pythonpython-3.xlibrosa

How to downcast sample rate of ndarray (without saving it as wav file)

I have a ndarray (which I got from microphone) with samples (sample rate = 48000).

I want to create a new ndarray with lower sample rate (16000).

One option is to save this ndarray as wav file and reload it with different sample rate (via librosa).

Is there a simple way to do it without saving the ndarray as wav file ?

Solution

  • there's a method resample in librosa and you can use it without saving the file

    import librosa
original_data, sr = librosa.load('0 .wav', sr=None)
print(original_data.shape)
resampled_data = librosa.resample(original_data, orig_sr=sr, target_sr=16000)
print(resampled_data.shape)