Tom Holman Interview

INSIGHTS: Interview with Tom Holman
"Audio Guru" and president of TMH Corporation

Interviewed by Mel Lambert in August 1996

It is no idle exaggeration to state that Tomlinson Holman's contribution to our understanding of theoretical and practical audio has had a profound impact on both creators and consumers alike. His distinguished career in audio, video and film now spans over 30 years; his influence has been felt by virtually all sectors of the motion-picture production and exhibition markets, in addition to home playback systems. As a theoretician well grounded in the practical realities of designing and building professional and consumer playback systems, Holman has the unique and rare ability to push the state of the art and, in doing so, challenge entire industries to achieve new quality standards. The legacy of his career is the creation of new markets, and the redefining of existing markets, by establishing new quality standards and designing products which perform to them.
All THX patents come directly from Tom Holman's experiments and research. Serving for 15 years as corporate technical director for Lucasfilm, Ltd., he spearheaded the conception, development, design, and implementation of the technical infrastructure for George Lucas' Skywalker Ranch, including the Skywalker Sound post-production complex. From 1987 thru 1992, he also designed many aspects of the Hollywood Bowl's sound system, including electronic enhancements
Tom Holman has also been teaching film sound at the USC School of Cinema since 1987, and is currently an Associate Professor. In late 1995 he resigned his position with Lucasfilm, to set up and head a new company, TMH Corporation, to research and develop new professional-audio and consumer-electronic systems. We caught up with this audio communicator during a rare break from his teaching at USC and ongoing research with TMH Corp.

Can it really be true that you've been involved with the audio industry for over three decades?
I started working full time as a professional the moment I got out of college; I have a degree in broadcasting and communications from the University of Illinois. I started in EE, and shifted midway through undergraduate school to broadcasting. Engineering, at that time, was a very narrowly focused view of the world. I wanted to work at the kind of things I'm doing now- engineering for entertainment; for the arts- but everyone around me in engineering school seemed to be on the straight and narrow for the corporate path, and that was something that didn't interest me.

I get the impression that you are more of a free-ranging spirit, and need to personally enjoy all that you become involved with. You combine inquisitiveness with the desire to communicate your discoveries. Is that a fair assessment?
Well, yes, I do enjoy being a university professor. The biggest problem with that job is that I came to it late in my career. You've so much shared knowledge and belief structures with your colleagues, and suddenly you have these students who today aren't even educated in the scientific method. Finding ways to distill [the information] down to the student level is quite difficult.
During my career I've been in some very different areas; while they have all been related to audio, they've ranged from film-sound production [sound mixer at the University of Illinois Motion Picture Production Center] to high-fidelity equipment, and back to post production and sound systems for theaters. Even though it is eclectic, the one thing that links it together more than anything else is inquiry. Once it becomes a boring job, I move on.

Without your focused attention to both technology and the creative aspects, many developments wouldn't have taken place. You seem to bring an academic structure to your deliberations, but with an innate fascination with the applications of advanced technology to audio?
One thing that links all this has been knowledge that already exists; I go to technical libraries frequently. It isn't so much that you have to be a genius- because I don't think that I am- just that you do need to know what the developments are in a variety of fields. While designing a sound system, for example, you're dealing with a lot of different aspects, from the mechanics of how it works, to the psycho-acoustics of how it's perceived. Experts get to be experts in fairly narrow fields; they don't have the time to look at what's been done in parallel fields, or ones that might be relevant.

It's great fun to cross all of these borders and have a competent conversation with the world's leading people in psycho-acoustics or loudspeaker design. I'm not a real expert in any of these things; I'm just knowledgeable enough to talk to people in all those areas.

