Stanley Kubrick’s decision to scrap the musical score originally written for 2001: A Space Odyssey and go with something different is what introduced me to the fantastic, creepy, and wonderful music of Gyorgy Ligeti. When I first signed on to Pandora internet radio, one of my first stations was seeded by Ligeti as an artist. Doing this introduced me to pioneers of early electronic music like Tod Dockstader, Morton Subotnick, John Cage, David Tudor. However, one name which was surprisingly overlooked was Karlheinz Stockhausen. It was only after I created my @mondayazura Twitter account and started connecting with other musicians that I was introduced to his works. Looking back, I now wonder how I managed to go so long without having discovered Stockhausen.
Stockhausen was the first person to publish a score for electronic music. I took it upon myself to recreate that score using the technology I have available to me, namely my computer and audio editing software. I wanted to understand the process Stockhausen used to create a seemingly simple and short piece of electronic music. The piece is only three minutes long, but it took me just over thirty days to realize those three minutes of sound. Here’s how I went about doing it.
Stockhausen created Studie II in 1954 using just a single electronic oscillator, a bit of reverb, and reel-to-reel tape. With only one oscillator, his method for creating chords was quite clever. He recorded a short length of each pure tone, then he cut these recordings into 4 cm strips and taped them end-to-end into a loop of five ascending tones. Then he played the loop through a ten-second reverb to produce each of the 193 chords. I wanted to try and reproduce the technique and sound to the best of my ability using digital technology, so my process for creating each chord was quite similar to Stockhausen’s despite the use of a computer.
I began in Berna 1.0 by recording 1 s clips of each tone, 81 total from 100 Hz all the way up to 17200 Hz. Each tone is separated from the other by a ratio of a twenty-fifth root of five 5^(1/25), and rounded to the nearest whole number. Stockhausen was limited by the technology of his time, but Berna allows for more precision in tone generation. Nevertheless, I rounded my numbers to the nearest whole in order to stay true to Stockhausen’s technique. My numbers turned out to be different from his because of my rounding methods. I used the intervals listed on the score (100, 138, 190, 263, 362, 500, 690, 952, 1310, 1810, 2500, 3430, 4760, 6570, 9000, 12500, 17200) and calculated the in-between frequencies using the ratio above.
||=~ 106.649 Hz
||=~ 113.741 Hz
||=~ 121.304 Hz
||=~ 129.370 Hz
||=~ 137.973 Hz
Repeat this process for each interval listed in the parentheses above. The fifth tone in each sequence should be close to the next one listed in the parentheses, in this case, 138 Hz.
With all the tones generated, the next step was to cut them all into 4 cm lengths. Reel-to-reel tape speed is 30 inches per second or 76.2 cm/s, making 4 cm about 5 milliseconds. One file at a time, I cropped each 1000 ms clip to 5 ms using Adobe Audition. Then I lined up each tone sequentially in groups of five the same way Stockhausen did to create his initial loops. Each chord is created by tone intervals of 1, 2, 3, 4, or 5 apart, and grouped together by their base frequency. Hopefully the chart below clarifies what I mean by that.
Actually, these aren’t grouped by base frequency. But it’s cool to see the pattern.
I had a bit of trouble with pops appearing in between each 5 ms tone. When I tried running the poppy sequence through the reverb, I wound up with a lot of white noise overwhelming the chord. So I had to figure out how to get rid of the pops. Eventually I settled on applying a linear envelope to each tone, fading in, then fading out, to eliminate the pops in the sequence. Running the cleaned up sequence through the reverb then produced the clean chord I was looking for.
Plate reverb: 10000 ms.
The reverb I used was in Berna 1.0. Playing each sequence I’d created in a loop through a ten-second reverb in Berna, just as Stockhausen did with the tape loop, generated the chord I needed for the final realization of the piece. Since the longest note in the Studie II score is almost 5.5 seconds, I recorded each chord at 6 seconds. I could then shorten as needed using Adobe Audition.
While Stockhausen had to manually adjust the audio gain to produce the envelopes for each note, and play the tape in reverse to get increasing volume, I was able to use the faders and gain control for each clip to generate the necessary envelopes for each note in the score. This no doubt has an effect on the way my recording sounds compared to Stockhausen’s method. Considering how much time I put into this piece doing it digitally, I can only imagine how much longer it must have taken Stockhausen to create this piece. After a month of recording, fiddling, manipulating, editing, and tweaking, I finally had the finished three-minute piece completed. Including the extra day I took to write this blog post, I worked on this recording over 33 days, from 29 November 2013 to 2 January 2014. The original piece was written and performed in 1954, and my version comes nearly 60 full years later. Cool!
My previous seven audio pieces on my Soundcloud page were all generated live. None of them has a score to follow, though I did try to document the settings I used to create each piece. This is the first piece of music I made that was from a score. What I hope to do in the future with my sounds is to write scores for them in much the same way Stockhausen did with his Studie II so that others can perform my works if they choose, and not merely listen to them. What this creation and performance of Stockhausen’s score has taught me is that electronic music generated from oscillators and tape is just as complex and time consuming as scoring music on traditional staves for traditional instruments. I have a long way to go before I’m writing music of my own, but I’m one step closer to doing so.