silence timestamps

babelbox_voice.py

@@ -9,6 +9,10 @@ from pathlib import Path
 from tqdm import tqdm
 import torchaudio
 from io import BytesIO
+import pydub
+from pydub import AudioSegment
+from pydub.silence import detect_leading_silence
+import numpy as np
 logger = datasets.logging.get_logger(__name__)
 
 _CITATION = """\
@@ -149,6 +153,28 @@ class BabelboxVoice(datasets.GeneratorBasedBuilder):
            
             sampling_rate = 16000
             
+            def get_audiosegment(array):
+                byte_array = np.int16(array * 2 ** 15).tobytes()
+                audio_segment = pydub.AudioSegment(byte_array, frame_rate=16000, sample_width=2, channels=1)
+                return audio_segment
+            
+            def get_leading_and_trailing_silence(audio_array):
+                audio_segment = get_audiosegment(audio_array)
+                
+                leading = detect_leading_silence(audio_segment)
+                trailing = detect_leading_silence(audio_segment.reverse())
+                return leading, trailing
+            
+            def get_timestamp(audio_array):
+                leading_silence, trailing_silence = get_leading_and_trailing_silence(audio_array)
+                start_len = leading_silence / 1000
+                start_time = round(start_len / 2, 2) * 2
+                
+                end_len = (len(audio_array) / sampling_rate) - (trailing_silence / 1000)
+                end_time = round(end_len / 2, 2) * 2
+                return (start_time, end_time)
+                
+            
             def filter_sentence(sentence):
                 if "... tyst under denna inspelning ..." in  sentence: return False
                 return True
@@ -172,10 +198,12 @@ class BabelboxVoice(datasets.GeneratorBasedBuilder):
                     audio_len = len(audio_array) / sampling_rate
                     audio_len = round(audio_len / 2, 2) * 2
                     
+                    print(get_leading_and_trailing_silence(audio_array))
+                    
                     text = {
                         "text": metadata_item['sentence'],
                         "offsets": [ 
-                            {"text": metadata_item['sentence'], "timestamp": (0.0, audio_len) }
+                            {"text": metadata_item['sentence'], "timestamp": get_timestamp(audio_array) }
                         ]
                     }