Bob Chidlaw

Interview by Brian Cowell.

Back in the late 1980's, unbeknown to its creators, a keyboard dynasty was about to begin. A method of synthesis that still, till this day, is unfathomable in its depth. Featuring a multitude of methods that could process sounds in many unimaginable ways, it is known on Kurzweil machines as VAST.

As for descriptive acronyms, VAST pretty much describes the heart of the Kurzweil operating system. One of the people who were the key components of the VAST architecture is long time employee Bob Chidlaw.

Sonikmatter got a chance to catch up with Bob while he is busy finishing up on the new KSP-8 effects module Kurzweil are about to release, here is what he had to say:


SONIK : What is your job title at Kurzweil and what do you do there?

BOB : My job title is Chief Scientist. I am also the head of the DSP department. I work on architecture and algorithms for synthesis and signal processing. This applies to both our custom DSP chips, and the way they are used inside the products.

Having started in Dec 1982, I have worked on pretty much all the products we have ever done, in addition to having been the entire software, soundware, and DSP departments for a fortunately very brief while back in the beginning.

SONIK : Is this the period of time when Young & Chang bought the original bankrupt Kurzweil?

BOB : Oh, no, this was back in the first few months of 1983.

SONIK : I understand there were very few left in the changeover from the original Kurzweil to the Young & Chang Kurzweil ; How did you feel when this changeover occurred?

BOB : Everyone in engineering was quite relieved when we found we would be able to carry on with our work. We had been in bankruptcy for several months when YC acquired us. And for many months before that, it was obvious that we were going down. Hal Chamberlin maintained a graph of our stock price on the wall outside his office. He eventually had to adopt a log scale as the price went down to 1/256 (from a high of around 7.)

SONIK : Would you consider it to be something better for Kurzweil (Y&C) in the long run?

BOB : Yes, we certainly have sold a lot more products under YC than with the original company. The original company suffered from really bad management; if anyone good came in, Ray Kurzweil would hypnotize them with his grandiose visions of total industry domination, and they would start spending money like drunken sailors.

Except for the CFO who held up purchase of chip design tools until our one chip designer left.

SONIK : Where did you work before becoming involved with Kurzweil?

BOB : My previous job was at Eikonix Corporation, which did work in photographic and optical systems, mostly for the government. I was there 9 years, which I thought was some kind of record, but I am now at over 18 at Kurzweil. I learned a lot at Eikonix. Sort of on-the-job training. Fourier transforms and digital filters. The underlying mathematics is very similar for image processing and audio processing. Whereas one samples audio signals in time, pictures are sampled in space. Except that you have two space dimensions, which makes all the DSP a lot harder. So going down to just one (time) dimension sure seemed a lot easier.

My last two years at Eikonix, I got to work on a system for making color separations, used to make the printing plates for four color printing. It scanned in color slides and wrote output on a home-brew laser writer. It permitted adjustments of color, sharpness, cropping, etc. (This cost $500,000 back in 1980!). I came up with some cool algorithms that made the whole thing possible. That was quite exciting, to be able to do some original work that related to a real product. I got a lot of self-confidence out of that. I was sorry to leave the project, but the chance to do something with synthesizers was too much to pass up.

SONIK : How did you become employed by Kurzweil?

BOB : The company sent off a mailing to subscribers of Computer Music Journal, announcing that a new company was being formed in the synthesizer business, and I sent in my resume.

After two interviews, and quite a lengthy delay, I was hired.

I had built a lot of signal processing boxes (some fuzzes, an envelope following filter, a six stage phasor, a flanger), a modular analog synth (which was the first synth I ever played), and a 4 voice polyphonic. The synths used SSM and Curtis chips. The poly keyboard was scanned with a circuit of about 30 digital logic chips. I had published some articles in Electronotes (which was Bernie Hutchins' project for people to share information on synthesizer design.) I had a lot of software and DSP experience at my current job, and was a versatile amateur musician. I even managed to look like I knew what I was doing fooling around with an Arp 2600 during the interview, although I had never even been close to one before. Never did figure out why they took so long to hire me!

SONIK : What is the level of your education?

BOB : I graduated from MIT in 71, with a physics degree. Sometimes I wonder why I didn't do EE , but I have always been very interested in physics. After two years, one is required to choose a major, and I had already taken a lot of physics classes. And everyone doing EE seemed to be really hung up over the course in Linear Systems, finding it incomprehensible. Perhaps it was just the MIT approach.

