Impro-Visor

[MikeRobinson]

Looks quite interesting and useful . . . what do you think?

http://www.cs.hmc.edu/~keller/jazz/improvisor/

http://en.wikipedia.org/wiki/Impro-Visor

[Lazz]

what do you think?

Hard to say while I struggle with deep-seated prejudice here.

Impossible to answer without a long test-drive, but these things worry me.

The web-site describes the purpose of the software as "to help jazz musicians compose and hear solos similar to ones that might be improvised".

My riposte is that if they can't compose and hear solos then how can we call them jazz musicians ?

Site also describes the objective as "to improve understanding of solo construction and tune chord changes" - which is something usually achieved by listening, by analysis, by transcription.... and that style of personal discovery process seems to be, to me, of most real fundamental importance in jazz education. The rest is done through interacting with your instrument and with other musicians - not with a machine. So I don't get the point of having a computer program improvise on our behalf. That just ain't what it's about.

See - I told you there was prejudice involved - and without trying the damn thing (do I have time?) my criticisms are only really validly aimed at the words on the web-site.

There must be good points involved though, I'm sure - for a start, it's free - which seems like a good deal.

There is also the suggestion that it might serve equally well as Band In A Box does - but then I know the guys who built that baby and why while I don't know any of the names attributed to this Impro-Visor piece of kit (sounds like a form of head-gear) and am immediately suspicious of its actual heuristic value............... as I said, there is some personal prejudice involved.

What do you think of it, Mike ?

Have you taken time to test it ?

.

[MikeRobinson]

Well, since it is written by a professor of music (and since I am, at heart, a computer-science geek ) I do intend to have a look-see at what it might have to offer. I don't perceive that the real thrust of this software is to be "Band-in-a-Box."

Maybe, such a tool is meant partly to study what can be done algorithmically. It's like the people who work with chess-playing software in order to study how humans play chess. (One of the grand-masters of chess is working with IBM on that.) I find that interesting, although I myself do not play chess. In a sense, the computer is made to produce a mirror against which we find our own selves reflected. When we construct such a thing, we notice what is missing or different. Our brains are especially good at spotting differences (any one of which could be a tiger in the grass, that's ready to eat you).

The folks who seriously expect "a band in a box," or that they can buy musical genius from Microsoft, probably also think that a hayseed farm-boy can impress an Italian supermodel just by buying a piece of software that calls itself Rosetta Stone. But then again, there's not much to be said or done about attitudes like that... except, of course, to sell them something. I don't think that this is what these folks had in mind. (And BTW, I also don't mean to sell-short a commercial tool that I know many people find to be useful.)

It also seems to me that they chose well, and deliberately, when they selected jazz improvisation as their field of study. This is extemporaneous composing, never the same thing twice and intended so to be. A computer would have a much better chance of doing that because it would largely consist of stringing-together riffs. So, they seem to be confining their "scope of work" carefully.

As I listen to their sample song (which, admittedly, has four or more alternate endings back-to-back), the repetition shows. The notes change (although in limited ways); the rhythm and pacing doesn't. And so on.

But, could a songwriter profitably learn from that anyway? Would it bust people out of writer's block? Would it introduce new ideas (taken from a library of riffs such as these programs have)?

Indeed, might it do so in a different way precisely because of what the digital computer can uniquely bring to the tale? The user can instantly both hear and see (simultaneously and in real-time) something that is, literally, "altogether new and unexpected." I'm quite intrigued to think what the actual impact of such an experience might be, say upon a serious music student or a young child (who is also serious about music).

Would it in any way substitute for the very valid things that you say? I think, absolutely not. Nevertheless, here is something that a very readily available piece of digital hardware can do, which, one might well argue, nothing else can do. How could such a tool, with such novel properties, be used in the context of music (self-?) education? I personally would love to find something that a computer can do for young people besides spoon-feed them expensive musical fantasies . . . (Pshaw! How can you call yourself a "Guitar Hero" when you can't even play the damm thing?)

Edited February 19, 2010 by MikeRobinson

[BigBubbaBrown]

I gave it a try. I'll probably mess around with it a little more, but it's not really worth much from what I've seen. On the surface it seems like it could allow you to learn about various styles, such as bluegrass and zydeco if you've never really worked with them, but in practice it's not even suitable for that. It does swing pretty well and just fakes the rest.

I created a new document, saw Irish jig was one of the styles, so I set the time signature to 6/8, the key signature to D, and entered a few chords. They were random so I knew it wasn't a proper progression, but it didn't sound anything like a jig. I'm talking about the aspects that you can't control like the rhythm and phrasing. Didn't even remotely sound like any type of jig I've heard.

I might explore the other styles, but I'm not familiar with a lot of them (such as two step, zydeco, and the like) so I don't know if I'll even be getting an accurate interpretation to learn about them using this software based on my experiences with the jig.