Hello, hello! In class last week we looked at a few cases of sonifying data, such as the Restless Genetic Mix in which bits of DNA sequences were assigned to instrument parts and it made some slick beats. We then took a glance at some algae songs which is what I wanted to take a deeper look at in my case study.
This microbial bebop project was done by Peter Larsen at Argonne National Laboratory. Larsen is a computational biologist, and the songs were put together as a method for making large sets of data easier to parse by scientists. They felt that sonifying it as classical music would be too rigid, so they borrowed from the improv style of jazz/bebop. I’ll talk a little bit about a few of the pieces:
Blues for Elle – in this composition (the one we heard a bit of in class), the chords are “generated from seasonal changes in photosynthetically active radiation. The melody for each measure is comprised of eight notes, each mapped to a physical environmental parameter… temperature, soluble reactive phosphate, nitrate, saline, silicate, and chlorophyll A concentrations.” Now this might not make a lot of sense to someone not versed in biology, but it gives those who are a new way to look (hear) vast amounts of data.
Far and Wide – this one focuses on a highly abundant plankton called Pelagibacter ubique that follows a distinctive seasonal pattern that is turned into music as follows: Two chords per measure (four beats in this case), generated from photosynthetically active radiation measurements and temperature. Melody is 6 notes a measure describing the abundance of the species. The melody in each measure follows the pattern of the species from a point in time, such as rise in pitch, fall in pitch, rise and fall, fall and rise. The more abundant the species, the fewer rests (where no notes are played) are in a measure. At the most abundant points, there is a cymbal crash.
Check out this picture, this is how they explain what’s going on: “The main advantage of Microbial Bebop over previous efforts to transform scientific data in music is the ability to highlight relationships between data types. The same melody, generated from the same set of data, sounds different when played in the context of chords generated from different data types. In the hypothetical example above, the notes in measure in (A) are comprised of six hypothetical data points. All twelve measures in (B) are derived from the same six data points, but each measure is rectified to different chords, representing the same data played in context of different hypothetical parameters. Each measure in (B) is subtly but audibly distinct, demonstrating ability of Microbial Bebop to represent data in a way that can potentially be interpreted by a listener.” They say it better than I could, haha, but yeah if you don’t read music it might not make much sense, basically each set of notes is going to sound different, so if you understand the relationship, you don’t have to read the music, you can hear it to interpret the data. Well, you don’t have to at all, you can just enjoy the music and leave the analysis to the scientists.
These types of sonification projects may be neat to listen to for us, and may be very useful in helping scientists parse large sets of data, but if they become more common and widespread and easy to understand, it could change the way we learn about things. Imagine picking up an album with a brief description of the parameters that the songs were made with, and then listening to some sweet science jams on your commute in the car while you expand your knowledge on something, (like the patterns in algae in this example, but it can go farther than that!)