The Synth That Survived ARP's FallBy Mark Vail
The following article is being reprinted with permission from Keyboard Magazine, October, 1993. It is similar to the chapter on the Chroma in Vintage Synthesizers (see the Bibliography for more information).
When the seas came crashing down on ARP Instruments in 1981, two-year-old plans for a programmable polyphonic synthesizer were seemingly squandered. The team that had several times tried to surf through internal company resistance to produce the instrument was scattered by the waves.
As detailed in the article that appeared in the April 1983 Keyboard (and Inc. magazine before that) [see The Rise and Fall of ARP Instruments], Philip Dodds was left at the helm of a nearly submerged ARP. While bailing out the lifeboat ... er, cleaning up the disorder that remained, he managed to sell the Chroma design to CBS -- and get himself hired as the director of its production.
The Chroma represents what could be thought of as the second generation of analog/digital hybrid synthesizer instruments. It came along at the peak of the Sequential Prophet-5's popularity, offering then-radical functions like multitimbral operation, voice layering, and keyboard splitting, not to mention velocity response -- features that we take for granted now. Voice structuring wasn't limited to the typical Minimoog configuration (two oscillators, a filter, two ADSRs, and a VCA); the Chroma's internal computer could route signals through two lowpass filters, in parallel or series, or position the VCA before or after the filters.
In addition, there was a computer interface that allowed Chroma performances to be digitally recorded and played back. Voice-editing software was even developed for the Apple II and IBM-PC. Digital sequencing and editor/librarian programs are commonplace today, but the Chroma came out at least a year before MIDI.
The Chroma also helped usher in a concept that annoyed some synthesizer programmers: menu-driven voice programming with a single data slider and multiple dual-function membrane switches (which probably influenced Yamaha in their design of the DX7). Previously, single-function knobs and sliders (one control per parameter) abounded, and single-line LCD displays were foreign and unfriendly to confirmed knob-twisters.
Philip Dodds is proud of the Chroma, as you'll see from the interview that follows. These days, he's the executive director of the Interactive Multimedia Association, which is made up of big guns like IBM, Apple, Kodak, and DEC. "I'm the head of the trade association that is setting standards for multimedia in desktop, workstation, and network computers," he says. Our sympathies! What this man needs is an extended programming session with an ARP 2600 or Chroma -- or even the ARP 2500 that he played to the aliens from out space in Close Encounters of the Third Kind. Sad to say, though, Philip Dodds doesn't own any of the synthesizers he helped develop.
How did the Chroma's development come about?
Conceptually, the Chroma began in the fall of '79. It started into serious development in the latter part of '79 and 1980. We were ready to bring it out in '81 when everything at ARP came uncorked. When ARP went into receivership, among the first things I had to do was shut down R&D, which was painful because I was the VP of engineering. Everything froze. People went off and tried to get other jobs, and many did. We lost a full six to eight months from the development cycle, because there was a lot of lost momentum internally while the company folded itself up and I was off trying to figure out how to get the Chroma back up and going. But at the time, I felt we had done enough development and that the product was still viable. Eventually, I was proved to be correct, even though development was interrupted twice, which normally doesn't occur in business.
Do you mean development was stopped once with ARP and once afterward?
Actually, it was stopped twice with ARP and we started again after CBS bought the rights to manufacture the Chroma. With ARP, I stopped the design a couple of times because it was heading down the wrong paths and because there were new developments and new technologies available that could put us significantly forward. Plus it just wasn't ready for prime time, and having suffered through releasing products that weren't ready for prime time, I wasn't about to allow that to happen again.
Did you get any resistance?
Oh, man, I got a lot of resistance internally.
From ARP president David Friend?
Principally, but then also from David Spencer -- who's not often heard of, but was chief operating officer right after David left, became president briefly, and was the chief financial officer for a period. He was a very good guy, but really had no idea about the technology. He tended to back my decisions, lacking any other sort of direction. Good guy, but not a music industry veteran. He was sort of left holding the bag. David Friend left ARP a substantial number of months before the company actually filed Chapter 11, and David Spencer took over in his absence.
When did you finally get back to working on the Chroma?
Through the summer of '81, I negotiated the sale of the Chroma to CBS, and in the fall we pulled the R&D team back together. We opened up the CBS R&D labs in Wolburn, Massachusetts, with 21 people. We had a pilot production area where we finished the design of the Chroma. We built the first 50 Chromas right there in Wolburn ourselves, because basically CBS was clueless about how to build electronic instruments. They built the Rhodes electric piano, but building instruments with tines that vibrate is a little bit different from making instruments that were built on an Intel 80186 computer chip. Worlds apart!
