change Audiosegment to ndarray for librosa to add effect

[bzvew]

I wonder how to change Audiosegment to float32 ndarray for librosa to manipulate.

[jiaaro]

Have you tried using ‘AudioSegment(…).get_array_of_samples()`?

[AbbyYangbb]

You can get "array.array"object by using AudioSegment(…).get_array_of_samples(), for example
,
ar = chunk.get_array_of_samples()

and if you want to get numpy.ndarray, try:
import numpy as np
ar = np.array(ar)

[zhenql]

You can get "array.array"object by using AudioSegment(…).get_array_of_samples(), for example
,
ar = chunk.get_array_of_samples()

and if you want to get numpy.ndarray, try:
import numpy as np
ar = np.array(ar)

the type of data x is int and |x| may more than 1, but the type of librosa is float32 and less than 1.
how to use x generated from pydub in librosa?

[mattc-eostar]

And how to change back to AudioSegment?

[jiaaro]

@mattc-eostar That's covered in the docs:

from pydub import AudioSegment
sound = AudioSegment.from_file("sound1.wav")

samples = sound.get_array_of_samples()

# then modify samples...

new_sound = sound._spawn(samples)

[mattc-eostar]

Ah, perfect. Thank you!