Mathieu,
Thank you for your well informed response.
I'll be sure to follow up on #dataflow IRC to ask more of your subjective, experienced input.
./d5
On Jan 30, 2006, at 12:48 AM, Mathieu Bouchard wrote:
On Sat, 28 Jan 2006, day 5 wrote:
I was wondering in what ways is it possible to calculate Mandlebrot sets using Gem/Pd/PDP/GridFlow/etc/etc ?
In GridFlow, the technique I use is to apply a pixel-remapping repeatedly using [#remap_image]. The big downside of it is that rounding errors will show because pixel coördinates are always whole. However there are several upsides, one being that it's incremental and that each increment is rather quick to compute. Classic Julia/Mandelbrot fractals are usually computed by repeating a formula on each point until the point gets kicked out of a circle, and then counts how many repetitions it took. I use one repetition per displayed frame, and the picture refines itself over time. At each increment I can feed in something different, e.g. I may inject live video into the fractal.
You can find out about it in gridflow/pd_examples/feedback_fractal.pd, which was the first real GridFlow patch. (I have other more convincing fractal patches but I've not released them.)
My technique is neither Julia nor Mandelbrot, even though it uses the same formula f(z)=z*z+c, because it uses it in a different way.
AFAIK Ben Bogart uses the same technique and/or the same formula but with GEM. However I haven't closely compared.
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801 - http://artengine.ca/matju | Freelance Digital Arts Engineer, Montréal QC Canada