File size: 5,120 Bytes
2f6b10b |
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 |
import collections
import os
import re
import shutil
import subprocess
import argparse
from pathlib import Path
from praatio.textgrid import openTextgrid, Textgrid, IntervalTier
korean_root = '/mnt/d/Data/speech/korean_corpora'
seoul_corpus = os.path.join(korean_root, 'seoul_corpus')
original_textgrids = os.path.join(seoul_corpus, 'original')
training_textgrids = os.path.join(seoul_corpus, 'seoul_corpus_benchmark')
reference_dir = os.path.join(seoul_corpus, 'seoul_reference_alignments')
speaker_info = {}
def fix_sample_rate(benchmark_directory: Path):
for speaker in os.listdir(benchmark_directory):
speaker_dir = os.path.join(benchmark_directory, speaker)
if not os.path.isdir(speaker_dir):
continue
for file in os.listdir(speaker_dir):
if 'resampled' in file:
os.remove(file)
if file.endswith('.flac') and 'resampled' not in file:
path = os.path.join(speaker_dir, file)
resampled_file = path.replace('.flac', '_resampled.flac')
subprocess.check_call(['sox', path, '-r', '16000', resampled_file])
os.remove(path)
os.rename(resampled_file, path)
def fix_textgrids(original_directory: Path, benchmark_directory: Path, reference_directory: Path):
for file in os.listdir(original_directory):
if os.path.isdir(os.path.join(original_directory, file)):
continue
speaker = file[:3]
speaker_gender = file[3]
speaker_age = file[4:6]
print(speaker, speaker_gender, speaker_age)
speaker_info[speaker] = {'gender': speaker_gender, 'age':speaker_age}
speaker_directory = os.path.join(benchmark_directory, speaker)
reference_speaker_directory = os.path.join(reference_directory, speaker)
os.makedirs(speaker_directory, exist_ok=True)
os.makedirs(reference_speaker_directory, exist_ok=True)
if not file.endswith('.TextGrid'):
shutil.move(os.path.join(original_directory, file), os.path.join(speaker_directory, file))
continue
tg_path = os.path.join(original_directory, file)
tg = openTextgrid(tg_path, includeEmptyIntervals=False, duplicateNamesMode='rename')
orth = tg._tierDict['utt.prono.']
new_intervals = []
new_tg_path = os.path.join(speaker_directory, file)
reference_tg_path = os.path.join(reference_speaker_directory, file)
new_tg = Textgrid(minTimestamp=tg.minTimestamp, maxTimestamp=tg.maxTimestamp)
word_intervals = []
for interval in orth._entries:
if re.match(r'^<[^>]+>$', interval.label):
continue
text = interval.label.replace('<VOCNOISE>', '')
if not text:
continue
begin = max(interval.start - 0.2, 0)
end = min(interval.end + 0.2, orth.maxTimestamp)
if new_intervals and begin <= new_intervals[-1][1]:
new_intervals[-1][1] = end
new_intervals[-1][2] += ' ' + text
else:
new_intervals.append([begin, end, text])
word_intervals.append((interval.start, interval.end, text))
tier = IntervalTier(speaker, new_intervals,minT=tg.minTimestamp, maxT=tg.maxTimestamp)
new_tg.addTier(tier)
new_tg.save(new_tg_path,includeBlankSpaces=True, format='short_textgrid')
reference_tg = Textgrid(minTimestamp=tg.minTimestamp, maxTimestamp=tg.maxTimestamp)
phone_intervals = []
for interval in tg._tierDict['phoneme']._entries:
if interval.label in {'<IVER>', '<SIL>', '<VOCNOISE>', '<NOISE>'}:
continue
text = interval.label
if text.startswith('<'):
text = 'spn'
phone_intervals.append((interval.start, interval.end, text))
word_tier = IntervalTier('words', word_intervals, minT=tg.minTimestamp, maxT=tg.maxTimestamp)
phone_tier = IntervalTier('phones', phone_intervals, minT=tg.minTimestamp, maxT=tg.maxTimestamp)
reference_tg.addTier(word_tier)
reference_tg.addTier(phone_tier)
reference_tg.save(reference_tg_path, format='short_textgrid', includeBlankSpaces=True)
def parse_directory(original_directory: Path, benchmark_directory: Path, reference_directory: Path):
# fix_sample_rate(benchmark_directory)
fix_textgrids(original_directory, benchmark_directory, reference_directory)
if __name__ == '__main__':
parser = argparse.ArgumentParser(
prog='create_seoul_benchmark',
description='Creates two directories of TextGrid files for use with MFA, '
'one as input with utterances (benchmark) and one for use in reference alignments (reference)')
parser.add_argument('original_directory')
parser.add_argument('benchmark_directory')
parser.add_argument('reference_directory')
args = parser.parse_args()
parse_directory(
Path(args.original_directory),
Path(args.benchmark_directory),
Path(args.reference_directory)
)
|