You spent some time at Advent, designing loudspeakers and amplifiers?
I came aboard in 1973 as an engineer; the chief engineer left about six months later and I got his job. The nice thing about Advent is that at that time I could take it as far as I could. The company was already successful in the video market; therefore the audio electronics and loudspeakers fell to me and others, including Andy Kotsatos, who is now president of Boston Acoustics. Those of us who worked at Advent often reminisce about the experience; the concentration of intelligence was amazing for a few years and crossed so many boundaries, from sales and marketing to music and engineering- a real amalgamation of talents.

I started in the film industry [as a mixer] and went into hi-fi because of the miserable state of sound. I would have stayed if the soundtrack hadn't wound up on mono Academy optical.

So the sidetrack to Advent was to learn more about electronics and loudspeaker design?
I wanted to extend what I knew. The reason they liked me over other candidates was my hands-on approach; I beat out MIT engineering graduates because I could built systems that worked. At the University of Illinois I did everything: production recording, editing, mixing for film documentaries. I stayed at school five years after graduation- that's the time I got to read everything in the library. Today people automatically become specialized, because the fields grow; people who are in the nitty-gritty of digital audio, for example, are unlikely to be as expert in psycho-acoustics, which is unfortunate.

So you had a broad foundation. How did the move to Lucasfilm come about?
Well, they looked for six months for someone to be the chief engineer; the contact was ultimately made through Dolby Laboratories, who were using my pre-amps and power amps at their lab. I was with Advent for four years, and left [to found] Apt Corporation, a company making pre-amplifiers and power amps. Apt got up to about 50 employees, and had about 30% market share in pre-amps. But the high-end had a "Pre-amp of the Month" syndrome, which was that certain units would be blessed for a certain time and then collapse. While ours went on for eight years and are still in service today, its was a very tough field.

Apt was a 50/50 partnership [Holman served as director of engineering], which made it very difficult to leave, but I decided to do it because the opportunity at Lucasfilm was so great. Here was the first movie studio to be built from the ground up since the Thirties. We had a music-scoring stage, mix-to-picture, sound editorial- a typical full-function film post-production facility.

Starting in 1980, I got the opportunity to examine the whole film-making process. Should we buy dubbers? Or should we buy multitrack digital machines? What technology should we get into? All of this was in parallel to a very significant project that eventually became DroidWorks; ultimately, it transitioned outside the company to make digital audio work stations. This was George's dream in 1980, to make films digitally, save money and increase quality.

In theory, that meant that post-production, editorial and mixing was going to be all-digital; analog was only meant to be a stop-gap measure. When we looked at film-dubbing consoles, they weren't very good audio quality. Then we looked at music-industry consoles, and found that those built in serial production in relatively larger numbers were much better quality. So we took a Neve quad console, and reconfigured it for film-sound mixing by designing a panner- we put over 3,000 man hours into that console with multichannel panners and the monitor matrix you need for film sound, which today are off-the-shelf items.

That year and a half that gave me the ability to look at everything in the production and theatrical chain; to clean up parts of it, to adopt others and, in the case of the theater loudspeaker, to start over. At that time Dolby Stereo had made such an obvious advance in film-sound quality that the loudspeakers and room acoustics of movie theaters needed attention. THX was the result of attempting to improve the overall playback quality of film sound, given the Dolby advances.

Did you continue to work with Dolby on the evolution of its 4-track 35 mm systems and 6-track, 70 mm formats?
To a certain extent. What we did on "Return of the Jedi" was change the magnetic oxide. When we looked at the magnetic oxide they were striping on 70 mm prints in 1983, it was equivalent to Scotch 102 from the early Fifties! The problem was there was only one vendor, but a second vendor was interested and had bought up the old MGM striping plant. We went with them because they were striping a much better oxide. The interesting thing was that the difference you got in headroom was 11 or 12 dB- a major improvement.