The keyboards were built by Kimball, then later by Pratt-Read before they went under. They had a wood-ply base. there was a nine-layer birch-ply bed with a full wooden keyboard action, which I designed. Actually, there were a lot of people who loved the feel of the keyboard. We also used solid cherry end blocks. I was trying to go for as much real, genuine wood as possible for feel and flavor.
The Chroma was the first keyboard with velocity-sensing keys that also sent velocity data out via an interface. It was the first commercial synthesizer with a keyboard interface with software that ran on an external personal computer, the first one being the Apple II, followed by the IBM-PC. Tony Williams programmed the sequencing software for both computer platforms. This was before MIDI existed. We actually developed the first desktop computer software for sequencing, as far as I know, anywhere. The Chroma had an external port that was fully documented and available to anyone.
There were 256 different velocity levels. We were the ones on Dave Smith's original committee that said that MIDI had to have velocity. Dave Smith said it was too complex and we couldn't do it. We dug our heels in and said it had to have velocity response. And he said, "Well, that's just because you're trying to favor the Chroma." "No," I told him, "One day keyboards will respond to velocity. Trust me on this."
Was it multitimbral?
The Chroma had 16 discrete synthesizer channels, complete analog synthesizers. There were two VCOs, a cool multimode VCF, a VCA, and software-created envelope generators. There were 64 sample-and-holds that controlled all of those channels, and they were real-time updated and driven directly off the central computer.
The heart of the Chroma was an Intel 80186. We were a beta site for Intel. We helped Intel debug the 186, which preceded the now-famous 286. We were ahead of the PC, which had an 8086. If you flip up the Chroma's front panel, on the right side you will see an 80186, a stack of RAM, and various other silicon glue: It's a CPU card. Take that card out, put in a SCSI controller and a hard disk drive, and you've got a desktop PC. It was a full-fledged programmable computer in a very proper sense. Paul DeRocco, the chief hardware architect, designed this system and wrote all the Chroma's assembly code. Paul selected the 186 because it had a very good I/O [input/output] structure that allowed a lot of simultaneous data to come in on a lot of data lines. The 80186 was designed as a peripheral controller, whereas the 80286, 386, 486, and so on are really designed more for desktop PC usage. But it's the same family and same instruction set.
Was it a new experience for you to incorporate a microprocessor in a synthesizer?
Not really. The predecessor product, the Quadra, was fakey, but it was among our first microprocessor-based instruments. It used the 8048, which was a single-chip microprocessor that scanned all the keys, scanned all the front panel controls, and set up all the patches. The Quadra was basically an old ARP Omni with a two-note synthesizer layered on top of it.
My guess is about 3,000. We built the first 50 in Massachusetts, until we could prove that everything was in good shape. Then we set up a production line in the Gulbransen organ factory, which was also a CBS company at the time, and had just all kinds of quality control problems.
Working with CBS wasn't easy. Their Gulbransen plant was in the middle of a cornfield in the middle of nowhere. "They know all about building stuff," CBS told me. "Okay," I said to myself. "I'm not sure how I feel about this, but they're the new bosses." So we tried to train the Gulbransen people to build synthesizers. These were people who had built home organs for 15 or 20 years, and that's all they knew.
The Chroma was a 16-channel instrument. It had eight dual-channel voice cards that plugged into a motherboard for easy servicing. The first thing the Gulbransen people did as a cost-cutting measure was to start processing the channel cards in a new way. They had just started using this water-soluble flux in their production line. The boards were soldered and put through a water-soluble solution. The process saved a lot of money and it was better than doing the old method of flux removal. But the process was still experimental at the time.
Anyway, they ran about 300 or 400 channel boards through this process, and we immediately started hearing about problems in the field from customers -- often major-name customers -- whose instruments would simply never stay in tune. It took us the longest while to figure out what was going on, and we finally figured out that it had to be the flux-removal process. But the manager of the plant told me, "No, it can't be. You don't know anything about manufacturing, you're just a nerdy engineer. We know about this stuff, you don't."
Finally Robert Hartford, who did all the detective work, took the boards up to a testing lab in New Hampshire and proved that each of the circuit boards had been contaminated through this process such that whenever the humidity rose, the oscillators would go sour. CBS ended up eating probably about $40,000 to $50,000 worth of channel cards, because there was no way to salvage them. They had to ditch the entire production line and move the whole thing to Fullerton, California, where they built tube amps and Fender Rhodes pianos.
