When talking about Computer Music, where else to begin than the very first computer to produce music. The CSIRAC (Commonwealth Scientific and Industrial Research Automatic Computer) was Australia’s very first digital computer, and world’s fifth stored program computer. It was constructed by the CSIR (Council for Scientific and Industrial Research) by a team under the supervision of scientists Trevor Pearcey and Maston Beard. In 1949 it ran its first program and less than 2 years later it played music for the first time.
Initially known as the CSIR Mark 1, the CSIRAC was in its time, at the forefront of technology. It was fast, efficient, and stored a lot of memory. Today the CSIRAC pales in comparison to the advances in computer technology achieved since then. It was a computer that needed to take up the space of an entire room, and by today’s standards was actually slow (1,000 cycles per second) with not very much memory at all (about 3KB of disk memory). To input information into the computer, a form of long hole punched paper was used. These holes would be converted later into text on another machine. Computer scientists added an output speaker to the machine that would make a sound once the computer had finished running the program letting them know the task had been completed. The machine’s output sounds are indeed the sounds that were later used to create music, but making music using them would prove to be a difficult challenge.
The CSIRAC used mercury acoustic delay lines meaning that a pulse (originally being inputted in the machine using the hole punched paper) would be sent into a memory tube that would then travel back in forth in the tube permitting storage for a number of bits and digital words in a single tube (about 20 memory tubes could function at any time). Unfortunately, musically this posed a problem. Not only was the computer processing information slowly and delayed, but each mercury tube accessed memory at a different time, therefore making it extremely difficult for any task to be performed in a time-crucial manner such as playing music.
The first to put the CSIRAC to the musical test was George Hill, a mathematician with a strong musical background; his mother was a music teacher, his sister a performer and himself possessing the rare ability of perfect pitch. This would prove useful when attempting to create music with the CSIRAC, a machine creating sounds by sending raw pulses through the computer and to the output speaker. The pulses sent would need to be carefully programmed in order to avoid random jabbers of sound. Hill realized that to create music, he would need to send the pulses in a structure to produce a steady pitch. Naturally, this task was easier said than done, considering that each memory access took a different amount of time.
After a long period of trial and error, Hill was finally able to use the CSIRAC to create tunes from already existing songs, such as Colonel Bogey or Girl with Flaxen Hair, and hence become the first computer to play music.
CSIRAC’s rendition of the Colonel Bogey tune
Unfortunately, none of the tunes played were recorded at the time. Since then the CSIRAC use was redirected to other areas and sciences, and was no longer used to play music. However, in 2018 the University of Melbourne used similar hardware to that of the CSIRAC (same mercury tube system) in order to run the old programs. Scientists then read all of the hole punched paper and input all of the data in Hill’s work, which did in fact end up reconstructing the pulses accurately and playing the reproduced tunes.
The work of the CSIR and George Hill helped discover a whole new way of making organized sounds and tunes. Although Hill did not create an extensive library of music or create new pieces as a composer using the CSIRAC, he was still able to create something that had not been done before, leading to the work of composers such as Lejaren Hiller (Illiac Suite, 1957) or Max Mathews who started looking into computers as a true musical instrument.
Video with additional information about CSIRAC and some footage of the machine and the scientists involved in the work at CSIR
Information Sources:
Busch, Paul. “Pearcey Interview about CSIRAC Music.” YouTube, uploaded by pauldoornbusch, uploaded 28 September 2009, Accessed September 9 2020. https://www.youtube.com/watch?time_continue=261&v=slr75sLhOCs&feature=emb_logo.
“CSIRAC: Computing in Australia begins with CSIRO.” CSIROscope, Accessed September 9 2020. https://blog.csiro.au/csirac-computing-australia-begins-csiro/.
“CSIRAC.” Wikipedia, Wikimedia Foundation, Accessed 9 Sept. 2020. https://en.wikipedia.org/wiki/CSIRAC.
“History of Internal Memory Technologies.” Memory and Database, November 18 2006, Accessed September 9 2020. http://memoryanddatabase.blogspot.com/2006/11/history-of-internal-memory.html.
ABC Science. “CSIRAC – The first computer to play music
” YouTube, uploaded by ABC Science, uploaded 6 May 2015, Accessed September 9 2020. https://www.youtube.com/watch?time_continue=261&v=slr75sLhOCs&feature=emb_logo.
“CSIRAC, Australia’s first computer” The Science of Everything: COSMOS, August 3 2018, Accessed September 9 2020. https://cosmosmagazine.com/technology/csirac-australia-s-first-computer/.
“CSIRAC – Colonel Bogey” YouTube, uploaded by TortoiseWrath, uploaded 1 Feburary 2016, Accessed September 9 2020. https://www.youtube.com/watch?time_continue=261&v=slr75sLhOCs&feature=emb_logo.
So did the CSIRAC convert the holes into text for another machine to interpret, or did it play sounds with its output speaker? I am not quite sure whether the output speaker was solely there to signal that the operation was done or if it was also used to play music according to the hole punches.
The CSIRAC was a computer that converted the holes into text and also interpreted the text. Originally, the output sound was simply there to tell the scientists when the computer had finished running the program. However, to then make music, the hole punches were input into the computer so that the raw pulses of the computer’s data words, the bit stream pulses, would be sent directly to an audio amplifier with a loudspeaker attached. The loudspeaker’s circuit would then produce a variety of frequencies depending on the pulses received.
I hope this makes sense. If not let me know.
I liked your thorough explanation of the technology behind the computer itself—it was particularly cool to learn about the mercury tubes used to send around signals. Nice job!