"Jedi" was pretty good sounding on 70 mm, but there were still a few orchestral passages where you could hear IM. What was happening was that the accumulated distortion over all those [dubbing] generations had suddenly become audible; it wasn't audible on the [print] master but became obvious on the releases print- it was just one generation too many. By the next year we pushed 3M into making much better mag-film stock for the post-production generations. When we had got to "Indiana Jones and the Temple of Doom," where was an incredible passage for IM where the bass drum is banging with choral music above it- the [new formulation] passed the test, and it did not intermodulate audibly.

The biggest problem in the chain was the theater loudspeaker. The [Altec] A4 was developed in 1947, and was very good for its era. But by the Seventies it was way out of date. It was a the chicken-and-egg situation, with Hollywood using it because it had 80% market share in theaters. I set out to design a more compact, better sounding systems.

A large amount of what makes [Lucasfilm's THX System] work was researched in the library by patching together various developments- spotting the right ones, discarding the wrong ones, and stitching it all together into a comprehensive system. THX's stock in trade was to get the room acoustics right, and then apply a good sound system. Then you make translation from the dubbing stage into the theater sound far better.

In other words, get the room sounding right, and then put in a system that will work correctly in that environment?
Right. There were some obvious observations. I took home a copy of "Star wars" and listened on a wide-range system that went down really low. There were rumbles on the soundtrack that would cut in and out from one shot to the next- the mixers would never have left that there if they'd heard it at the dubbing stage! So the first THX design was merely an attempt to build a system for the dubbing stage.

Was there an attempt to keep the design relatively simply? It's a pretty basic, two-way system, with a patented active crossover design?
Yes, it's all in the two-way active crossover. As a matter of fact, I did look at three-way designs rather extensively- the later-generation Apogee Sound design for smaller rooms uses a three-way configuration- but a two-way was serviceable. I couldn't figure out a way to get the three-way to work as well- I couldn't get a match in amplitude, phase and directivity plus timing in that environment. Given the screen and its action, the top octave response of modern compression drivers is ragged but is there. It would be nice to have some device that worked as a piston at those frequencies- you'd get more consistent results- but I still don't know of one.

But isn't the key aspect of the THX design that it uses both electronic and acoustic functions at the crossover point? Using the propagation properties of the LF drivers and the 90-by-40-degree MF/HF horn to dramatically smooth the transition between the two component's frequency ranges?
It's not really a new idea; Don Keele wrote about it in an AES pre-print, and it was used in the standard JBL studio monitor with a bi-radial horn. JBL built that horn to match the directivity of the 15-inch driver crossover. The directivity index increases the amount of the direct sound to the user- which helps us localize sounds- compared to the reflected sound. What people don't realize about film-sound systems is that we are our own worst enemies in terms of intelligibility! We put up all this interfering music and effects in front of the dialog, which gets buried. If you don't deliver it with a sufficient direct-to-reflected ratio, and an adequately low background-noise level, you are going to loose it in the mix.

It's not only a directivity match, but the phase alignment at the crossover point. To be in-phase at the crossover, Linkwitz-Riley filters depend on the acoustical response and not just the electrical response. The system's electrical response is buttered up in order to make it come out such that acoustically it is fourth-order, which means tailoring the crossover to the exact drivers. And that's what hardly anyone does.

You have also emphasized the importance of using a hard defined center channel while mixing for film and video; and ensuring that three identical, front-channel loudspeaker systems are used for domestic playback.
Of course that's not a new concept; it started in 1932. The biggest problem is that even though center-channel loud speakers are selling in large numbers for the home now, people think it's the dialog channel- that it has different requirements than the others. The truth is that it should be perfectly matched to the other channels, because sounds move. What most consumers don't understand is that the music elements are made three-channel, and ought to be reproduced via a three-channel systems.

You also make a good case for using more than the traditional four LCRS channels for playback.
And of course that's what was found in the Thirties, where three front-oriented channels were used for orchestral music playing in concert halls; interestingly they really had a surround channel. If you look at the drawings, they used microphones placed close to the close to the orchestra, and picked up relatively little hall sound. (Today, if we were doing a two-channel recording there would be mikes capturing the hall sounds.)