Wasn't there a slave unit for the Chroma as well?
Yes, the Chroma Expander, which provided another 16 voice channels. We basically chopped off the keyboard and connected it to the Chroma through the ports on the back to go from 16 to 32 channels.
Herbie Hancock was one of the first of many early Chroma buyers. I ran into him a year ago and he said he still has his Chroma. I remember the first time he came up to the keyboard at a NAMM show in Anaheim. He laid down some riffs that were classic Hancock. Of course, I'd captured what he played using the computer. Then I started changing the voice as it played back, and you should have seen his eyes pop. This was before MIDI. I sped it up, slowed it down, reallocated the voices. This is all stuff that's perfectly routine now.
Was the computer just for sequencing?
Well, it had a precursor to the MIDI interface. It was very similar in some ways. You could send out multi-channel information, velocity data, and time-stamp information. In fact, it was on the basis of what came out of the Chroma that we got into arguments with Dave Smith on what MIDI should look like.
As a parenthetical story to that, I was on the first MIDI committee, but CBS came down and said, "Uh oh, we don't like the way in which they're forming this committee." They were worried about anti-trust. As a result of that, they forbade me as a CBS employee from participating, and I was blocked from attending all but the first three MIDI meetings. After that, I wasn't allowed to go. But our velocity stuff got in there, and our opto-isolator [the circuits in the MIDI jack] concepts got in there and were picked up by the Japanese.
There was something about the Chroma that caused the R&D people who were involved with it to say, "This is something really worthwhile to do. This is something that the world needs." There's something magical about the Chroma that really bonded people, making them say, "Let's at all costs (and sometimes those costs were high, personally) make the best instrument we possibly can." We didn't make the best possible instrument in the world, but the devotion to trying was sincere and that's what kept that team together. That's what kept it moving and, even when ARP died, the Chroma lived on. It transcended a corporation's death and kept that team together, and was the sole reason everything moved forward from there.
The Chroma Polaris, a descendant of the Chroma, cost $1,495 and sported MIDI, a built-in sequencer, and interfaces for Apple IIe computer and the Chroma. The Polaris was designed and executed after Dodds' departure from CBS. "although its seeds were sown right before I left," he explains. "The Polaris was a spin-off from the Chroma, conceived on the drawing board as a cut-down, slimmed, low-cost version." Street value of the Polaris today  runs from $200 to $400.
Description: 16-channel programmable polyphonic synthesizer with built-in computer interface, 64-note velocity-sensitive wooden weighted-action keyboard (E to G), 75 front-panel membrane switches, data slider, two programmable spring-loaded performance levers, three-band graphic EQ, volume, and tuning sliders. Destinations for velocity control included pitch, filter cutoff, the rate of the LFO, the waveshape, amplitude, and attack time. The computer interface allowed pre-MIDI sequencing and voice editing functions via Apple IIe and IBM-PC computers. A tape interface allowed program storage on any ordinary cassette tape machine.
Manufacturer: Original designs completed at ARP. CBS bought the Chroma design, assigned it to their Rhodes division, and hired Philip Dodds and the former ARP engineering team to produce the synthesizer.
Production Dates: 1982 to late '83. Development began in the fall of 1979.
Approximate Number Made: 3,000 (including Chroma Expanders). [See the FAQ]
Original Retail Price: $5,295.
Current Street Value: $300 to $450. ($350 for Chroma Expander.)
Current Dealer Price: $450 to $800. ($500 for a Chroma Expander.)
Note that prices have increased substantially since this article was published. See Price History.
Insider Information: CBS picked up the Chroma and ARP electronic piano designs for a mere $350,000. ... Philip Dodds reports that the best Chromas are serial numbers zero through 50 [presumably 21010000 through 21010050], because they were hand-built by the development team in Woburn, Massachusetts. ... Dodds appears in Close Encounters of the Third Kind (1977) as ARP musician Jean Claude. ... Dodds went from vice-president of engineering at ARP to the Chroma development team for CBS, then to Kurzweil, where he "helped birth the Kurzweil 250 and dealt with Ray Kurzweil's personal form of reality distortion." Does Dodds own a Chroma? "I don't have a Chroma," he admits. "That's something that I really want. And a 2600. The reason I don't have either is that back in those days, I did that for a living, and when I went home I didn't want to see the instruments. Now, I regret not keeping one of each."