Interesting story. I’m wondering at what point he decided he wanted to have 35,000 tunes written before publishing any of them.
I’d have thought it more worth his while to put them out there now in batches, maybe 1,000 tunes per volume, while he’s still alive to hear them played (and to reap some of the royalties).
That could very well be true. I have a friend who went to a lecture he gave some years ago on this topic, and that’s my only source for my information. Sorry, and thanks for keeping me honest.
I’m not sure how well MacMahon knows Vallely’s music either which was my main problem with his comment. Mind you it was a few years ago and the comment may have been aimed at what MacMahon viewed as a whole raft of young musicians who he perceived as playing too fast rather than Niall’s playing in particular. To direct it at Niall Vallely (or any specific musician) was a mistake however. Anyone who thinks he plays without soul and musicality either isn’t listening hard enough or doesn’t know what they are talking about. It might have been the fact that MacMahon just couldn’t get look past the tempo. Another thing one has to be careful of when criticising a fellow musician is that there is no doubt that the person doing the criticizing has the same level of technical alibility as the subject. Otherwise the comments may be misconstrued as sour grapes.
But I didn’t mean to hijack the thread. It was an interesting enough topic without dragging the colourful MacMahon character into the mix.
Actually, looking further at it (I had no idea he was lecturing on this stuff), it looks like maybe it isn’t his research but he has lectured and written about it. Very interesting stuff, especially in context of the current discussion. Thanks for the tip!
Hello there,
I’m Bryan Duggan, the author of the paper in question. The paper was an abstract for a proposed PhD. In fact I ended up working on something kind of similar, but related. I.e. trying to develop a system that could learn “musical style” of flute players. I have had mixed results so far, though I’ve only been working on it for the first year. I anticipate this time next year I should have some better results to report. This paper (recently presented at the ECAI Computational Creativity workshop) gives a summary of my progress:
http://www.comp.dit.ie/bduggan/Research/MATT.pdf
Cheers,
Bryan
Hi Bryan;
thanks for posting the info on the current project. Maybe you could let us know when you have completed your work. I’d be very interested to hear how you get on.
I just heard a really interesting radio show that focused on different aspects of music and how it touches us.
In the show, they gave samples of a composer who had written an intellingent (or non-intelligent, if you ask him) program that analyzes the music of a composer, and then writes more in the style.
http://www.wnyc.org/shows/radiolab/episodes/2006/04/21
Scroll down to the “Musical DNA” section to hear some of what EMI (“Emmy”) has written. Very interesting!
Ok,
I’ve eventually got around to doing something about this idea and in fact a student of mine this year has developed some software that “learns” how to compose reels by analysing a corpus of about 800 tunes. I have had the program compose some tunes and I am looking for people to help evaluate the tunes for me. The tunes need be evaluated by:
To make the survey more interesting, one of the tunes was composed by a human rather than by the software and after listening to all the tunes,
you have to try and guess which one was composed by a human.
I need at least 10 responses for the survey to be valid, and I need them
within the next couple of days, so feel free to pass on this email to
any other musicians you know.
You can download all the tunes in midi format here:
http://www.bryanduggan.com/nan/MIDI.zip
And you can access the survey (and the tunes) here:
http://www.bryanduggan.com/phpESP/public/survey.php?name=GeneratedTunes
It uses quite an interesting technique and we’ll be reporting on it in a paper for AICS. I’ll put a link up to the paper when its written if anyone is interested in learning more.
Many thanks!!
Done.
Interesting stuff. Well done.