What they did was to play it back in another hall. They now had the direct sound of the orchestra picked up from the original hall and created the reverberation in the receiving hall. This was so much smarter in 1932 than what we do today to trick the consumer into thinking they need added reflections in their playback equipment using DSP algorithms, which is foolish.

You're no longer associated with Lucasfilm. Tell me about your new focus these days. How did TMH Corporation come about?
I was an employee of Lucasfilm until 1987, and then a consultant until a couple of years ago when my company became a contractor. The employment and consulting was exclusive to Lucasfilm until about a year and a half ago. But it was a pretty tough life commuting every week from Marin to Southern California; being down here [at the USC Campus] three to four days a week and up there for many years got to be quite a drag.

What was behind the foundation of TMH? To give you a vehicle to explore new professional and consumer electronic designs?
I felt it was constraining the have to justify everything to George's executives and so forth. Now I am working in a number of fields, including multi-channel audio. I'm making professional products that probably wouldn't offer big enough markets to interest companies like Lucasfilm- one thing we're making is a multi-channel panner. And there are other products I plan to make in that series which will be relatively simple and inexpensive. People have already asked me for a six-channel fader.

Are you going to be handling manufacturing of these products?
While we won't physically manufacture them, we will certainly QC and pack them. The THX crossover is made from a kit of parts that are QC'd from Lucasfilm; an outside firm stuffs the boards, wave-solders the components, and then Lucasfilm tests them. TMH would also do that. If I'm going to put my name on it I'm going to have to test it myself.

You have two other partners in TMH Corporation: Fritz Koenig and Ross Hering. What is their particular expertise?
Fritz and I have been working together since 1980. He ran Apt in my absence and then came to work at Lucasfilm in the Theater Alignment Program [TAP], where he became the database expert. We were using big UNIX database to nail down the parameters of theaters. Ross was the third employee of TAP, and later ran the laser disc program at Lucasfilm; he has good contacts in the studios and is well respected for his managerial skills.

One of your major projects is the Microtheater monitoring system. How did that come about?
The Microtheater desktop-based dubbing project is a monitor system that allows true film mixing in a DAW environment. C41, as it is called, will let you make competent temp dubs, pre-mixes, print masters, etc. In this way, facilities can do a lot of things that traditionally require dubbing stages; time required on the stage can be spent instead in editing rooms. We have some tricks that get rid of the room-acoustics problems of small rooms, and how to scale the sound from one environment to another.

Today in Hollywood it's largely perceived that digital audio work stations are for editorial work and not for mixing. But, as these DAWs become more sophisticated, with control surfaces and more automation, it's clear that they are capable of doing more jobs- especially when they have a multi-channel panner. So the direction we're headed in is for the sound designer editor/mixer to produce more finished goods in the editing room, and being able to do that [simultaneously] in multiple rooms.

Because of the nature of the system in smaller rooms, you have to employ a phantom center. You asked me earlier about a dedicated center channel. I've come full circle; now I have a phantom center. If you're doing 4-channel work, C41 uses left and right satellites, a subwoofer and one surround; two surrounds if you're handling five-channel dubs. It's currently mostly analog electronics, although there is some digital in it. We intend to incorporate more DSP as time goes by.

One of the jobs it's doing is lengthening films for television, by extending ambiance and dialog, for instance. A [Studer Editech] Dyaxis workstation is used in that case to edit the transitions and perform scene extensions; getting everything to match across the scene extensions is quite a trick. C41 is the monitoring system. We also have a system working with a [Digidesign] Pro Tools five-channel system with multi-channel panning; from what we have heard this week, [the results] are translating well to the dubbing stages.

