[sdiy] The Coupland

[karl dalen]

Just to complete this elephant graveyard thing, the Coupland digital synth.

-------------------------------------------------------------------

The Coupland Digital Music Synthesizer is a 16-voice polyphonic real-time instrument with a full 88 key keyboard, introduced in the 1970’s.

The idea was first conceived and the basic concepts invented in 1973 by Rick Coupland and John Moore, old friends and systems programmers who were working at Ramada Inns Micor division (Phoenix, AZ) at the time. The project lasted from 1973 until 1979, included two product version, but never reached commercial success due to inadequate funding.

1n 1975, Coupland left Micor to work full-time on creating the instrument, building the first, and 8-bit version. Micor subsequently funded the project and a team including a consulting professor of music and consulting physicist, along with the usual engineers. They designed and built a prototype of the 12-bit version.

Due to funding constraints, the prototype was rushed to a showing at the NAMM conference where the fragile, wire-wrapped prototype failed to function, leading to significant embarrassment. However, the concept and design was solidly proven, and with continued funding would have been a commercial instrument representing a significant improvement in the state of the art.

The performer used a frequency domain specification of the musical voice harmonics, and created an ADSR function to modify it.

The instrument used a waveform buffer, now a widespread practice, but independently conceived by Moore and Coupland. This was driven with high precision phase generator, of which only the high order bits addressed the buffer. A subtle logarithmic algorithm (log eighth root of two) was invented to apply audio level and ADSR amplitude modulation to the signal generated for each voice. This avoided the cost and heat dissipation of available commercial multiplier modules.

A Texas Instruments TI-990 minicomputer performed the non-real-time Fast Fourier transform processing necessary to convert from the frequency domain specification to the time domain waveform buffer format. It also handled user interaction such as specifying sound parameters. The second version used a TMS-9900 microprocessor for the same functions

The first version used only 8 bit precision for its output digital-to-analog conversion (DAC). The instrument suffered from aliasing, which was not due to the typical cause of too low a sample rate or inadequate post-D/A filtering. Rather, it was an inherent effect of low precision calculations in the digital circuitry - round-off error produced digital non-linearity and extra harmonics resulting in digital mixing with the sample rate. The problem was solved by dithering, which did not noticeably affect the intended output, but through the natural characteristics of human hearing caused the narrow-band high frequency alias to become imperceptible. In general, in-depth study of psychoacoustics was used to solve such problems and provide some technical constraints.

The 12 bit instrument was more advanced than the 8 bit prototype, and was enclosed in a stylishly modern plastic case with a futuristic touch-sensitive console above the keyboard. It was small and light enough to be a practical performance instrument. It is still in Coupland's possession.

------------------------------------

I've seen the Coupland called vaporware enough times that I figured I should set the record straight. While the slick looking, performance oriented unit pictured in Vail's book may never have made it past the brochure stage, there were several working prototypes built, and I had the privilege of spending 3 months in a 16-track studio with one of them.

In the Fall of 1980, I was poking my nose into some dim and dusty corners of the music building at the University of Arizona in Tucson when I came across what was obviously a computer with a keyboard attached. Being an EE major with a CSC minor playing keyboards in a band by night, I immediately went to the department head and asked if I could play with it. I was told that the professor who had brought it to the University had left, taking with him all knowledge about the unit. They said that if I could get it to work I could have 6 hours / week in the studio with it for the rest of the semester.

The unit consisted of a card cage with 8 or 9 wire-wrapped cards (one with a TI-9900 processor on it), a dual 8" floppy, 25x80 CRT "dumb" terminal, an 88 key pressure sensitive keyboard, a stack of unlabeled floppy disks, and a one page xeroxed "instruction sheet". Amazingly enough, when I plugged it all together and turned it on, the floppy drive read light came on, so I started feeding it the disks from the pile until I found one that booted to a command menu. After a couple of very late nights it was actually making sounds.

The control software on those disks must have been in a very early state, because it was nothing at all like the brochure  The voice and the modulation waveforms were entered as mathematical functions of the form "sin(x)+cos(2*x)/2+sin(4*x)/4". You could get a square wave with round((sin(x)+1)/2), and with the mod function you could create sawtooth waves. There was no "filter" block, and no way (that I ever discovered) to use the ADSR to control any aspect of the voice besides amplitude.

Those limitations aside, though, it was a truly amazing synth for its day.

At the risk of being too long winded, I'll pass on a little history as well. One evening I took a girl I was dating into the studio. She saw the Coupland and told me that her father had showed her one just like it. It turned out that her father worked for Micor, knew Rick Coupland, and knew quite a bit of the history of the unit. He told me that there were 5 working prototypes: one at Arizona State University in Phoenix, the one I was working on, one that Rick had kept, and he didn't know what happened to the other two. 

He also told me that Rick was very bitter over Micor pulling the funding for the project just as it was nearing completion. He said that when they pulled the funding, Rick quit Micor and took every shred of project documentation with him. Micor sued to get the documentation back, but Rick said that he had destroyed it, and no one could prove otherwise. 

I asked if he could arrange for me to meet Rick, and he called a couple weeks later to say that Rick was not interested in talking to anyone about the project. Period.

I hope this sheds a little light into another dusty corner of electronic music history.

Dave Hamara (ishkabbible at gmail.com)

__________________________________________________
Använder du Yahoo!?
Är du trött på spam?  Yahoo! E-post har det bästa spamskyddet som finns 
http://se.mail.yahoo.com