Maybe start with your x values somewhere negative, since exp(0) = 1. In your graph it looks like around 0.6 would be 0, so subtract 0.6 from x, looks like maybe 1000*exp(x-0.6).
Martin
On Wed, Oct 22, 2014 at 2:03 PM, Dan Wilcox danomatika@gmail.com wrote:
Howdy all,
I’m *almost* done with my rjlib compatible s_minimoog abstraction for the rc-patches, but I need help finding the right curve for the ADSR controls. I used block diagrams and the Minimoog Service Manual to build the dsp chain. What I have now can give a *reasonably* accurate rendition of Minimoog patches from sheets found online (ELP Aquatarkus, for instance). What I did notice in comparing it to the Arturia Mini V AudioUnit minimoog simulation, is that the knob curves for frequency are different form what I’ve been able to do, so some patches aren’t quite *right* yet.
Basically, I’ve tried a number of different exponential and power curves but can’t get one that has the long, low slope on the low end. All of my curves rise too quickly. In any case, this is a curve fitting problem and this is the data I’m working with. The mapping I want is 0 - 1 to 20 0 20kHz. And no, I don’t to just stick a log hsl in there :D
Any help would be appreciated so I can get this out there and start using it!
pos freq 0.000 0 0.125 83 0.250 362 0.375 456 0.500 768 0.625 1750 0.750 6300 0.875 12750 1.000 25000
Dan Wilcox @danomatika danomatika.com robotcowboy.com
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list