Of course, Linear Systems is what I now specialize in, but I have had no formal training in it, or indeed in most of what I do. I'm not recommending this, by the way. I then spent a year at Northeastern University, as a physics grad student. Eventually I just got tired of the same old problem sets, and left to get a job programming (my only marketable skill.)

SONIK : Who works in your team at Kurzweil?

BOB : Chris Perkins is the only one left in the DSP department, although we will have a new guy starting from England as soon as his visa comes through. Chris has written a lot of the KDFX algs, and also a microcode compiler for our effects chip (Lisa), and the linker and read-time control software that runs in the product. KDFX wouldn't exist without him.

SONIK : What aspect of the K2-series do you look back on feel like "you've done the impossible"?

BOB : The most impossible thing was to get YC interested in doing high-end synths, when they didn't really even know what they were. They were an acoustic piano company, making some lame home digital pianos (with Gulbransen technology) when they bought us. They were mostly interested in the home digital piano market.

SONIK : You wrote the original FM Algorithm that wasn't used in the K2-series?

BOB : Oh, yeah, I wrote an FM alg. It was really great sounding, being able to use samples as modulators. A lot of the sample characteristics would come through, but interestingly transformed. A pity we couldn't use it.

SONIK : Were there any other Algorithms that didn't make it to the original K2000?

BOB : There weren't very many other alg's that we didn't put in. For various architectural reasons, there had to be just one set of alg's, and we had to get it finalized so that soundware could do the presets.

SONIK : What DSP functions have been the hardest to get working properly?

BOB : The two pole resonant low pass filter was definitely the hardest. It took several months work spread over two years before I was satisfied. Sorry I can't describe the nature of the difficulties or the solutions. Or maybe it was the reverb algorithms in KDFX. We have worked on reverb for over ten years, with quite a number of people participating at various times. The evaluation of reverb quality has to be done by ear, which makes it all the harder. We are still working on new reverb Algorithm's ; currently surround Algorithm's for the KSP-8.

SONIK : Do you ever design a DSP and find it has a side effect that becomes a feature?

BOB : The best accidental discovery was made by Chris Perkins, who works for me in the DSP Department, who discovered LaserVerb (a KDFX Algorithm) while performing experiments relating to reverb. Pitcher (another KDFX Algorithm) was also a bit of a surprise to me, even though I wrote it.

Some of the stranger things that can be done with Live Mode have astonished everyone, including John Teele. Who pushed for us to do it, and wrote all the code necessary. Still, there really aren't a lot of accidents that go into the products; it's mostly pretty clear what's going to happen before we start.

SONIK : Is it possible to put new DSP functions into VAST?

BOB : It is theoretically possible to write more DSP functions and we did for KB-3 mode and the vocoder. But we had to chose the DSP functions for the initial release of the K2000. Due to the architecture of our chips, if we wanted to add another function, an existing one would have to come out. (Yeah, I'm still kicking myself over that architectural decision, but, hey, we started the design in 1988.)

And then we're faced with the problem of breaking existing programs that use the old, removed function. This just seems like quite a nightmare, and although the users that just do one track at a time wouldn't mind selecting and loading the precise set of DSP functions they need for a single, particular program, I think it would really annoy a lot of people. There would be programs floating around that wouldn't work together. KB-3 and the vocoder seemed cool enough to break this rule. But I don't think it's likely we'll do it again.

SONIK : Do you ever run out of ideas for things to put in the machines?

BOB : We generally have far more ideas for products and features than we could ever hope to do. People write memoes, and we have meetings (sometimes fairly heated), and eventually form a consensus of what things we should do.

SONIK : Do you think that the original K2000 was a surprise with its popularity?

BOB : We all knew it was really great, but I don't think we ever managed to communicate that in any of our ads. If someone got a one hour demo from someone who knew the machine well, they would just have to buy one. So I think we saw exponential growth of sales, starting extremely slowly, but picking up once people got to see one themselves, either at a friend's, or at a studio.

There is a strong fashion element to the music business; some things become popular for no particular reason. But a good sounding, flexible product will have appeal to a good portion of the marketplace.

SONIK : What was the in-house project name for the K2500?

BOB : The working name for the K2500 was the K2048. I really liked it, but marketing didn't. (A 48 voice version of the K2000. And 2048 is a power of two. What could be cleverer?)

