karlmac at peabody.jhu.edu
Tue May 1 18:51:31 CEST 2001
You can also use my knn object:
It uses the K-nn algorithm to match vectors. Here is a paper about the
use of the object for realtime timbre recognition:
On Tue, 1 May 2001, Miller Puckette wrote:
> Hi Evan,
> You might find the "choice" object useful. It's an extern in the "extra"
> directory of the Pd release. The basic idea is you give it a bunch of
> vectors (expected behaviors) in advance. Then you give it a real "behavior"
> from a visitor (for instance) and it tells you which if the pre-stored
> ones is closest to it. So you can use it for a simple classifying tool.
> On Tue, May 01, 2001 at 03:24:23AM -0700, Evan Sugerman wrote:
> > Greet.
> > I am using IR sensors [of different types] -> feeding 12V relays ->
> > creating contact, of individual keys, on a disassembled MIDI keyboard. So
> > far things are quite well. This is my first attempt at programming of any
> > sort so my resources are self generating (the list has been helpful for the
> > most part).
> > My question is in regards to the idea of an array and its relationship to
> > the 'tabwrite' and 'tabread' objects. I am hoping to to generate a
> > reservoir of the information gathered by a certain proximity sensor i am
> > using (increments vs. time between increments),plus some other
> > idiosyncratic in formation which is not relavent to my question, and store
> > in hopes of using it to make Pd make decisions. Is this too simple an
> > application for the array? I am having trouble figuring out how to get
> > separate x y (z?) information into and out of it. My need is to generate
> > different response mechanisms depending on the type of movement detected
> > (agitation level) and to do so by subtly altering certain characteristics
> > of the system. These alterations will be modeled after an aggressive
> > animal system.
> > I have began to use the msgfile and i think i am beginning to understand
> > its application........a little.
> > I re-read my question; it seems unclear. At this point with my current
> > knowledge on such programmatic logistics I cannot do any better. I hope
> > someone has some idea of what i think i am trying to ask.
> > bye.evan.
| Karl W. MacMillan |
| Computer Music Department |
| Peabody Institute of the Johns Hopkins University |
| karlmac at peabody.jhu.edu |
| www.peabody.jhu.edu/~karlmac |
More information about the Pd-list