KateI have collected data from 60 participants over the last two days, and the preliminary data analysis looks really good. Overall, participants are performing as hypothesized or better. I suspected that the study I am basing this study on underestimated the ability of people without musical training to do this task (or a similar one), and it appears that I was right! They’re doing even better than I expected.
This experiment is looking at how many instruments people can count when they are playing together—for example, if listening to a recording of a string quartet, how well can people tell if it is 1, 2, 3, or 4 instruments playing at any given moment, and how readily can they identify when the number changes (one or more start or stop). My musical examples go up to 5 instruments (or musical “voices”). The hypothesis is that trained musicians would be able to identify 1-3 quite well, and then that accuracy and ability to identify changes would go down fairly rapidly with textures of more than 3 voices, and that people without musical training would perform less well than those with musical training.
Very preliminary data analysis seems to be showing that people, regardless of musical training, are able to count up to 3 pretty well, and accuracy drops considerably for 4 and even more for 5, and it drops a bit less for people with musical training than those without.
Klara@musiciankate are the instruments identical? Like 5 soprano violins with more or less the same sound?
@Klara Great question, I left that part out but it is a key part of my study. The instruments are not identical because the range of the musical excerpts is too wide, but I tried to make the timbres similar enough to seem the same. I used fugues by Bach and Musescore MIDI instrument sounds for five different instrument timbres, and in this experiment all of the instruments in a given trial were the same general instrument timbre. My follow up study will have the same musical examples but mix up the timbres, so that no two instruments in a given trial have the same instrument timbre.
@Klara There is a lot of evidence that timbre differences make it easier to separate musical “streams” or “objects,” and I am curious to see *how much* easier it makes this task.
@musiciankate I sing all kinds of choir music, right now I'm trying to study Monteverdi's Maria Vespres, which has 8 voices and more in places. The differences in human voices helps.
I also found that having the sheet music for the first time, while listening to a well known peace (and being able to read) makes a big difference in what my ears discern.
I also found that having the sheet music for the first time, while listening to a well known peace (and being able to read) makes a big difference in what my ears discern.