将 WAV 格式从 32 位浮点数更改为 16 位 PCM

Question

I'm trying to convert my 32-bit floating point sampled WAVE file to 16-bit integer WAVE file.

I want to generate 16-bit integer WAV file from ultrasonic sine wave which range from -1.414 to 1.414. In SciPy.io.wavfile , the output format is fixed according to type array which I generated. But I changed my 32-bit float format array to 16-bit by multiplying '32767' with casting to get integer format value, but the result frequency is not similar at all.

Is there any way to change my from 32-bit floating point to 16-bit PCM using SciPy library?

Answer 1

Not sure about using SciPy for this but you could easily get your data into a 16bit integer and then use what ever audio package you like to write it to a wav file.

Divide your data by 1.414 so that it is in the range [-1, 1], then multiply it by 32767 so that it is within the range of an int16 variable. Then cast it using numpy.astype() . The solution would look something like this:

import numpy as np

float32_data /=1.414
float32_data *= 32767
int16_data = float32_data.astype(np.int16)