The Micropiano was also called the Nanopiano during part of its development. (This was before the Nanopiano by Alesis (was it Alesis?) came out.) I liked the name, marketing hated it. I just stay away from all the naming stuff now.

Although I did manage to get Dave Fox and his wife Lucy calling their unborn child "Sly" for a while. But it got changed to "Zachary" before the shipping date.

SONIK : When the K2000 was released, people spent years programming and learning VAST.Do you think that is happening all over again with the introduction of TMP in the K2600?

BOB : It certainly takes a while to learn how to use VAST, and to get an idea of the range of sounds it can produce. If you listen to our original K2000 presets, and compare them to the latest 2500 presets, you can hear that we got a lot better at using VAST. (This is a hard experiment to do, since there probably aren't many 2000's out there that haven't been upgraded.) TMP offers so many possibilities beyond VAST that it's hard to imagine anyone ever exploring more than a tiny fraction of it. Still, if one plays around with TMP long enough, some vague feeling for what is likely to happen for different wiring diagrams and different DSP functions will become apparent, which should be (at least slightly) useful when trying to create a specific sort of sound.

It's more fun to just play around quasi-randomly, of course, to see what kind of sonic mayhem you can produce, as long as you're not trying to find the right sound for some mellow ballad. I am seldom surprised by anything done in VAST these days, but TMP is refreshingly unpredictable.

I have recently gotten to hear a set of TMP programs made by Jens Dibowski for us. They are totally unbelievable. He has managed to achieve amazing responsiveness in a number of different sounds, that all have a very interesting complexity to them, reminiscent of the complexity of acoustic intruments. These are sounds which are useful for conventional music, i.e., not just weird noises. We will be distributing these; I'm not sure how. Everyone must check them out. Which raises the question, what happens in future products (by us or some other company) that have even more possibilities? I can imagine people working for weeks on just one program, tweaking the sound and all the modulations, and learning how to play it. But then does one have to be an instrument designer and builder, when one started off wanting to be a performer and/or a composer? Beethoven didn't build his own pianos, let alone design them. And I don't believe Christofori ever wrote any music.

When someone is really fully utilizing the sound design ability of one of our synthesizers, then I become not an instrument designer, but rather the guy who made saws and chisels for Christofori. Which is OK by me; I love to see people really using all the capabilities of the instrument. (Perhaps I should mention that Christofori built the first modern piano.)

SONIK : What is like being part of building the best synthesizer out right now?

BOB : It's a really interesting job, of course. Much better than, say, designing tele-communications routers. There are a lot of challenging problems to work on, with a nice range of difficulties. And there is no end in sight. And yet, because the market is so small, the field isn't crowded with thousands of researchers, so one has a chance to be the first do something. The downside to the small market is that we're never going to become filthy rich. Certainly there are frustrating aspects. We never really have enough resources to do everything it seems we should be doing. There's a lot of competition out there. And everything takes longer to finish than expected, even when you started out expecting it to take longer than you expected. Sometimes when some new instrument or feature starts working, and you hear something really cool, that you know no one has ever heard before, then it's all worth it.

Currently the KSP-8 is providing that kick; a couple nights ago I heard it do a several minute long piece of electronic music when fed a single snare hit. If a different sound was put in, it would do a different piece.

SONIK : Can you tell us what instruments you have in your studio at home?

BOB : Well, there's an awful lot of guitars, basses, mandolins, banjos, fiddles, and lap steels. Favorites are a 67 Gibson SG, a 93 Taylor acoustic, and a cheap 90's Japanese-made Fender Jazz fretless. For synthesizers, naturally a lot of Kurzweil gear. In addition to my home-brew synths, I have a CS-80, a Prophet-10, and a Cat. Two Hammond organs, and parts from a third. A Hohner Pianet, a Fender Rhodes. And too many microphones.

SONIK : What do you get up to outside of work Bob?

BOB : I have a collection of vacuum tube guitar amps, around 60 I guess. Most of them have been restored, although I do have a bit of a backlog! The oldest ones go back to the late 30's. I read a lot. A lot of science books, sordid biographies of dead musicians, some science fiction. Used to read a lot of science fiction, but I'm getting bored with most of the new stuff. Iain Banks is really great. Martin Amis and Irvine Welsh for non-sf fiction. Don't watch much television. I will admit to watching the Simpsons, Futurama, and Earth the Final Conflict (never met anyone else who watches that one.)(Editor's note: Sleen watches it)

There's the occasional jam session. I'm taking voice lessons, and I'm working on violin, although I did cheat and have a luthier make me a fretted violin. I putter about in my home studio. I have finished one CD's worth of material, and I'm now waiting for inspiration to strike to start another.

