f

main.py

@@ -9,6 +9,8 @@ import uvicorn
 import threading
 import asyncio
 import logging
+import io
+import soundfile as sf
 from typing import List
 
 logging.basicConfig(level=logging.INFO)
@@ -76,8 +78,16 @@ def extract_features(audio):
     return combined_features
 
 async def process_audio_data(audio_data):
-    audio_np = np.frombuffer(audio_data, dtype=np.float32)
-    features = extract_features(audio_np)
+    try:
+        audio, sr = sf.read(io.BytesIO(audio_data), dtype='float32')
+    except RuntimeError as e:
+        logger.error(f"Failed to read audio data: {e}")
+        return
+
+    if audio.ndim > 1:  # If audio has more than one channel, average them
+        audio = np.mean(audio, axis=1)
+
+    features = extract_features(audio)
     features = features.reshape(1, -1)
     prediction = model.predict(features)
     is_fake = prediction[0]

templates/index.html

@@ -63,7 +63,7 @@
                 const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
                 logMessage('Microphone access granted', 'info');
 
-                mediaRecorder = new MediaRecorder(stream);
+                mediaRecorder = new MediaRecorder(stream, { mimeType: 'audio/webm' });
                 mediaRecorder.ondataavailable = function(event) {
                     if (websocket && websocket.readyState === WebSocket.OPEN) {
                         websocket.send(event.data);
@@ -96,6 +96,7 @@
             }
         }
 
+
         async function stopDetection() {
             if (!isDetecting) {
                 logMessage('Detection is not running...', 'info');