Too much time on my hands, and theirs.

This is easy to do with gnuplot and is done with a simple sequence of operations. The goal is to create some images to be imported into gimp and analyzed with a script that I am creating. What is done is to create the fourier transform of a signal then convert that to specific sine waves and compute the "mass" or length of the waves by summing pixels. (Number of dots)= length of curve. I was trying to see if there were some easy unusual transforms that would provide analytical information about a signal without performing a de-constructive analysis.

sudo apt-get install gnuplot

Below is the file saved as "gnup.plot"

set xrange [-pi/2:pi/2]; set yrange [-pi/2:pi/2] set contour set isosamples 100,100; set samples 100,100 splot sin(x)*cos(y) with lines palette linewidth 4 pause -1

It is invoked by the command line :

gnuplot gnup.plot

That is just the start of it and I will generate sine waves, open them as layers, merge and then run my script to analyze them. It seems to me that a fourier analysis implies a certain amount of perfection that is never present in the world. Any system that oscillates would be damped to some degree and the most common waveform that is likely would be a distorted sine of continually decaying amplitude.

## 0 comments:

Post a Comment