script to amplify voice tracks

scripts/amplify-voice-tracks.py

@@ -0,0 +1,48 @@
+#! /usr/bin/env python
+# pip install -r requirements.txt
+usage = 'python amplify-voice-tracks.py <cut json> <?wav filename>'
+
+from imports import *
+import util
+from time import sleep
+import string
+from os.path import exists
+from os import system
+system('color')
+
+json_filename = util.arg(1, usage)
+default_wav_name = json_filename.replace('.json', '.wav')
+wav_filename = util.arg(2, usage, default_wav_name)
+
+cutter = util.AudioCutter(wav_filename, json_filename)
+
+new_wav = wav_filename.replace(".wav", f"-amplified.wav")
+
+def save():
+    cutter.save_and_quit(new_wav)
+
+def process_chunk(audio_guess, timestamp):
+    audio, length = cutter.audio_and_length(timestamp['start'], timestamp['end'])
+    
+    audio_left = [channels[0] for channels in audio]
+    audio_right = [channels[1] for channels in audio]
+
+    def max(l):
+        m = 0
+        m_abs = 0
+        for v in l:
+            if abs(v) > m_abs:
+                m = v
+                m_abs = abs(v)
+        return m
+    ml = max(audio_left)
+    mr = max(audio_right)
+
+    MAX = 32767
+
+    can_multiply = abs(min(MAX/ml, MAX/mr))
+    print(can_multiply)
+    cutter.take_audio(audio_guess, {'start': cutter.current_sec, 'end': cutter.current_sec + length}, timestamp['start'], timestamp['end'], can_multiply)    
+
+cutter.process_audio(process_chunk, new_wav)
+save()

scripts/util.py

@@ -63,8 +63,15 @@ class AudioCutter:
         end_frame = int(end * self.framerate)
         return self.data[start_frame:end_frame], end - start
 
-    def take_audio(self, tag, info, start, end):
+    def take_audio(self, tag, info, start, end, amplify_by=1.0):
         audio, length = self.audio_and_length(start, end)
+        if amplify_by != 1.0:
+            for idx in range(len(audio)):
+                audio[idx][0] *= amplify_by
+                audio[idx][0] = int(audio[idx][0])
+                audio[idx][1] *= amplify_by
+                audio[idx][1] = int(audio[idx][1])
+
         self.new_data = vstack((self.new_data, audio))
         self.current_sec += length
         self.new_json_info[tag] = info