Unusual Notes
When the radio can’t offer you good music anymore and Spotify offers you too much advertising, where do you look for something new? Obviously many prefer instruments played with a method (even if at times questionable) by musicians. However, there are those who get bored with the usual music and decide to leave things to chance.
Sound isn’t the only type of wave we know of. To tell the truth, the entire cosmos is in constant vibration: from the largest to the most imperceptible oscillations, at very high and very low frequencies, from mechanical to electromagnetic oscillations. Therefore, with the necessary precautions, everything that can be represented with a graph like this can be translated into sound:
This is the case of the sound of two colliding black holes, or of sections of tree trunks that are directly played like vinyl records. The sound that comes out, in this case, is usually far from being considered pleasant.
You don’t know how to play any traditional instrument and are you too introverted to become a DJ? No fear! As you may have noticed by now, you are certainly not alone, and over time others have developed extremely creative, as well as complex, methods to create music starting from things that may seem just crazy.
For example, there are programming competitions in which participants compete to write a code to perform a certain task, but using the least number of characters. Just like in golf, you try to get the ball into the hole with as few strokes as possible. This discipline is in fact called Code Golf.
You may wonder what this has to do with music. Well, then you should know that Rob Miles, a British programmer, wrote a program capable of generating character sequences and then applied a player capable of associating sounds to the generated output. His work, in addition to guaranteeing him victory in the competition, has produced a song called Bitshift Variations in C minor which you can listen to below.
In this video, the programmer explains in person what he did, for those who want to learn more about the computer science side of the matter.
This wasn’t the only time Miles turned to convert data to music. His more recent work makes it possible to hear the sound of the economic collapse of our civilization, associating sounds with the changing values of some stock market indices.
An equally interesting method that can be pursued consists of decoding a set of symbols or objects that repeat themselves in some sequence, relating each object to a specific musical note (without a criterion). A wonderful example comes from the amino acid sequence in proteins.
Also, in this case, the almost complete randomness of the sequences and the arbitrary choice of the notes often creates compositions that do not even remotely respect the musical rules with which Beethoven’s symphonies were composed.
To correct the helm, we have tried to create some simple strategies to convert literary texts in Italian and English into music. The choice of the passages, rich in rhymes, alliterations and repetitions, also serves to reduce the stochasticity of the sequences of letters. The general principles underlying our solutions are the following:
- the notes to be played all fall on a major, minor or whole tone scale
- the attribution of notes to letters follows Zipf’s law
- if possible, we try to add more melody lines to the main one to create harmony
It will be appropriate to explain these principles a little more in detail because not all of us have had much luck playing the flute at school.
The first rule allows us to do nothing overly casual and to choose, broadly speaking, whether the melody produced will evoke euphoria, depression or wonder. The major scale (we used the famous C scale), when used wisely, creates cheerful or euphoric music. The minor (harmonic) scale will leave us a little sad or melancholy.
The whole tone scale, on the other hand, allows you to make mental trips of considerable value if you have also taken psychotropic substances.
In reality, the experiments with the whole tone scale did not produce anything catchy enough so we decided to exclude them from the article. (But who knows the whole tone scale anyway?)
We leave it to the reader, as an exercise, to try the chromatic scale and pentatonic scale.
To understand the second rule, you need to know what the Zipf’s law is.
Take a set of THINGS (let say, words, cities, NBA teams, …) and see how often a certain feature occurs.
For example, consider the words within a text: you will notice that some (few) words are repeated very often, while others (most) appear only once or twice in the entire text.
Now, try to sort the words by decreasing frequencies, that is, from the word that appears most often to those that appear only once. You will find that
the frequencies decrease with the inverse of the position in the ordered set.
This is, in a nutshell, the famous Zipf law: an empirical and mysterious law, but which is very useful in our case for the choice of notes.
Below you can see how Zipf’s law applies in other contexts, such as the number of likes on the Facebook pages of teams in the NBA.
The second rule, therefore, allows us to give more importance (by making them play more often) to the relevant notes of the C chord: C, E, G, C. The less frequent letters take notes of the scale with less logical rigour and an equally special place has been given to D, F and A.
The first attempts made us use a single frequency scale while for another case we tried to separate the vowels from the consonants, promoting the latter forward in the scale.
The third rule is not trivial to solve. Both for the tools we use and of which we are not experts (SonicPi). Both for the criterion with which to possibly assign 3 notes to a single letter. It would be very nice to make sure that it is conditioned by the preceding and subsequent letters and possibly respect the rules of the canon. But it’s a slam.
The simplest procedure consists in assigning other notes randomly, always on the scale, but strictly lower than the main melodic line, giving more possibilities to draw the three structural notes of the C chord and the sixth (A) of the scale. We suggest listening to the Iron Maiden to understand how lustful the sixths could be.
An alternative method to the previous one consists in separating the syllables in the words and attributing at most 3 notes to the 3 or fewer letters in the syllable. However, there are syllables that can be much longer than three letters so we had to cut syllables too long to get around the problem. And here is the result.
An interesting future development could be to try to match different instruments to different texts, for example by basing the choice on parameters that can vary from one language to another such as the number of vowels or consonants per sentence. As you may have understood, there is a lot of room for manoeuvre to experiment!
And have you ever tried your hand to create music from something unconventional? As we hope to have been able to describe to you, it is often starting from the most unexpected ideas that the most interesting results are obtained.
