forked from mangelroman/audio2score
-
Notifications
You must be signed in to change notification settings - Fork 0
/
prepare.py
237 lines (188 loc) · 9.54 KB
/
prepare.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
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
import os
import sys
import argparse
import multiprocessing as mp
import subprocess
import csv
import json
import pickle
import numpy as np
from pathlib import Path
from tqdm import tqdm
from sklearn.model_selection import train_test_split
from data.data_loader import load_audio
from data.humdrum import Kern, Labels, LabelsMulti
def parseList(string):
if string and len(string) > 0:
return string.split(',')
return None
def parseIntList(string):
if string and len(string) > 0:
return [int(x) for x in string.split(',')]
return None
parser = argparse.ArgumentParser(description='Spectrum preparation')
parser.add_argument('--data-dir', metavar='DIR', help='path to data', required=True)
parser.add_argument('--out-dir', metavar='DIR', help='path to output', required=True)
parser.add_argument('--min-duration-symbol', type=float, help='Select the minimum duration per symbol', required=True)
parser.add_argument('--max-duration', type=float, help='Select maximum duration of audio input files in seconds', required=True)
parser.add_argument('--num-workers', default=None, type=int, help='Number of workers used in data preparation')
parser.add_argument('--sample-rate', type=int, help='Select sampling frequency of WAV files', default=22050)
parser.add_argument('--bit-rate', type=int, help='Select bit rate of MP4 audio files', default=128000)
parser.add_argument('--soundfont', metavar='FILE', help='path to soundfont', default='/usr/share/sounds/sf2/FluidR3_GM.sf2')
parser.add_argument('--resynthesize', dest='resynthesize', action='store_true', default=False, help='Resynthesize audio from midi')
parser.add_argument('--instruments', type=parseList, help='Override kern defined intruments', default='piano')
parser.add_argument('--tempo-scaling', type=float, default=0.06, help='Select tempo random scaling')
parser.add_argument('--chunk-sizes', type=parseIntList, help='Select chunk sizes separated by commas', default=[sys.maxsize])
parser.add_argument('--test-split', type=float, default=0.3, help='Select train-test split ratio')
parser.add_argument('--train-stride', type=int, help='Select the stride of overlapped training samples', default=None)
parser.add_argument('--id', default='manifest', help='Id of output manifest and label files')
parser.add_argument('--labels-multi', action="store_true", default=False, help="Use multichar labels to reduce sequence size")
def process_sample(q, samples, args, labels):
while True:
score_path = q.get()
if score_path is None:
break
# Remove grace notes, ornaments, etc...
kern = Kern(Path(args.data_dir) / score_path)
kern.spines.override_instruments(args.instruments)
try:
if not kern.clean():
print(f'Cannot clean kern {score_path}')
continue
except Exception as e:
print(f"Exception while cleaning {score_path} audio. Reason: {e}")
continue
root_path = Path(args.out_dir) / score_path.parent
root_path.mkdir(parents=True, exist_ok=True)
krn_path = Path(args.out_dir) / score_path
# Set seed to ensure same chunk sizes and tempo scaling
np.random.seed(bytearray(score_path.name, 'utf-8'))
try:
kern_chunks = kern.split(args.chunk_sizes, args.train_stride)
except Exception as e:
print(f'Exception {e} while splitting {score_path}')
continue
# random scale between +ts and -ts
ts = 1 + args.tempo_scaling * (2 * np.random.rand(len(kern_chunks)) - 1)
for i, kern in enumerate(kern_chunks):
chunk_path = krn_path.with_suffix(f'.{i:03d}.krn')
kern.save(chunk_path)
# Fix ties with tiefix command
process = subprocess.run(['tiefix', chunk_path], capture_output=True, encoding='iso-8859-1')
if (process.returncode != 0):
print(f"tiefix error={process.returncode} on {chunk_path}")
print(process.stdout)
continue
kern = Kern(data=process.stdout)
kern.save(chunk_path)
audio_path = chunk_path.with_suffix('.flac')
if args.resynthesize or not audio_path.exists():
mid_path = chunk_path.with_suffix('.mid')
# Tempo and instrumment extracted from *MM and *I indications
status = os.system(f'hum2mid {str(chunk_path)} -C -v 100 -t {ts[i]} -o {str(mid_path)} >/dev/null 2>&1')
