Freesound Forums

The Semantic Audio team over at MTG needs your help: They are trying to "build up a model relating subjective musical complexity and some acoustic-musical features that they can automatically compute from the audio".

Uhm, okay... It all sounds a hell of a lot more complicated than it really is, you just go over here: http://mtg100.upf.es/simac/websurvey/
and take the survey they created. It takes about 30 minutes, mostly spent listening to music. What they are trying to do is teach a computer to think like a human when it comes to "complexity" in music.

I did it and it's kinda fun Once you have finished you will probably have discovered some new music too! Feedback and comments are welcome, you can send them to Sebastian who created this survey.

Thanks a lot!!!

- bram

I took it, it was fun.
My main concern is on the method: complexity is not just a matter of harmony/rythm/melody/timbre, but also of structure. And the fragments are too short to determine structure, you can only guess...
It was interesting, though

I got right thru to giving my feedback at the end, hit some wrong key, and it all went belly up. I could go back thru the pieces, but could never return to the feedback page. The "continue" button took me, for example, to "song 34 of 30" etc. ... ever increasing numbers and previously unheard pieces. I eventually gave up and quit.

Mike

to dobroid:
Nice to hear that you enjoyed the survey and thank you for your time. I agree that long term structure can be relevant for music complexity judgements. However, it was not practical to provide complete tracks for this experiment, due to bandwidth and answering time issues. So we decided to keep the structure out in favour of a higher number of ratings per user.

to acclivity:
Thank you very much for your feedback. I'm sorry that you experienced these technical problems. Using the "back" and "forward" buttons is in general not completely safe with this survey. But the "song 34 of 30" problem is indeed a flaw in the programming, my apologies.

Sebastian

I took the survey too, it was interesting although I did get a bit impatient with it towards the end.
WEll worth taking though as it gets the braincells working a bit. It was also interesting to try and listen to the complexity of the music rather than its emotional impact, as I normally do.

grilla

A cool survey! One thing is missing and it is whether the Semantic Audio team will report back to its respondents with some results, if possible at all. At least I'm curious to know what they learned from this.