You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
183 lines
6.3 KiB
183 lines
6.3 KiB
2 years ago
|
"""
|
||
|
|
Various functions for finding/manipulating silence in AudioSegments
|
||
|
|
"""
|
||
|
|
import itertools
|
||
|
|
|
||
|
|
from .utils import db_to_float
|
||
|
|
|
||
|
|
|
||
|
|
def detect_silence(audio_segment, min_silence_len=1000, silence_thresh=-16, seek_step=1):
|
||
|
|
"""
|
||
|
|
Returns a list of all silent sections [start, end] in milliseconds of audio_segment.
|
||
|
|
Inverse of detect_nonsilent()
|
||
|
|
|
||
|
|
audio_segment - the segment to find silence in
|
||
|
|
min_silence_len - the minimum length for any silent section
|
||
|
|
silence_thresh - the upper bound for how quiet is silent in dFBS
|
||
|
|
seek_step - step size for interating over the segment in ms
|
||
|
|
"""
|
||
|
|
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 successive (1 sec by default) chunk of sound for silence
|
||
|
|
# try a chunk at every "seek step" (or every chunk for a seek step == 1)
|
||
|
|
last_slice_start = seg_len - min_silence_len
|
||
|
|
slice_starts = range(0, last_slice_start + 1, seek_step)
|
||
|
|
|
||
|
|
# guarantee last_slice_start is included in the range
|
||
|
|
# to make sure the last portion of the audio is searched
|
||
|
|
if last_slice_start % seek_step:
|
||
|
|
slice_starts = itertools.chain(slice_starts, [last_slice_start])
|
||
|
|
|
||
|
|
for i in slice_starts:
|
||
|
|
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 + seek_step)
|
||
|
|
|
||
|
|
# 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, seek_step=1):
|
||
|
|
"""
|
||
|
|
Returns a list of all nonsilent sections [start, end] in milliseconds of audio_segment.
|
||
|
|
Inverse of detect_silent()
|
||
|
|
|
||
|
|
audio_segment - the segment to find silence in
|
||
|
|
min_silence_len - the minimum length for any silent section
|
||
|
|
silence_thresh - the upper bound for how quiet is silent in dFBS
|
||
|
|
seek_step - step size for interating over the segment in ms
|
||
|
|
"""
|
||
|
|
silent_ranges = detect_silence(audio_segment, min_silence_len, silence_thresh, seek_step)
|
||
|
|
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,
|
||
|
|
seek_step=1):
|
||
|
|
"""
|
||
|
|
Returns list of audio segments from splitting audio_segment on silent sections
|
||
|
|
|
||
|
|
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 or True/False) leave some silence at the beginning
|
||
|
|
and end of the chunks. Keeps the sound from sounding like it
|
||
|
|
is abruptly cut off.
|
||
|
|
When the length of the silence is less than the keep_silence duration
|
||
|
|
it is split evenly between the preceding and following non-silent
|
||
|
|
segments.
|
||
|
|
If True is specified, all the silence is kept, if False none is kept.
|
||
|
|
default: 100ms
|
||
|
|
|
||
|
|
seek_step - step size for interating over the segment in ms
|
||
|
|
"""
|
||
|
|
|
||
|
|
# from the itertools documentation
|
||
|
|
def pairwise(iterable):
|
||
|
|
"s -> (s0,s1), (s1,s2), (s2, s3), ..."
|
||
|
|
a, b = itertools.tee(iterable)
|
||
|
|
next(b, None)
|
||
|
|
return zip(a, b)
|
||
|
|
|
||
|
|
if isinstance(keep_silence, bool):
|
||
|
|
keep_silence = len(audio_segment) if keep_silence else 0
|
||
|
|
|
||
|
|
output_ranges = [
|
||
|
|
[ start - keep_silence, end + keep_silence ]
|
||
|
|
for (start,end)
|
||
|
|
in detect_nonsilent(audio_segment, min_silence_len, silence_thresh, seek_step)
|
||
|
|
]
|
||
|
|
|
||
|
|
for range_i, range_ii in pairwise(output_ranges):
|
||
|
|
last_end = range_i[1]
|
||
|
|
next_start = range_ii[0]
|
||
|
|
if next_start < last_end:
|
||
|
|
range_i[1] = (last_end+next_start)//2
|
||
|
|
range_ii[0] = range_i[1]
|
||
|
|
|
||
|
|
return [
|
||
|
|
audio_segment[ max(start,0) : min(end,len(audio_segment)) ]
|
||
|
|
for start,end in output_ranges
|
||
|
|
]
|
||
|
|
|
||
|
|
|
||
|
|
def detect_leading_silence(sound, silence_threshold=-50.0, chunk_size=10):
|
||
|
|
"""
|
||
|
|
Returns the millisecond/index that the leading silence ends.
|
||
|
|
|
||
|
|
audio_segment - the segment to find silence in
|
||
|
|
silence_threshold - the upper bound for how quiet is silent in dFBS
|
||
|
|
chunk_size - chunk size for interating over the segment in ms
|
||
|
|
"""
|
||
|
|
trim_ms = 0 # ms
|
||
|
|
assert chunk_size > 0 # to avoid infinite loop
|
||
|
|
while sound[trim_ms:trim_ms+chunk_size].dBFS < silence_threshold and trim_ms < len(sound):
|
||
|
|
trim_ms += chunk_size
|
||
|
|
|
||
|
|
# if there is no end it should return the length of the segment
|
||
|
|
return min(trim_ms, len(sound))
|
||
|
|
|
||
|
|
|