eddie-water / lightbeats

1 stars 0 forks source link

Implement the continuous wavelet transform #3

Open eddie-water opened 2 months ago

eddie-water commented 2 months ago

Created a mixed signal that has three different frequency signals present but with different amounts of time support. Plotted the time series and the continuous wavelet coefficients as a function of frequency and time. Image

eddie-water commented 1 month ago

I've been stuck on something. I want to try and use the Blit Manager to draw the time-freq coefficients of the pywavelets CWT function. In order to be blitted, matplotib says the data structure should be of type Artist. In order to plot just a single frame of the audio, I have had success using the pcolormesh object, but I don't think it is compatible for blitting. At least my first attempt didn't work. So I've been trying to use other objects like meshgrids and imshow types from matplotib, and supposedly they should work? By now I've tried so many different things that I my code is all muddy, and I was being bad and haven't done a good job of committing on the way. I also have examples that work, but I'm not exactly sure how the examples work but mine doesn't. It's prob some small little dimension, size, type, structure or etc that's just different but I can't see which one of these attributes of the one object is the incompatible culprit.

Going to take a break. But when I come back I'm going to replicate this example written by a dude who literally works for matplotlib https://stackoverflow.com/questions/40126176/fast-live-plotting-in-matplotlib-pyplot/40139416#40139416