User Tools

Site Tools


tracy444a:todo

TODOS

— June 14, 2011

— March 11, 2011

  • I'd like a palette of expansion controls, such as transpose invert and so on. How they interact with the current expansion requires Deep Thought
  • Serialize Sequences: add a control section called SAVE, button for GEN, button for EXP, to save any given sequence to a file
  • Obviously (but at )less high priority) is reading them back in, but yeah LOAD a GEN or EXP
  • Eventually an output format could be massaged into a form that could be read by other MIDI players
  • Eventually++ the system could be used without exp audio, but producing gobs of expansions that can be played/analyzed at leisure

— March 4, 2011

  • Framework for specifying a scale (root, mode, maj0r/minor) at constructor time and eventually startup
  • Comp tool based on grid idea for specifying initial seed melody (option)
  • Can/should user tell grid that we're in minor/pentatonic/etc?
  • Where the are the modes in impromptu's Scheme code?

—- November: SEMESTER PROJECT

<Nels> I did have a thought about this, on the Processing side you can change the code so that it only sends the pitch info etc, when the size changes, and that would be an interesting way to generate note lengths. The other major issue you address is the tracking of individual blobs. I've been wondering about how best to handle this, and I'm not sure of the best way since openCV does some strange things with the numerical heading on each blob. If you perhaps tracked the position of a blob for a limited amount of time you could over come at least part of this issue. On the pd side, it should be just an issue as you say, of handling each osc message as it is passed from processing.

This week we updated Nels' CVBlobber to use the Pd implementation of MIDI Synth from Miller Puckette's book. We identified the following problems:

  • The synth takes its pitch information from the center point of each movement contour (blob), but the contours are presented sequentially in index order, so pitches do not conform to movement directly. How might this information be preserved and presented in some coherent manner?
  • One possibility is to use a different voice for each contour. Study up on subpatches/externals to see a nice way to do this (Ch 4 in Puckette's book, for example). This opens up a new set of problems, however. For example, are each frame's contours guaranteed to have the same index each time through the draw loop?
  • Duration information does not seem to be computed. Each note is approximately the same length.
  • Look into using buffering to preserve pitch information across, say, a few seconds at a time.
  • (NTS: complete study of Blob class and contour recognition)

—- Sometime last week (let's just say Oct. 13)

  • Are Henon maps mentioned in the music/fractal book?

I found a PhD thesis that covers some algorithmic music systems from fractals as well. I should post a link to that soon.

—- Oct. 7, 2010

  • Continue with the Pd study. Remember the idea of changing the octaves to overtone series in Evil But Cool, in fact you can probably try this out first in Processing without much trouble. We'll spend some time on connecting it to Pd later.
  • Study the four Golan Levin computer vision algorithms presented in the Processing book (Extension 3: Vision), pages 547-561. Try them and the other Video (Capture) examples in WL 119.
  • Read and give a brief/summary critique of the 3 GA papers (and 1 AlgComp) paper. It would be nice to have one per week or better.

—- Sept. 23, 2010

  • How is the experiment using self-similar tinted rectangles containing an image like, and how is it unlike, a Sierpinski carpet? What would have to change (if anything) to make it a true fractal? Post a writeup on this kind of fractal.
  • Study the Pd tutorial with a view toward using it to generate tones a la the Evil but Cool version, moving beyond single sine waves. Figure out a way to generate an overtone series using Pd; eventually we'd like to implement fractals in music this way.
  • Soon we'll into connecting data generated by Processing sketches to Pd as pitch and/or control values.

—- Sept. 16, 2010

  • Work through the code for DrawYerself (PainterV1), etc.
tracy444a/todo.txt · Last modified: 2011/06/14 21:20 by scarl