dsp/lib/smoothfft.py
2018-02-20 04:04:25 -08:00

60 lines
2.0 KiB
Python

from . import xsp, lament
import numpy as np
def smoothfft(xs, ys, bw=1, precision=512):
"""performs log-lin smoothing on magnitude data,
generally from the output of averfft."""
lament("smoothfft(): DEPRECATED; use smoothfft2 instead.")
xs2 = xsp(precision)
ys2 = np.zeros(precision)
log_xs = np.log(xs)
for i, x in enumerate(xs2):
dist = np.exp(np.abs(log_xs - np.log(x + 1e-35)))
window = np.maximum(0, 1 - (dist - bw))
# at this point we could probably
# normalize our *triangular* window to 0-1
# and transform it into *another* windowing function
wsum = np.sum(window)
ys2[i] = np.sum(ys*window/wsum)
return xs2, ys2
def smoothfft2(xs, ys, bw=1, precision=512, compensate=True):
"""performs log-lin smoothing on magnitude data,
generally from the output of averfft."""
# this is probably implementable with FFTs now that i think about it
xs2 = xsp(precision)
ys2 = np.zeros(precision)
log2_xs2 = np.log2(xs2)
for i, x in enumerate(xs):
# before optimizations: dist = np.abs(np.log2(xs2/(x + 1e-35)))/bw
dist = np.abs(log2_xs2 - np.log2(x + 1e-35))/bw
# window = np.maximum(0, 1 - dist) # triangular
window = np.exp(-dist**2/(0.5/2)) # gaussian (untruncated)
ys2 += ys[i]*window
if compensate:
_, temp = smoothfft2(xs, np.ones(len(xs)),
bw=bw, precision=precision, compensate=False)
ys2 /= temp
return xs2, ys2
def smoothfft3(ys, bw=1, precision=512):
"""performs log-lin smoothing on magnitude data"""
size = len(ys)
xs = np.arange(0, 1, 1/size)
xs2 = np.logspace(-np.log2(size), 0, precision, base=2)
ys2 = np.zeros(precision)
comp = np.zeros(precision)
log2_xs2 = np.log2(xs2)
for i, x in enumerate(xs):
dist = np.abs(log2_xs2 - np.log2(x + 1e-35)) / bw
window = np.exp(-dist**2 * 4) # gaussian (untruncated)
comp += window
ys2 += ys[i] * window
return xs2, ys2 / comp