Spaces:
Running
on
Zero
Running
on
Zero
File size: 2,140 Bytes
4efe6b5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
import librosa
import gradio as gr
import os
from matplotlib import pyplot as plt
from rvc.lib.predictors.F0Extractor import F0Extractor
from assets.i18n.i18n import I18nAuto
i18n = I18nAuto()
def extract_f0_curve(audio_path: str, method: str) -> tuple:
print("Extracting F0 Curve...")
image_path = os.path.join("logs", "f0_plot.png")
txt_path = os.path.join("logs", "f0_curve.txt")
y, sr = librosa.load(audio_path, sr=None)
hop_length = 160
librosa.note_to_hz("C1")
librosa.note_to_hz("C8")
f0_extractor = F0Extractor(audio_path, sample_rate=sr, method=method)
f0 = f0_extractor.extract_f0()
plt.figure(figsize=(10, 4))
plt.plot(f0)
plt.title(method)
plt.xlabel("Time (frames)")
plt.ylabel("Frequency (Hz)")
plt.savefig(image_path)
plt.close()
with open(txt_path, "w") as txtfile:
for i, f0_value in enumerate(f0):
frequency = i * sr / hop_length
txtfile.write(f"{frequency},{f0_value}\n")
print("F0 Curve extracted successfully!")
return image_path, txt_path
def f0_extractor_tab():
with gr.Accordion(label=i18n("Extract F0 Curve")):
with gr.Row():
audio = gr.Audio(label=i18n("Upload Audio"), type="filepath")
f0_method = gr.Radio(
label=i18n("Pitch extraction algorithm"),
info=i18n(
"Pitch extraction algorithm to use for the audio conversion. The default algorithm is rmvpe, which is recommended for most cases."
),
choices=["crepe", "fcpe", "rmvpe"],
value="rmvpe",
)
button = gr.Button(i18n("Extract F0 Curve"), variant="primary")
with gr.Accordion(label=i18n("Output Information")):
txt_output = gr.File(label="F0 Curve", type="filepath")
image_output = gr.Image(type="filepath", interactive=False)
button.click(
fn=extract_f0_curve,
inputs=[
audio,
f0_method,
],
outputs=[image_output, txt_output],
api_name="f0_extract",
)
|