/usr/lib/python3/dist-packages/pydub/silence.py is in python3-pydub 0.20.0-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 | from .utils import (
db_to_float,
)
def detect_silence(audio_segment, min_silence_len=1000, silence_thresh=-16):
seg_len = len(audio_segment)
# you can't have a silent portion of a sound that is longer than the sound
if seg_len < min_silence_len:
return []
# convert silence threshold to a float value (so we can compare it to rms)
silence_thresh = db_to_float(silence_thresh) * audio_segment.max_possible_amplitude
# find silence and add start and end indicies to the to_cut list
silence_starts = []
# check every (1 sec by default) chunk of sound for silence
slice_starts = seg_len - min_silence_len
for i in range(slice_starts + 1):
audio_slice = audio_segment[i:i+min_silence_len]
if audio_slice.rms < silence_thresh:
silence_starts.append(i)
# short circuit when there is no silence
if not silence_starts:
return []
# combine the silence we detected into ranges (start ms - end ms)
silent_ranges = []
prev_i = silence_starts.pop(0)
current_range_start = prev_i
for silence_start_i in silence_starts:
continuous = (silence_start_i == prev_i + 1)
# sometimes two small blips are enough for one particular slice to be
# non-silent, despite the silence all running together. Just combine
# the two overlapping silent ranges.
silence_has_gap = silence_start_i > (prev_i + min_silence_len)
if not continuous and silence_has_gap:
silent_ranges.append([current_range_start,
prev_i + min_silence_len])
current_range_start = silence_start_i
prev_i = silence_start_i
silent_ranges.append([current_range_start,
prev_i + min_silence_len])
return silent_ranges
def detect_nonsilent(audio_segment, min_silence_len=1000, silence_thresh=-16):
silent_ranges = detect_silence(audio_segment, min_silence_len, silence_thresh)
len_seg = len(audio_segment)
# if there is no silence, the whole thing is nonsilent
if not silent_ranges:
return [[0, len_seg]]
# short circuit when the whole audio segment is silent
if silent_ranges[0][0] == 0 and silent_ranges[0][1] == len_seg:
return []
prev_end_i = 0
nonsilent_ranges = []
for start_i, end_i in silent_ranges:
nonsilent_ranges.append([prev_end_i, start_i])
prev_end_i = end_i
if end_i != len_seg:
nonsilent_ranges.append([prev_end_i, len_seg])
if nonsilent_ranges[0] == [0, 0]:
nonsilent_ranges.pop(0)
return nonsilent_ranges
def split_on_silence(audio_segment, min_silence_len=1000, silence_thresh=-16, keep_silence=100):
"""
audio_segment - original pydub.AudioSegment() object
min_silence_len - (in ms) minimum length of a silence to be used for
a split. default: 1000ms
silence_thresh - (in dBFS) anything quieter than this will be
considered silence. default: -16dBFS
keep_silence - (in ms) amount of silence to leave at the beginning
and end of the chunks. Keeps the sound from sounding like it is
abruptly cut off. (default: 100ms)
"""
not_silence_ranges = detect_nonsilent(audio_segment, min_silence_len, silence_thresh)
chunks = []
for start_i, end_i in not_silence_ranges:
start_i = max(0, start_i - keep_silence)
end_i += keep_silence
chunks.append(audio_segment[start_i:end_i])
return chunks
|