Add audio processing features: implement silence removal and audio length calculation

Author: Jared-Gross

audio_file.py

@@ -0,0 +1,8 @@
+from pydub import AudioSegment
+
+
+def get_audio_length(file_path: str) -> float:
+    file_format = file_path.split(".")[-1]
+    audio = AudioSegment.from_file(file_path, format=file_format)
+    duration_minutes = len(audio) / (1000.0 * 60.0)  # milliseconds to minutes
+    return duration_minutes

main.py

@@ -18,9 +18,11 @@
 from tornado.ioloop import IOLoop
 from tornado.web import Application, RequestHandler
 
+import audio_file
 import filebrowser_uploader
 import firebase_android_notification
 import firebase_web_notification
+import remove_silence
 import send_email
 import zip_file
 
@@ -168,8 +170,10 @@ def convert_delta_time(self, duration: timedelta) -> str:
         )
 
     def process_file(self):
+        self.backup_stream(self.recording_file_name)
         app_log.info(f"Processing {self.recording_file_name}")
-        self.audio_file_length = self.get_recording_time_minutes()
+        remove_silence.remove_silence_everywhere(self.recording_file_path)
+        self.audio_file_length = audio_file.get_audio_length(self.recording_file_path)
 
         time_delta = timedelta(minutes=self.audio_file_length)
         final_delta_time = self.convert_delta_time(time_delta)
@@ -241,9 +245,16 @@ def delayed_notification():
                 )
                 app_log.info(f"Notification sent for {self.host}")
 
+        email_body = f"""
+            URL: {self.url}<br>
+            Title: {self.title}<br>
+            Description: {self.description}<br>
+            Host: {self.host}<br>
+            Date: {self.starting_time.strftime('%B %d, %A %I:%M %p')}<br>
+        """
         send_email.send(
             f"{self.title} just started a stream!",
-            f"{self.description}<br>{self.url}",
+            email_body,
         )
 
         if "test" in self.title.lower():
@@ -318,7 +329,7 @@ def run(self):
                     is_recording = source.get("genre", "various") == "RECORDING"
 
                     if (
-                        host not in self.active_streams and "test" not in host.lower() and "test" not in description.lower() and not is_private
+                        host not in self.active_streams
                         # and not is_recording # It is being recorded by HBNI Audio
                     ):
                         stream = Stream(

remove_silence.py

@@ -0,0 +1,29 @@
+from pydub import AudioSegment, silence
+
+
+def remove_silence_everywhere(file_path: str, silence_thresh: float = -40.0, min_silence_len: int = 1000, keep_silence_ms: int = 500):
+    print(f"Removing silence inside {file_path}")
+    file_format = file_path.split(".")[-1]
+    sound = AudioSegment.from_file(file_path, format=file_format)
+
+    chunks = silence.split_on_silence(
+        sound,
+        min_silence_len=min_silence_len,     # only cut silences longer than 500ms
+        silence_thresh=silence_thresh,       # treat anything below -35dBFS as silence
+        keep_silence=keep_silence_ms         # leave 200ms of silence at each cut
+    )
+
+    if not chunks:
+        print("No sound detected above silence threshold. Skipping.")
+        return
+
+    combined = AudioSegment.empty()
+    for chunk in chunks:
+        combined += chunk
+
+    combined.export(file_path, format=file_format)
+    print(f"Silence removed inside and saved to {file_path}")
+
+if __name__ == "__main__":
+    file_path = r"CURRENTLY_RECORDING\Test - Pineland - April 25 Friday 2025 09_14 PM - 0m 19s.mp3"
+    remove_silence_everywhere(file_path)