Dubstep's great-great-grandad

September 27, 2016 3:01 AM

More famous for helping to crack the Enigma code during World War 2, Alan Turing also created the first ever computer-generated musical notes in 1948. In 1951, a recording - the first ever of computer-generated music - was made at the BBC. The recording was restored this year at the University of Canterbury in new Zealand and can be heard here [mp3]. via @v21
posted by EndsOfInvention (21 comments total)

- I like how the off-key notes make it sound like a kid practising the violin.
- I really want someone to sample this for a modern electronica/EDM/dubstep/whatever piece.
posted by EndsOfInvention at 3:02 AM on September 27, 2016

What a beautiful story! Thanks!
posted by Joe in Australia at 3:18 AM on September 27, 2016

Daisy, Daisy, give me your answer, do. I'm half crazy [giggling] all for the love of you.
posted by steef at 4:27 AM on September 27, 2016 [2 favorites]

Of course they start with God Save The Queen
posted by timdiggerm at 4:37 AM on September 27, 2016 [7 favorites]

Those voices are incredibly lifelike!
posted by Thorzdad at 4:57 AM on September 27, 2016

Of course they start with God Save The Queen

It's (I presume) really quite handy for restoration purposes.

(Also, I like that the apparently only readily available picture of Strachey is exactly the sort of picture everyone seems to have of their relatives in the 1950s. It's a contrast to the "standard" picture of Turing, which I think might have been a passport photo.)
posted by hoyland at 5:42 AM on September 27, 2016 [2 favorites]

That "In the Mood" will totally be my ringtone.
posted by scruss at 5:50 AM on September 27, 2016

This is a classic piece of reverse engineering, using the same sorts of research and process as people do when resurrecting old games.

Digital archaeology is fun, and if it's not a genuine field already it will be thoroughly established in twenty years. If any historically-minded young Mefites are considering where to head off with a good chance of making a name for themselves, I'd get stuck in.

Plus, what a great story. Every bit of it played out in my mind as clearly as a movie.
posted by Devonian at 5:56 AM on September 27, 2016 [3 favorites]

Computers incidentally being able to make music was a common misfeature in the 70's as well. Here is a PDP/8 playing music through noise picked up by a nearby AM radio.

Max Mathews took a different approach - he modeled music and generated digital samples written out to magnetic tape. The tape was then run through a D/A converter to produce analog output. Max's first version was generating music in 1957. Here is a piece from 1960 called Numerology.

In 1979, the Apple II had about the cheapest possible sound hardware, which was a 1 bit D/A converter. You reference a memory address and the speaker would move from in to out or vice versa, which is similar to what Turing was doing with the instructions on the Mark II. There was an older MeFi thread here about the Apple II, with a section on convincing it to make two voice music.
posted by plinth at 6:43 AM on September 27, 2016 [1 favorite]

Of course they start with God Save The Queen

Also via @v21, an earlier example of machine-assisted music:

In 1845, John Clark built the Eureka Machine for Composing Hexameter Latin Verses, which he exhibited it at Egyptian Hall in Picadilly. Words were pre-encoded on turning cylinders, whose projecting pins would cause letters to drop, forming Latin verses. The cylinders were turned random amounts, so that a different verse was formed each time, but each verse had the same grammatical shape: adjective, substantive, adverb, verb, substantive, adjective.
[...]
During the composition of each line, a cylinder in the interior of the machine performs the National Anthem. As soon as the verse is complete, a short pause of silence ensues.

posted by EndsOfInvention at 6:47 AM on September 27, 2016 [1 favorite]

The Eureka Machine for Composing Hexameter Latin Verses was also the first ever Twitter bot.
posted by EndsOfInvention at 6:48 AM on September 27, 2016 [1 favorite]

Another favorite post from the same blog.
posted by in278s at 7:08 AM on September 27, 2016

This is amazing, I'm amazed this was reconstructed. It's so early in the history of computers--Max Mathews wouldn't write "The Digital Computer as Musical Instrument" until 1963.
posted by LooseFilter at 8:04 AM on September 27, 2016

Of course they start with God Save The Queen

Pretty sure that's "My Country, 'Tis of Thee"!

/runs away cackling
posted by Celsius1414 at 8:19 AM on September 27, 2016 [1 favorite]

I usually listen to classical while I work, and it blended in quite nicely.

Thanks for posting!!!
posted by JoeXIII007 at 9:25 AM on September 27, 2016

So yeah I got a bit testy at the "it's an urban legend" about Bell Labs. Is it Urban Legend or more that is what we have safely understood for many years until new evidence and material arises, perhaps not as well known?

Anyways. The descriptiong of using audio to figure out what's going on, is interesting to me, because it sorta reminds me of the cassette tapes/modem noises, that the signals are the audio themselves (due) but that they contain not noise, but actual information.

What happens when you load a song from cassette (not a program encoded as "noise)) into the memory of a TRS-80 and view the bytes directly. Would they correspond to the same mapping. Is it possible to load the signals into the system without crashing it? Like could I load a song into memory the same way I would with software (doubtful, I'd *hope* they had some error correction stuff involved in transcoding audio to data, but who knows back then), and then peek/poke my way around on the bits themselves? And what would they write back out as upon rewrite to disk?
posted by symbioid at 10:50 AM on September 27, 2016

Computers incidentally being able to make music was a common misfeature in the 70's as well.

If you have a TI-83+ graphing calculator lying around, I remember it being able to do the same trick.
posted by BungaDunga at 1:19 PM on September 27, 2016

As recently(?) as the Atari 2600, computer musicians have had to deal with synthesizers that could produce only sorrowful approximations of the western 12-tone scale. On that console, each voice only had 32 different tuning numbers covering 5 octaves, and frequencies from two different voices (often called "distortions" in the documentation I ran into) seldom match up squarely.

For instance, according to this source you can see that for "distortion 1"on an NTSC machine you can get f4 spot on, c4 pretty close, but the only note in between is a not particularly good approximation of d4.

Elsewhere, you'll be advised:

You should know that it's *much* easier to write your own music for the 2600 than it is to port a song to the 2600 and end up with something that sounds good. The pitch is so limited that often you will not be able to find notes that are reasonably in tune in the sound types you want, and this will lead to compromises. If you write your own music, you can write based on the notes you know that the Atari can play in-tune and what sounds best.

Good advice, when you consider this rendition of the opening of the famous "Toccata and Fugue in D Minor"

Holy cow, the Atari 2600 (introduced 1977) is closer in time to Turing's activities in 1948 than to 2016.
posted by the antecedent of that pronoun at 7:15 PM on September 27, 2016 [1 favorite]

Since it was mentioned in the title, when I was in San Francisco, I encountered a street busker doing beatbox dubstep. Just him, a mic, and an amp. It was insanely amazing.
posted by Thorzdad at 7:22 PM on September 27, 2016 [1 favorite]

Of course they start with God Save The Queen

Actually, no. In 1948, it would have been God Save The King. Which He didn't, because George died 4 years later

Nice to know they've upgaded the computer at UofC because when I was there the best it could do was tell you what day of the week you were born on.
posted by arzakh at 3:41 AM on September 28, 2016

Thanks for posting this. If only because it's opened the doors of the British Library's sound archives. I could listen to this all day. (Hovercraft full of Eels notwithstanding.)
posted by not_on_display at 9:37 PM on September 28, 2016