how to normalise scipy.signal.correlate output to be between -1 and 1

Question

Does anyone know how to normalise the output of scipy's signal.correlate function so that the return array has numbers between -1 and 1. at the moment its returning numbers between -1 and 70000.

Answer 1

AFAIK scipy.signal.correlate does not have an option for auto normalize, however you can easily normalize the signal yourself:

import numpy as np
def normalize(tSignal):
      # copy the data if needed, omit and rename function argument if desired
      signal = np.copy(tSignal) # signal is in range [a;b]
      signal -= np.min(signal) # signal is in range to [0;b-a]
      signal /= np.max(signal) # signal is normalized to [0;1]
      signal -= 0.5 # signal is in range [-0.5;0.5]
      signal *=2 # signal is in range [-1;1]
      return signal

And more general function, normalizing a vector to range [a,b]:

import numpy as np
def normalize(signal, a, b):
      # solving system of linear equations one can find the coefficients
      A = np.min(signal)
      B = np.max(signal)
      C = (a-b)/(A-B)
      k = (C*A - a)/C
      return (signal-k)*C