Add mel spec visulaization

app.py

@@ -6,6 +6,7 @@ from transformers import AutoFeatureExtractor, AutoModelForAudioClassification
 import logging
 import sys
 import librosa
+import librosa.display
 import os
 
 logging.basicConfig(
@@ -83,10 +84,26 @@ with gr.Blocks(title="AST Model") as demo:
     def wrapper(audio_path):
         waveform, probs, id2label = predict_audio(audio_path)
 
-        fig, ax = plt.subplots(figsize=(10, 3))
-        ax.plot(waveform)
-        ax.set_title("Waveform")
+        mel_spec = librosa.feature.melspectrogram(
+                y = waveform, sr = AST_SR, n_mels = 128
+        )
 
+        mel_db = librosa.power_to_db(mel_spec, ref=np.max)
+        
+        fig, ax = plt.subplots(1, 2, figsize=(10, 3))
+        ax[0].plot(waveform)
+        ax[0].set_title("Waveform")
+
+        # mel spectrogram
+        img = librosa.display.specshow(
+            mel_db, sr = AST_SR, 
+            x_axis = 'time', y_axis = 'mel',
+            ax = ax[1]
+        )
+
+        ax[1].set_title("Mel Spectrogram")
+        fig.colorbar(img, ax=ax[1], format="%+2.0f dB")
+        
         label_dict = {
             id2label[i]: float(probs[i])
             for i in range(len(probs))