I recorded some audio with
sd.rec(frames=int(duration * fs), samplerate=fs, channels=2, dtype='float32')
This returned a numpy array of dtype np.float32.
When recording some clapping (i.e. a fairly loud noise), min and max where around -14.0 and 14.0 respecively. What format is this? It does not appear to correspond to any format listed in scipy.io.wavfile.read. I need to know the range so I can scale this to [-1, 1].