SONIK : Who are your musical influences?

BOB : The first song I ever remember liking was "Purple People Eater" by Sheb Wolley. Around that time I started taking accordian lessons.I really wanted to play guitar (because I wanted to be a cowboy, and all the cowboys played guitar.) But when my mother went to a music store to sign me up for lessons, they convinced her to have me study accordian! (Well, this was Los Angeles in the 50's.) Actually, I am forever grateful for this decision, because it got me started on a keyboard (well, at least in the right hand) and reading music, all at a young age. I made pretty good progress, but I didn't keep it up after we left LA.

Then in seventh grade the music teacher got me very interested in classical music. I started taking classical piano lessons a year later. I was really into Chopin, Debussy, and Beethoven. Never got good enough to play Chopin's Ab Polonaise, though. Started improvising, and writing pieces. Not very good, I'm afraid. I kept that up until I went off to college.

By that time I was getting interested in pop music. Simon and Garfunkel, the Beatles, the Stones, Bob Dylan, Jefferson Airplane. I thought I was pretty hip, but my first roomate at college had Muddy Waters, Howling Wolf, the Jim Kweskin Jug Band, the Grateful Dead, the Velvet Underground. So I got hipper. I started playing guitar, bass and organ during college. No real bands, just jam sessions. Got a Rhodes and my own Hammond during grad school.

Ah, yes, let me share my most mystical music experience, and it doesn't even involve drugs. I had been listening to a lot of Commander Cody and the Lost Planet Airmen, who played western swing/rockabilly music. At this point I didn't have a keyboard of my own. I went to look at a Clavinet I saw advertised (this was way before "Superstition" came out, so don't even think about that...), I stepped up to the keyboard, and without even thinking about what I was going to play, I started playing their boogie-woogie tune "Beat Me Daddy, Eight to the Bar." Isn't the human brain wonderful?

So I kept up with all the trends, psychedelic, country rock, fusion, even some disco. (Sorry, no prog rock.) By 76 I was pretty bored with rock and started getting into swing era music. In 78 I started listening to some of the excellent college radio stations here in Boston, and discovered the whole punk/new wave thing, which I really got into. This was also about the time I heard the German electronic bands like Kraftwerk, Tangerine Dream, and Klaus Schulze.

I think my tastes have pretty much stayed the same since then. Mass market music has become pretty uninteresting to me, although the grunge thing was promising. There's still a lot of good stuff out there, if one looks hard enough.

Current favorites, in no particular order:

Man or Astro-Man?, Tom Waits, Nico, Hawkwind, The Bevis Frond, Sisters of Mercy, Steve Roach, Robert Rich, Danielle Dax, Dock Boggs, Freakwater, Nuggets compilation, Grateful Dead live stuff from 68-72, Love Spirals Downward, Spacemen 3. I'm not apologizing for any of these.

I'm not sure what the underlying theme is. I don't believe any of these recordings ever used a Kurzweil synthesizer, though. One of the exciting things about this business is seeing my heroes use our gear. I remember watching the 1985 Live Aid show on television, and seeing Chris Stainton with Eric Clapton, and Billy Currie with Ultravox, both using K250's. Alas, I seem to be too old for hero-worship at this point.

SONIK : How do you see "software" affecting "hardware" in the future?

BOB : Well, that's a tough question. We discuss that issue periodically, of course. Could we be successful doing software-only products? We would have to use the same CPU's and/or DSP chips that everyone else was using. Could we do any better? I suppose we would make different trade-offs, and probably get higher audio quality at the expense of less voices.

We like the idea of a musical instrument that seems more like a musical instrument than a computer. We can also be more uncompromising when it comes to audio quality issues, since we use our own DSP chips. Now, it would be quite another matter if a single Pentium could indeed do everything a musician wanted, but fortunately the perceived needs of the marketplace seem to grow exponentially. So I'm not terribly worried about our style of products becoming non-viable in the near future. Still, I will admit to some level of background nervousness.

SONIK : MARCH 2001