in mind. We’dlike to produce the most number of watts for the least dollars. Yet,you have to be able to drive, what I call, elbow room, which is to beable to drive the severity of excursions caused by the phase anglesof the load.

    I was in an amateurloudspeaker listening test recently at Just Speakers in San Franciscowhere seven hobbyists had loudspeakers built for a contest. One ofthose loudspeakers shut down these very expensive Boulder poweramplifiers, made them snap like crazy&emdash;very, very bad sound. Sowe eliminated that pair from competition, and when we measured later,they dropped to half an ohm 140Hz. Half an ohm! That wasn’t evena phase angle problem. That was just a pure and simple resistanceproblem.

    In THX, by the way, wehave qualified the front channel loudspeakers to be a certainimpedance curve, and we test every amplifier into that curve. We putup tone bursts versus frequency, as loud as possible, up to clipping,and make sure they make no protection circuitry to speak of, so weknow the amp will drive those loads. If an amplifier appears in ourlist (Table 1 ) it doesn’t say that it will drive allloudspeakers. It will, however, drive the loudspeakers we certify,because we know that the interaction is acceptable.

    THX specifiesloudspeakers and amplifiers, matched to each other?

    TH: Yes, exceptfor power level. You must pick the right power for the room size. Apower graph in the manual tells you how to pick the power.

SPEAKER CABLES. What aboutinterconnect cable?

    TH: We designedthe cable for frequency response variations. Wire gauge is mostimportant. In theaters where we have very long runs, up to 150 to 250feet, we have to get down to 12- and 10-gauge cable, respectively. Iset our standard as plus/minus one eighth of a dB ripple, due toimpedance variations. That’s probably inaudible. The only thingI believe about cable is wire gauge, and wire gauge with strangeimpedances can matter, because we can get frequency responsedifferences of even half a dB over an octave or more. Then theybecome audible. That’s where it becomes important.

    It’s nothingmystical if you can measure it.

    TH:Yes.

    Like resistance,capacitance, and inductance.

    TH: Well,capacitance and inductance are not that important because, after all,it’s not a transmission line. But, they will become problems ifthey have peculiar characteristics which cause problems for theamplifier. If the cable is built to be low inductance, and thereforehigh capacitance, it can be a capacitive load to the amplifier whichmoves the poles around in the amp’s transfer function. Whenincorrectly connected this causes trouble. It may causeoscillation.

    Shorting at highfrequencies...

    TH: Shorting it athigh frequencies, right. Such problems are certainly possible. Wegenerally use limp, many stranded 10 gauge wire so you can pull itand get good connections. That’s the main reason.

    The speakers, by the way,can’t be in physical contact with the wall. They sit onvibration absorbers, and are gasketed into the wall, air-tight, withabout a half inch rubber gasket between box and wall so they do nottransmit to the structure.

    So, you don’thave a structural radiation factor as part of theequation?

    TH: That’sright.

REFERENCE LEVEL. What’s thereference fluxivity level for 70mm films?

    TH: The referencelevel is 185 nano-Webers per meter, the standard.

    And that gives you85dB SPL?

    TH: And that makes85dB, right. The headroom is about 21dB to the MOL, but we use16 of that. In Willow, there’s a limiter, a slow bandlimiter operating at +16dB over 185. So, that’s about 3%distortion. The reason for the limiter is to make the 70 mix morelike the 35 final resolve. So, when you make the two of them youdon’t have to spend all your time pulling the 70 mix down intothe realm that 35 can handle.

    Is the 70 mixed firstand then kind of compromised for the 35?

    TH: Yes. Usuallywe make the 70 print masters first. Also, because it takes longer tomake the 70 prints.

    Those are usually madeone at a time and not just one in a batch.

   TH: One of the jobs TAP does is the reel check-outs. Somebodywill have to sit there and watch the same reel as many as 200 times,over, and over, and over, to make sure there aren’t any blotchesor drop outs or clicks.

    I heard a problem whenI saw Willow, and I don’t remember whether it was a THXtheater or not. The recorder (flute, not reel-to-reel) was playing the main theme.

    TH: Yes, it wasreal Ocarina, or something.

    It was a breathy,swishy sound as though the film was skating across the tapehead.

    TH: That’sprobably a head contact problem. You saw the 70, probably.That’s certainly one of the things that should be checked out.In fact, that’s what the 10kHz tone is for: to set azimuth andcheck level.

SYSTEM COST. How much does it costa 500-seat theater to upgrade to THX? Assume they already have a goodsound system installed.

    TH: That’s acomplete unknown be-cause of the acoustic requirements. If the placemeets the acoustic requirements it’s pretty straightforward. Ifit’s new construction they can design it in from the beginning.The difference in cost is minimal. But, if it’s an older theaterwith rattly air conditioning and high reverb time for a 70mm housethen it gets to be a lot of money. So it’s a huge range. Thebasic system Dolby included, with MPU’s, Dolby CP-200, ourelectronic crossover, eight (to 12) power amps, surroundspeakers...

    Surrounds arebi-amped, also?

    TH: No,they’re just usually wired split-surround. Wired in groups ofsome kind. The entire cost will be around $50,000 to $60,000,something in that range. Home hi-fi’s are that much(laughs).

    How much further isfilm sound going to improve?

    TH: I thinkit’s a distribution period for the available technology. Thenext revolutionary step will be digital film sound. But whether thatmakes sense, and when it makes sense, is still inquestion.

    Could DAT recordingsqueeze enough information on the existing tracks?

    TH: The linearspeed is high enough. You don’t need a rotary head. You couldreplace the mag stripes with optical bits on the film from a negativeor directly written by a laser or something. There s enough space todo it. It’s a simple matter of hardware.

    What about reading itoff?

    TH: That’snot even so hard with CDs.

    It can al be done, today.It would be a multi-million dollar development effort. But who’sgoing to pay? It would mean the upper end of the market would get alittle better. It wouldn’t be widely or quicklydisseminated.

    It’s still thespeaker.

    TH: It’s thetheater acoustics, actually.

    That’s right,you’ve taken care of the speakers!

    TH: Yes. Actually,I’d say the A4, even the A4 [Photo2 ], is not the limit inmost theaters. It’s mostly room acoustics. So THX is as muchconcentrated on room acoustics as it is on loudspeakers, andbackground noise levels. We were in a theater recently that had an[a noise level of] NC 47. What that usually specifies is highschool gyms, or light industrial manufacturing spaces. Not exactlymotion picture theaters.

PHOTO1: Tomlinson Holman, Director ofTechnical Operations, Lucasfilm, Ltd.

PHOTO2: The classic “A4” foundin most “other” theaters.

FIGURE1: Harmonic distortion of hornloaded (triangle with gray line) vs. vented box (square with blackline) at three frequencies: 40Hz, 80Hz, and 400Hz. Drawn by authorfrom data chart in THX… InstructionManual….⁴

FIGURE2: Amplifier power requirements vs.room volume. Discontinuity is where a second woofer channel is added.Redrawn by author from THX … Instruction Manual…. ⁴

PHOTO3: THX crossover boards and monitorcontrol panel. Shown is a five-channel system in a rack that allowsmonitoring each channel’s crossover and amplifieroutput.

REFERENCE

4. Holman,Tomlinson, THX Sound System Instruction Manual: Architect’sand Engineer’s Edition, Fourth Edition, Theatre Operations,a division of Lucasfilm Ltd., October 1987, p.15.