Frank da Cruz
Columbia University*
February 2001 Most recent update:
Sat Sep 2 19:37:37 2023
The RCA Mark II Synthesizer at the Columbia-Princeton Electronic Music
Center at Columbia's Prentis Hall on West 125th
Street in 1958. Pictured: Milton Babbitt, Peter Mauzey, Vladimir
Ussachevsky.
Studio 317, one of four composition studios at the C-PEMC, circa 1970.
Clockwise from the center front, Vladimir Ussachevsky (seated), Milton
Babbitt, Bülent Arel,
Pril Smiley, Mario Davidovsky, Alice Shields, Otto Luening.
I was a student from 1948 to 1958 and taught electrical engineering from 1952
to 1962. But I don't remember any computer facilities for students or staff
during those years. Your document has 1962 or so as the founding of CUCC,
about the time I joined what was then Bell Telephone Laboratories, owned by
AT&T. I have still some association with Columbia.
There was at what was then The Columbia-Princeton Electronic Music Center a
Z-80 based computer that used big (8-inch "frisbee") floppies, an LSI CRT
terminal, and a dot-matrix line printer. I think the name of the computer
maker was Cromemco. We installed a switch in the LSI terminal to reset the
computer and built a small rack-mounted panel to simplify transferring files
between the Cromemco and other computers, some of which were reached by modem.
Lim-Lip Yuen constructed the panel and
graduated in 1992 so I know the Cromemco was acquired well before then. The
Cromemco ran a good BASIC. Some code was also written in assembler. The
Cromemco was still at Prentis (632 West 125), and working, when my association
with the Center faded "under new management" following the death of Professor
Vladimir Ussachevsky.
I'm pleased that I was able to get Virgil deCarvalho, now in charge at Abacus
Consulting, Inc., to write about some early computer activity at the
Columbia-Princeton Electronic Music Center. Virgil was in charge of
engineering at the C-PEMC after I joined Bell Labs in 1962.
From Virgil deCarvalho (18 February 2001):
Virgil deCarvalho and Brad Garton in Studio 324, late 1980s.
Charles Dodge: The Earth's Magnetic Field,
LP record jacket.
Charles Dodge ran the computer music program at Columbia until, I think, the
late 1970's. Earth's Magnetic Field (1970) and
Changes (1970) were two computer pieces by Dodge that I know of
that were produced using facilities at the Columbia Computer Center and the
C-PEMC; both were published on vinyl LPs by Nonesuch Records. [Howard
Eskin of Watson Lab collaborated on these productions; Howard was also
Virgil's advisor.]
I met Charles Dodge when he asked me if I could build some filters for the
DAC (digital-to-analog converter) he was using at Columbia's Nevis Labs in
Irvington, NY, a division of the Physics department. There was a large IBM
360 (360/44 ?) machine at Nevis that was hooked up to a DAC. Dodge was able
to get time on this machine and he used it to convert his digital music to
analog tape. We installed a four-track Ampex 354 tape deck in the computer
room and eventually built filters and a replacement DAC (the IBM DAC was
really more suited for driving analog mechanical devices than for sound). The
process: music (and speech) was directly synthesized using the [IBM 360/91]
mainframe at the Columbia Computer Center; the synthesized output was written
to open-reel digital tape; the tapes (and also punch cards) were transported
to Nevis where they were converted to analog sound (for a reason I didn't
understand, the IBM 360 had to stop whatever it was doing to do the
conversions) and recorded onto analog tape using the Ampex 354; the analog
tape was brought to Prentis for editing.
Mark Zuckerman took over the Computer Music program in the late 1970s and
(at some point) purchased a used Data General Supernova computer and
Kennedy tape drive unit (we built a new DAC and filters and the Ampex
was brought over from Nevis) and installed them at Prentis to do music
conversions. This machine's sole purpose was to convert music that was
written to digital tape (at the Computer Center) into music that was
written to analog tape. Now the Computer Music program owned its own
music conversion equipment and no longer had to beg time from Nevis,
where the physics people were getting increasingly irritated by the
use of their computing equipment for something as trivial as music.
