main.py

import numpy as np
from scipy.io import wavfile
import ctypes
from timeit import default_timer as timer
import matplotlib.pyplot as plt

num = 1
infile = f"./data/test{num}.wav"
outfile = f"./data/out{num}-f.wav"

fun = ctypes.CDLL("./bin/main")

fun.process.argtypes = [ctypes.POINTER(
    ctypes.c_int16), ctypes.c_uint32, ctypes.c_uint16]

samplerate, data = wavfile.read(infile)

flat = data.flatten("A")
# flat = (flat / (2**32)) * (2**15)
flat = flat.astype(np.int16)
flat_p = flat.ctypes.data_as(ctypes.POINTER(ctypes.c_int16))

len_c = ctypes.c_uint32(len(data))
sample_c = ctypes.c_uint16(samplerate)

acc = 0
count = 1
for i in range(count):
    start = timer()
    returnVale = fun.process(flat_p, len_c, sample_c)
    end = timer()
    acc += (end - start)

print(
    f"processed {2 * len(data)} samples {count} times in {round((acc) * 1000, 3)}ms")
print(f"{2 * len(data) * count} samples, with {round(((acc) * 1000 * 1000*1000) / (2*len(data)*count), 3)}ns per sample")

restored = flat.reshape(-1, 2)

wavfile.write(outfile, samplerate, restored)

print(data[:, 0])
print(restored[:, 0])

plt.figure(1)
plt.subplot(411)
plt.specgram(restored[:, 0], Fs=samplerate, scale="dB",
             mode="magnitude", cmap="rainbow")
plt.subplot(412)
plt.specgram(restored[:, 1], Fs=samplerate, scale="dB",
             mode="magnitude", cmap="rainbow")
plt.subplot(413)
plt.specgram(data[:, 0], Fs=samplerate, scale="dB",
             mode="magnitude", cmap="rainbow")
plt.subplot(414)
plt.specgram(data[:, 1], Fs=samplerate, scale="dB",
             mode="magnitude", cmap="rainbow")
plt.show()