if (os.WEXITSTATUS(status) != 0):
print(f"hum2mid error={status} on {krn_path}")
continue
status = os.system(f'fluidsynth --sample-rate={args.sample_rate} -O s16 -T raw -i -l -F - {args.soundfont} {str(mid_path)} | '
f'ffmpeg -y -f s16le -ar {args.sample_rate} -ac 2 -i pipe: '
f'-ar {args.sample_rate} -ac 1 -ab {args.bit_rate} -strict -2 {str(audio_path)} 2>/dev/null')
try:
y = load_audio(str(audio_path))
except Exception as e:
print(f"Exception while loading {chunk_path} audio. Reason: {e}")
continue
duration = len(y) / args.sample_rate
krnseq = kern.tosequence()
if krnseq is None:
#print(f"Discarded {chunk_path} for double dots/sharps/flats")
continue
try:
seq = labels.encode(krnseq)
except Exception as e:
print(f"Discarded {chunk_path}. Reason: {e}")
continue
seqlen = labels.ctclen(seq)
krnseq_path = chunk_path.with_suffix('.krnseq')
krnseq_path.write_text(krnseq)
seq_path = chunk_path.with_suffix('.seq')
with seq_path.open(mode="wb") as f:
f.write(pickle.dumps(seq))
if duration > args.max_duration or duration < seqlen * args.min_duration_symbol:
#print(f"Sequence too long in {chunk_path} len={seqlen} duration={duration:.2f}")
continue
samples.append([str(audio_path), str(seq_path), duration])
def process_scores(scores, args, labels):
manager = mp.Manager()
samples = manager.list()
q = mp.Queue(maxsize=args.num_workers)
pool = mp.Pool(args.num_workers, initializer=process_sample, initargs=(q, samples, args, labels))
for score in tqdm(scores, ascii=True):
if score.is_symlink():
continue
q.put(score)
# stop workers
for i in range(args.num_workers):
q.put(None)
pool.close()
pool.join()
x = np.array([x[0] for x in samples])
y = np.array([x[1] for x in samples])
durations = np.array([x[2] for x in samples])
# SortaGrad
sorted_indexes = np.argsort(durations)
x = x[sorted_indexes]
y = y[sorted_indexes]
return x, y, np.sum(durations)
if __name__ == '__main__':
args = parser.parse_args()
print("Preprocessing humdrum data...")
outdir = Path(args.out_dir)
outdir.mkdir(parents=True, exist_ok=True)
root = Path(args.data_dir)
scores = sorted([x.relative_to(root) for x in root.rglob('*.krn')])
print("Spliting train/test samples...")
scores_train, scores_test = train_test_split(scores, test_size=args.test_split, random_state=45)
middle = round(len(scores_test) / 2) # Favor validation if number of samples is odd
scores_val = scores_test[:middle]
scores_test = scores_test[middle:]
labels = Labels() if not args.labels_multi else LabelsMulti()
if args.num_workers is None:
args.num_workers = 4
print("Processing training samples:")
x_train, y_train, train_dur = process_scores(scores_train, args, labels)
# Force no overlap for validation and test samples
args.train_stride = None
print("Processing validation samples:")
x_val, y_val, val_dur = process_scores(scores_val, args, labels)
print("Processing test samples:")
x_test, y_test, test_dur = process_scores(scores_test, args, labels)
print("Number of train samples: {} ({:.2f} hours)".format(len(x_train), train_dur / 3600))
print("Number of validation samples: {} ({:.2f} hours)".format(len(x_val), val_dur / 3600))
print("Number of test samples: {} ({:.2f} hours)".format(len(x_test), test_dur / 3600))
total_samples = len(x_train) + len(x_val) + len(x_test)
total_dur = train_dur + val_dur + test_dur
print("Total samples: {} ({:.2f} hours)".format(total_samples, total_dur / 3600))
train_manifest_path = Path(f'train_{args.id}.csv')
val_manifest_path = Path(f'val_{args.id}.csv')
test_manifest_path = Path(f'test_{args.id}.csv')
labels_path = Path(f'labels_{args.id}.json')
with train_manifest_path.open(mode='w') as csvfile:
writer = csv.writer(csvfile)
print(f"Creating train manifest {train_manifest_path}...")
for x, y in zip(x_train, y_train):
writer.writerow([x, y])
with val_manifest_path.open(mode='w') as csvfile:
writer = csv.writer(csvfile)
print(f"Creating val manifest {val_manifest_path}...")
for x, y in zip(x_val, y_val):
writer.writerow([x, y])
with test_manifest_path.open(mode='w') as csvfile:
writer = csv.writer(csvfile)
print(f"Creating test manifest {test_manifest_path}...")
for x, y in zip(x_test, y_test):
writer.writerow([x, y])
print("Creating label JSON file with {} symbols".format(len(labels.labels)))
print(labels.labels)
with labels_path.open(mode='w') as jsonfile:
json.dump(labels.labels, jsonfile)
sys.exit(0)