Purchasing the Data General saved a lot of time because students
could get the results of their work without having to wait for
things to go to Nevis and come back. And there were days when large
experiments were crunching at Nevis and conversion could not happen at
all -- this was not real-time. However, the Data General had one little
quirk. When it was powered off (which happened sometimes), we often had
to boot-strap it back into operation by manually keying in a small
program using the many toggle switches on the front panel -- a real
pain.
Bootstrapping minicomputers from their console switches was
standard procedure through the mid-1970s. You literally had to load 16-bit
binary words -- the instructions of the bootstrap program -- one bit at a time
("touch typists" could do three at a time in octal) into the switch register,
deposit them in successive memory locations, and then load the starting
address of the program into the program counter, then press "start". -ed.
The Cromemco Z80 was bought in 1978 for a project (The CPEMC Digitally
Driven Studio) involving computer control of analog synthesizer devices
in the studio. I believe the Z80 used a disk operating system that was
earlier than CP/M (I may be wrong). Ussachevsky was interested in
bringing computer automation to the EMC and in 1977 asked me for an
idea. I proposed that we build a computerized device that would drive
the voltage-controlled synthesizer equipment in the studio. I wrote
a proposal that was submitted for an NEA grant of $2000 or so. While
we were waiting for the grant approval, I worked on the design of the
amplifier stages (32 of them) that would turn the computer signals into
control voltages. When we got the NEA money we bought the Cromemco and
all of the stuff we needed to build the device. The Cromemco was chosen
for two reasons: 1. it was the most advanced microprocessor on the
market, 2. it was a kit and highly customizable and would allow us to
insert the circuitry for the device we were building. Jim Constantine
helped with the digital interface. Mark Zuckerman wrote the software in
Assembly code. The computer operated on an S-100 bus architecture and we
had to tap into that to get at the digital signals we needed. After many
fits and starts I got the digital-to-analog amplifiers designed and we
put the whole thing on perforated boards that went into the Z80 computer
chassis itself. A large 25 foot long cable hung out the back of the
computer and this was wired into a patch panel in studio 317 (the
computer had to live outside the studio because of fan noise).
Ussachevsky experimented with it for a few years and I think managed
to produce a few sounds that he used in a composition. I'm not sure
if anyone else tried to use it. When I was at the EMC two years ago
it was still there in one of the back rooms.
Prior to the Sun and NeXT machines there was a Data General Supernova in
the computer studio in room 324. Mark Zuckerman brought in this machine
(1977?) for the computer music synthesis program he was running. The
analog end for the machine and the smoothing filters that were used to
produce usable sound were custom-built at the EMC. I think this was the
first computer that the Center owned for music synthesis. Prior to this
the computer music program was using an IBM 360/? mainframe at Nevis
Laboratories for music conversion. We built the analog end for this
machine too. Somehow Charles Dodge was able to get time on the Nevis
machine at night and on weekends. His students did their programming
at the computer center on campus and sent their jobs on punch cards up
at night with Dodge to Nevis for conversion.
Howard Eskin 1997
Howard Eskin** recalls (in 2007):
I found also 2 dusty LPs (1969 and 1976) of computer music done at CU mostly
by Charles Dodge. I helped on parts as did Richard Garland at Nevis Labs. As
I recall I mostly did speech recognition and synthesis software (I had been
doing some R&D in that area) and I gave his class a few guest lectures on the
subject.
The programs were run at the Columbia Computer center
(IBM 360/91) late at night. Digital tapes driven
to Nevis Physics labs [in Irvington NY], where they were converted to audio
via D to A converter the size of a clothes closet (now done on a chip in a
generic PC).
C-PEMC album covers from 1969-70
Charles Dodge
The photo of Charles Dodge is from 1976, then assistant Professor of Music at
Columbia (now at Dartmouth I think).
*
Retired from Columbia in 2011.
**
Howard Eskin was professer of Electrical Engineering and Computer
Science at Columbia University, 1970-75, and Director of Academic Computing
at Columbia 1975-84. During this time he collaborated with members of
the Electronic Music Center, mainly Charles Dodge. Howard left Columbia in
1984 for General Electric. Howard
died
in 2022.
