mirror of
https://github.com/gradio-app/gradio.git
synced 2024-12-15 02:11:15 +08:00
52 lines
1.2 KiB
Python
52 lines
1.2 KiB
Python
# Demo: (Audio) -> (Label)
|
|
|
|
import gradio as gr
|
|
import numpy as np
|
|
from scipy.fftpack import fft
|
|
import matplotlib.pyplot as plt
|
|
from math import log2, pow
|
|
|
|
A4 = 440
|
|
C0 = A4*pow(2, -4.75)
|
|
name = ["C", "C#", "D", "D#", "E", "F", "F#", "G", "G#", "A", "A#", "B"]
|
|
|
|
def get_pitch(freq):
|
|
h = round(12*log2(freq/C0))
|
|
n = h % 12
|
|
return name[n]
|
|
|
|
def main_note(audio):
|
|
rate, y = audio
|
|
if len(y.shape) == 2:
|
|
y = y.T[0]
|
|
N = len(y)
|
|
T = 1.0 / rate
|
|
x = np.linspace(0.0, N*T, N)
|
|
yf = fft(y)
|
|
yf2 = 2.0/N * np.abs(yf[0:N//2])
|
|
xf = np.linspace(0.0, 1.0/(2.0*T), N//2)
|
|
|
|
volume_per_pitch = {}
|
|
total_volume = np.sum(yf2)
|
|
for freq, volume in zip(xf, yf2):
|
|
if freq == 0:
|
|
continue
|
|
pitch = get_pitch(freq)
|
|
if pitch not in volume_per_pitch:
|
|
volume_per_pitch[pitch] = 0
|
|
volume_per_pitch[pitch] += 1.0 * volume / total_volume
|
|
return volume_per_pitch
|
|
|
|
iface = gr.Interface(
|
|
main_note,
|
|
"audio",
|
|
gr.outputs.Label(num_top_classes=4),
|
|
examples=[
|
|
["audio/recording1.wav"],
|
|
["audio/cantina.wav"],
|
|
],
|
|
interpretation="default")
|
|
|
|
if __name__ == "__main__":
|
|
iface.launch()
|