How to resample a signal with SciPy resample_poly

Audio and sensor pipelines combine signals recorded at different rates. SciPy's resample_poly converts evenly spaced samples to a new sample rate using an up/down ratio and an antialiasing filter.

Passing the target rate as up and the input rate as down converts 800 Hz to 1200 Hz as a 3/2 ratio, so 200 input samples become 300 output samples over the same quarter-second interval.

The default boundary treatment assumes zeros beyond both ends, which can affect values near the edges. scipy.signal.resample() is a separate Fourier method that assumes a periodic signal and takes a target sample count; resample_poly takes integer upsampling and downsampling factors.

1. Create signal_resample.py with the sample rates and evenly spaced input signal.

   signal_resample.py

   import numpy as np
   from scipy.signal import resample_poly
    
   input_rate = 800
   target_rate = 1200
   duration = 0.25
    
   time = np.arange(0, duration, 1 / input_rate)
   samples = (
       np.sin(2 * np.pi * 40 * time)
       + 0.25 * np.sin(2 * np.pi * 120 * time)
   )

2. Append the polyphase resampling call to signal_resample.py.

   resampled = resample_poly(samples, up=target_rate, down=input_rate)

   resample_poly() applies its low-pass FIR filter during the 3/2 conversion. For a multidimensional array, axis= selects the dimension containing consecutive time samples.

3. Append a line that prints the input and output sample counts to signal_resample.py.

   print(samples.size, resampled.size)

4. Run signal_resample.py to confirm the sample count.

   $ python3 signal_resample.py
   200 300

   The 300 output samples at 1200 Hz cover the same quarter second as the 200 input samples at 800 Hz.