initialversion

app.py

@@ -0,0 +1,45 @@
+import gradio as gr
+import torch,torchaudio
+from PIL import Image
+import matplotlib
+import matplotlib.pyplot as plt
+import torchaudio.functional as F
+def plot_kaldi_pitch(mic_used,audio_path,mic_file):
+  SAMPLE_RATE = 16000
+  if(mic_used==False):
+   target_audio = audio_path
+  else :
+   target_audio = mic_file
+  wavform, sample_rate = torchaudio.load(target_audio)
+  pitch_feature = F.compute_kaldi_pitch(wavform, sample_rate)
+  pitch, nfcc = pitch_feature[..., 0], pitch_feature[..., 1]
+  figure, axis = plt.subplots(1, 1)
+  axis.set_title("Kaldi Pitch Feature")
+  axis.grid(True)
+  end_time = wavform.shape[1] / sample_rate
+  time_axis = torch.linspace(0, end_time,  wavform.shape[1])
+  axis.plot(time_axis, wavform[0], linewidth=1, color='gray', alpha=0.3)
+
+  time_axis = torch.linspace(0, end_time, pitch.shape[1])
+  ln1 = axis.plot(time_axis, pitch[0], linewidth=2, label='Pitch', color='green')
+  axis.set_ylim((-1.3, 1.3))
+
+  axis2 = axis.twinx()
+  time_axis = torch.linspace(0, end_time, nfcc.shape[1])
+  ln2 = axis2.plot(
+      time_axis, nfcc[0], linewidth=2, label='NFCC', color='blue', linestyle='--')
+
+  lns = ln1 + ln2
+  labels = [l.get_label() for l in lns]
+  axis.legend(lns, labels, loc=0)
+  plt.savefig("test.png")
+  image = Image.open("test.png")
+  return image , target_audio
+inputs = [gr.Checkbox(label="使用麥克風錄音嗎?"),gr.Audio(source="upload",type="filepath"),gr.Audio(source="microphone",type="filepath")]
+outputs = [gr.Image(type="pil"),gr.Audio(type="filepath")]
+demo = gr.Interface(fn=plot_kaldi_pitch,
+                    inputs=inputs,
+                    outputs=outputs
+                    )
+
+demo.launch(debug=False)