The C41 Microtheater monitoring system will be leased; it comes complete with everything you need, and TMH will handle final tuning and adjustment of the playback environment. Obviously, we can't fix background noise and other big problems- like the editor next door with a louder sound system than you- but we do fix the standing waves and so forth in small-room acoustics. We have very tight control over its installation because we're supplying the hardware. There are various application areas for C41: post production, DDD Compact Disc and CD-ROM mastering, for example.

One of my problems might be convincing accountants and so forth that these far-out ideas I have are practical. We have a way around that. The National Science Foundation has just made USC The Engineering Research Center for Multimedia- we beat out 140 other schools. It's about $12 million of Federal Government money, which has opened the flood gates on donations from industry totaling about $35 million for the first 5 years through USC. So we can do the more far-out research in cooperation with the USC engineering faculty.

You and John Eargle recently organized IAMM '96, "International Alliance for Multi-Channel Music," a two-day colloquium that looked the basics of multichannel audio, hardware requirements for studio and consumer users. At the time, I quoted you as saying that "The challenge for the music and recording industries is to define the best audio-only uses for such emerging technology [as DVD], and to prepare the infrastructure required to service and expand the existing market for multi-channel music."
There was some time pressure to get something to happen, because the theory was that people in smoke-filled back rooms were going to make a decision on the audio-format Digital Versatile Disc. We wanted to be certain that all opinions were aired and given an overview, including the use of 5.1-channel formats with split surrounds, and more.

I may be seen as tilting at windmills when I ask for more than 5.1 channels on a new Audio-DVD format, but really I'm not. What I'm trying to do there is to produce some kind of future-proofing of this disc medium. There are three things that determine the bit rate on an audio disc: sample rate, word length and number of channels.

While we could go on all day about sample rate and word length, the truth of it is that they will make a very small difference to most people compared to what 16-bit and 44.1 kHz can achieve if done properly. What will make a bigger difference than anything else is the number of audio channels; how many we have and how they're disposed. As I have pointed out in various articles, three or five up-front channels and split surrounds can achieve wonderful results for music, while overhead channels- where available- can add enhanced realism.

From a psycho-acoustic point of view- that your perception is better in some parts of space than others- you should put more channels in the front hemisphere than in back. (Not that you don't want them in the back; it's just more important in the front sector because you have better resolution in that sector.) There's a lot of evidence which says that more channels would be nice- dating back from the work done at Bell Labs in the Thirties on up through [the late] Michael Gerzon's work. But vinyl, FM broadcasting and the Compact Disc only offer two channels, so we have become used to two-channel stereophony. But now there are maybe 20 million home theaters around the world that can play back at least four-channel [Dolby ProLogic] soundtracks; and we now have a greater potential of using such 5.1-capable systems with audio-DVD.

How would you categorize the improvement in replay quality moving from two-channel stereo to "5.1" formats?
It is plainly audible to the casual observer. Most people who get a surround-sound system tend to exaggerate the surrounds- we are in the "ping-pong" era. Yet when it's done right, the [surrounds] are probably unnoticeable a good deal of the time.

As you add more channels, everyone can hear a difference from one to two; no question. Under the right comparisons, everyone can hear the transition from two to five but I don't know where we reach the point of diminishing returns. It's clearly above five channels. In terms of sample rate and word length, increases here only satisfy a vanishingly small part of the audience. Going from 48 kHz to 96 kHz is so infinitesimal that it's simply not worth the bits, compared to adding more channels.

And what about enhanced bit resolution, from 16- to maybe 20- or 24-bit?
Well, the difficulty with that is that we need extra bit capacity for the recording and mastering medium, compared to the release medium. In the early days, we had a two-channel, 16-bit Soundstream recorder and a two-channel, 16-bit releases; no problem. The first time you use a 16-bit multitrack recorder you've got a problem [because] you can't have a 16-bit result if you add channels together. The professional always has to have greater word length than the release format. So it's kind of crazy to jump the release format to 24-bits because you'd have to increase the professional requirement very much greater than that.

