forked from IS2AI/SpeakingFaces
-
Notifications
You must be signed in to change notification settings - Fork 0
/
trim_audio.py
118 lines (92 loc) · 4.76 KB
/
trim_audio.py
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
110
111
112
113
114
115
116
117
118
from speakingfacespy.imtools import make_dir
from imutils import paths
import scipy.io.wavfile
import argparse
import glob
import os
def trim_audio(audio_trim_filename, length_audio_trim, to_trim_mic_id, to_trim_new_dir):
# construct a filename for the audio which we wanna trim
audio_to_trim_filename = '_'.join(audio_trim_filename.split('_')[:-1])+'_{}.wav'.format(to_trim_mic_id)
# construct a path to save the audio file
audio_to_trim_new_filepath = os.path.join(to_trim_new_dir, audio_to_trim_filename)
# extract the sample rate and audio data
print("[INFO] Reading an audio file to trim: {}".format(audio_to_trim_filename))
sample_rate_to_trim, audio_to_trim = scipy.io.wavfile.read(os.path.join(to_trim_new_dir[:-5], audio_to_trim_filename))
# estimate the duration of the file
duration_to_trim = audio_to_trim.shape[0] / sample_rate_to_trim
print("[INFO] Duration of the original file = {} sec".format(duration_to_trim))
print("[INFO] Writing trimmed audio to {}".format(audio_to_trim_new_filepath))
scipy.io.wavfile.write(audio_to_trim_new_filepath, sample_rate_to_trim, audio_to_trim[:length_audio_trim])
def trim_audio_by_sub_trial(path_to_dataset, sub_id, trial_id):
# construct a path to the data for the given subject and trial
path_to_dataset = os.path.join(path_to_dataset, "sub_{}/trial_{}".format(sub_id, trial_id))
# find a folder with audio files
for mic_id in [1, 2]:
# construct a path to the trimmed audiofiles
path_to_audiofiles = os.path.join(path_to_dataset, "mic{}_audio_cmd_trim".format(mic_id))
# grab all audiofiles for the given directory
audio_file_paths = list(paths.list_files(path_to_audiofiles))
# if the folder is not empty
# then exit from the loop
if len(audio_file_paths) > 0:
break
# createa a directory to save new trimmed audio files
if mic_id == 1:
to_trim_dir = os.path.join(path_to_dataset, "mic{}_audio_cmd_trim".format(2))
to_trim_mic_id = 2
else:
to_trim_dir = os.path.join(path_to_dataset, "mic{}_audio_cmd_trim".format(1))
to_trim_mic_id = 1
make_dir(to_trim_dir)
# loop over the trimmed audio files
for audio_file_path in audio_file_paths:
# extract the filename
audio_trim_filename = audio_file_path.split(os.path.sep)[-1]
print("[INFO] Reading already trimmed audio file: {}".format(audio_trim_filename))
# extract the sample rate and audio data from
# the already trimmed file
sample_rate_trim, audio_trim = scipy.io.wavfile.read(audio_file_path)
# estimate the duration of the already trimmed file
duration_trim = audio_trim.shape[0] / sample_rate_trim
print("[INFO] Duration of the already trimmed file = {} sec".format(duration_trim))
#trim the corresponding audio file from the other
# mic accordingly
trim_audio(audio_trim_filename, audio_trim.shape[0], to_trim_mic_id, to_trim_dir)
def trim_audio_by_range(path_to_dataset, sub_id_i, sub_id_l):
print("[INFO] trim all audio files in the range of sub_id [{} ... {}]".format(sub_id_i, sub_id_l))
for sub_id in range(sub_id_i, sub_id_l + 1):
for trial_id in range(1, 3):
trim_audio_by_sub_trial(path_to_dataset, sub_id, trial_id)
# construct the argument parser and parse the arguments
ap = argparse.ArgumentParser()
ap.add_argument("-d", "--dataset", required=True,
help="path to the SpeakingFaces dataset")
ap.add_argument("-i", "--sub_info", nargs='+', type=int, default = (0,0),
help="subID(1,...,142) trialID (1...2)")
ap.add_argument("-r", "--sub_range", nargs='+', type=int, default = (0,0),
help="process more than one subject (1...142)")
args = vars(ap.parse_args())
# initialize a path to the dataset
path_to_dataset = args["dataset"]
# if we are going to extract frames for a specific subject
if args["sub_info"][0] > 0 and args["sub_info"][0] <= 142:
# initialize the subject ID and
# trial ID
sub_id = args["sub_info"][0]
trial_id = args["sub_info"][1]
if trial_id >0 and trial_id <= 2:
# extract frames for the given subject and trial
trim_audio_by_sub_trial(path_to_dataset, sub_id, trial_id)
else:
print("[INFO] Number of trials are invalid!")
# if we want to process more than one subject
elif args["sub_range"][0] > 0 and args["sub_range"][0] <= 142:
# initialize IDs of the initial
# and the last subjects
sub_id_i = args["sub_range"][0]
sub_id_l = args["sub_range"][1]
# extract frames
trim_audio_by_range(path_to_dataset, sub_id_i, sub_id_l)
# otherwise ask to enter proper arguments
else:
print("[INFO] --sub_info or/and --sub_range was/were not defined properly!")