File size: 781 Bytes
a8daa4b b9442ae c065b14 87d610a b9442ae 027cd67 6437819 027cd67 abca20d 027cd67 b5b432f d6d39f0 fcecdda daab81b b5b432f 7fdcdff b90c087 027cd67 b90c087 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 |
import gradio as gr
import numpy as np
import librosa
from scipy.io.wavfile import write
# import the librosa library for converting a Mel spectrogram image to audio
def mel_to_audio(mel_spectrogram):
# convert the Mel spectrogram image to a grayscale image to avoid the string indices must be integers error
# convert the Mel spectrogram image to audio using librosa
# specify the data type of the output array to avoid the same_kind casting error
audio = librosa.feature.inverse.mel_to_audio(mel_spectrogram, dtype=np.float64)
# Save the audio to a file
output = 'output.wav'
sample = 44100
print (output)
write(output, sample, audio.astype(np.float64))
return output
# create the gradio app
app = gr.Interface(mel_to_audio, "image", "audio")
app.launch() |