Recent papers have shown that 16 bits are inadequate because the noise level is audible in the frequency range around 2 kHz, where we are most sensitive. When the replay level for zero dB FS is 120 dB SPL, okay. But how many people play their systems aligned so that 0 dB FS = 120 dB SPL? So if you turn it down from there and you have 100 dB- those are pretty loud levels. Not levels you use in studios but levels civilians use. Then that 16-bit noise is already below minimum audible threshold.

I think there will be a mass market in five-channel material simply because you've got millions of people asking: "How do I light up these other loudspeakers?" And not being satisfied necessarily with the results they're getting now.

What would you conclude from March's IAMM '96 conference? Was it timely?
Oh, absolutely. The interest level was higher than we thought. We really struck pay dirt in the idea of it not being an engineering conference. We had people from many countries and from parallel industries, but they were all talking about multi-channel formats. George Massenburg [engineer/producer and president of GML Laboratories] was one of our advisors, so he got comp'ed. But he sent us a check afterwards because he said it was one the of best events he'd ever been to! Probably the most important session was the business discussion, including representatives from [record] labels. The sessions was scheduled to go 90 minutes, take a lunch break and then another 90 minutes- in the end, they went for six hours straight and had their lunch brought in!

Do you plan to organize another one?
Yes, we are currently formalizing that. We think it will bracket the AES Conference; we're not trying to compete with AES in any way, but make it easy for foreign guests. One of the most important things we did were multi-channel demos with multiple replay formats. John [Eargle] and I also co-chair the AES Subcommittee looking at new multichannel DVD, amongst other considerations. We had another big meeting at the recent AES Convention [in Copenhagen], at which the European contingent raised a lot of [reactions] to the proposal that had come out of the L.A. group. The fact that DVD is behind schedule is actually very helpful, since the pressure is not quite as much as it was. It will be several years before we'll see an audio-only disc.

How will today's audio graduates fare in the future world? What sort of skills are they going to need? Are we training them in the right direction?
They're much better trained than they were in my day, because there wasn't any training in my day. I left engineering in my junior year, so I spent nearly six years doing the kind of training that today you might be able to get in a concentrated course at, for example, The University of Miami, because it's focused. Our film graduates [from USC's School of Cinema-Television] are very different because they are going to film makers.

It is a very complex environment, and we do things because we think we can make a contribution. We think that we have expertise in the area to do it, and so there are many factors involved. I want to do it because I enjoy it, because I find it interesting, and because I will find out something I didn't know before.

Tom Holman's Nineties Manifestation: TMH Corporation
Tom Holman's latest focus is TMH Corporation, a start-up technology firm that creates, manufactures, and licenses products for the entertainment industry and related consumer markets. "A core asset of the new company," Holman offers, "is a comprehensive system-design philosophy that links artists' original conception with its final presentation. Initial products are targeted for producers of entertainment software, whose needs for sophisticated and refined production tools is ever increasing."
   TMH Corp. will examine market opportunities created by the converging cinema, computer and communications industries. "We are actively developing multi-channel sound products for a variety of professional and consumer applications. High-definition video systems are also under investigation. In addition, TMH consults for digital production studios, exhibition, consumer electronics companies, and high-tech companies wishing to apply digital technology to entertainment."
   Offices are located in Marin County, CA, and University of Southern California, Los Angeles, TMH was the first partner of USC's Egg Company 2 (EC 2), a multimedia laboratory recently opened by the University's Annenberg Center for Communication. EC 2 is designed to incubate digital business development and to assist high-tech entrepreneurs with taking their products out of the lab and into the marketplace.
   Company founders include Friederich (Fritz) Koenig, who first worked with Holman in 1980 as GM of Apt Corporation; the duo were instrumental in Lucasfilm's Theater Alignment Program. S. Ross Hering is the former director of business development for Lucasfilm's THX Division, and worked closely with Holman while expanding TAP and in developing the THX Laser Disc Program.