Upload 2 files
Browse files
asr.py
ADDED
@@ -0,0 +1,233 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from typing import *
|
2 |
+
import logging
|
3 |
+
import time
|
4 |
+
import logging
|
5 |
+
import sherpa_onnx
|
6 |
+
import os
|
7 |
+
import asyncio
|
8 |
+
import numpy as np
|
9 |
+
|
10 |
+
logger = logging.getLogger(__file__)
|
11 |
+
_asr_engines = {}
|
12 |
+
|
13 |
+
|
14 |
+
class ASRResult:
|
15 |
+
def __init__(self, text: str, finished: bool, idx: int):
|
16 |
+
self.text = text
|
17 |
+
self.finished = finished
|
18 |
+
self.idx = idx
|
19 |
+
|
20 |
+
def to_dict(self):
|
21 |
+
return {"text": self.text, "finished": self.finished, "idx": self.idx}
|
22 |
+
|
23 |
+
|
24 |
+
class ASRStream:
|
25 |
+
def __init__(self, recognizer: Union[sherpa_onnx.OnlineRecognizer | sherpa_onnx.OfflineRecognizer], sample_rate: int) -> None:
|
26 |
+
self.recognizer = recognizer
|
27 |
+
self.inbuf = asyncio.Queue()
|
28 |
+
self.outbuf = asyncio.Queue()
|
29 |
+
self.sample_rate = sample_rate
|
30 |
+
self.is_closed = False
|
31 |
+
self.online = isinstance(recognizer, sherpa_onnx.OnlineRecognizer)
|
32 |
+
|
33 |
+
async def start(self):
|
34 |
+
if self.online:
|
35 |
+
asyncio.create_task(self.run_online())
|
36 |
+
else:
|
37 |
+
asyncio.create_task(self.run_offline())
|
38 |
+
|
39 |
+
async def run_online(self):
|
40 |
+
stream = self.recognizer.create_stream()
|
41 |
+
last_result = ""
|
42 |
+
segment_id = 0
|
43 |
+
logger.info('asr: start real-time recognizer')
|
44 |
+
while not self.is_closed:
|
45 |
+
samples = await self.inbuf.get()
|
46 |
+
stream.accept_waveform(self.sample_rate, samples)
|
47 |
+
while self.recognizer.is_ready(stream):
|
48 |
+
self.recognizer.decode_stream(stream)
|
49 |
+
|
50 |
+
is_endpoint = self.recognizer.is_endpoint(stream)
|
51 |
+
result = self.recognizer.get_result(stream)
|
52 |
+
|
53 |
+
if result and (last_result != result):
|
54 |
+
last_result = result
|
55 |
+
logger.info(f' > {segment_id}:{result}')
|
56 |
+
self.outbuf.put_nowait(
|
57 |
+
ASRResult(result, False, segment_id))
|
58 |
+
|
59 |
+
if is_endpoint:
|
60 |
+
if result:
|
61 |
+
logger.info(f'{segment_id}: {result}')
|
62 |
+
self.outbuf.put_nowait(
|
63 |
+
ASRResult(result, True, segment_id))
|
64 |
+
segment_id += 1
|
65 |
+
self.recognizer.reset(stream)
|
66 |
+
|
67 |
+
async def run_offline(self):
|
68 |
+
vad = _asr_engines['vad']
|
69 |
+
segment_id = 0
|
70 |
+
st = None
|
71 |
+
while not self.is_closed:
|
72 |
+
samples = await self.inbuf.get()
|
73 |
+
vad.accept_waveform(samples)
|
74 |
+
while not vad.empty():
|
75 |
+
if not st:
|
76 |
+
st = time.time()
|
77 |
+
stream = self.recognizer.create_stream()
|
78 |
+
stream.accept_waveform(self.sample_rate, vad.front.samples)
|
79 |
+
|
80 |
+
vad.pop()
|
81 |
+
self.recognizer.decode_stream(stream)
|
82 |
+
|
83 |
+
result = stream.result.text.strip()
|
84 |
+
if result:
|
85 |
+
duration = time.time() - st
|
86 |
+
logger.info(f'{segment_id}:{result} ({duration:.2f}s)')
|
87 |
+
self.outbuf.put_nowait(ASRResult(result, True, segment_id))
|
88 |
+
segment_id += 1
|
89 |
+
st = None
|
90 |
+
|
91 |
+
async def close(self):
|
92 |
+
self.is_closed = True
|
93 |
+
self.outbuf.put_nowait(None)
|
94 |
+
|
95 |
+
async def write(self, pcm_bytes: bytes):
|
96 |
+
pcm_data = np.frombuffer(pcm_bytes, dtype=np.int16)
|
97 |
+
samples = pcm_data.astype(np.float32) / 32768.0
|
98 |
+
self.inbuf.put_nowait(samples)
|
99 |
+
|
100 |
+
async def read(self) -> ASRResult:
|
101 |
+
return await self.outbuf.get()
|
102 |
+
|
103 |
+
|
104 |
+
def create_zipformer(samplerate: int, args) -> sherpa_onnx.OnlineRecognizer:
|
105 |
+
d = os.path.join(
|
106 |
+
args.models_root, 'sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20')
|
107 |
+
if not os.path.exists(d):
|
108 |
+
raise ValueError(f"asr: model not found {d}")
|
109 |
+
|
110 |
+
encoder = os.path.join(d, "encoder-epoch-99-avg-1.onnx")
|
111 |
+
decoder = os.path.join(d, "decoder-epoch-99-avg-1.onnx")
|
112 |
+
joiner = os.path.join(d, "joiner-epoch-99-avg-1.onnx")
|
113 |
+
tokens = os.path.join(d, "tokens.txt")
|
114 |
+
|
115 |
+
recognizer = sherpa_onnx.OnlineRecognizer.from_transducer(
|
116 |
+
tokens=tokens,
|
117 |
+
encoder=encoder,
|
118 |
+
decoder=decoder,
|
119 |
+
joiner=joiner,
|
120 |
+
provider=args.asr_provider,
|
121 |
+
num_threads=args.threads,
|
122 |
+
sample_rate=samplerate,
|
123 |
+
feature_dim=80,
|
124 |
+
enable_endpoint_detection=True,
|
125 |
+
rule1_min_trailing_silence=2.4,
|
126 |
+
rule2_min_trailing_silence=1.2,
|
127 |
+
rule3_min_utterance_length=20, # it essentially disables this rule
|
128 |
+
)
|
129 |
+
return recognizer
|
130 |
+
|
131 |
+
|
132 |
+
def create_sensevoice(samplerate: int, args) -> sherpa_onnx.OfflineRecognizer:
|
133 |
+
d = os.path.join(args.models_root,
|
134 |
+
'sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17')
|
135 |
+
|
136 |
+
if not os.path.exists(d):
|
137 |
+
raise ValueError(f"asr: model not found {d}")
|
138 |
+
|
139 |
+
recognizer = sherpa_onnx.OfflineRecognizer.from_sense_voice(
|
140 |
+
model=os.path.join(d, 'model.onnx'),
|
141 |
+
tokens=os.path.join(d, 'tokens.txt'),
|
142 |
+
num_threads=args.threads,
|
143 |
+
sample_rate=samplerate,
|
144 |
+
use_itn=True,
|
145 |
+
debug=0,
|
146 |
+
language=args.asr_lang,
|
147 |
+
)
|
148 |
+
return recognizer
|
149 |
+
|
150 |
+
|
151 |
+
def create_paraformer_trilingual(samplerate: int, args) -> sherpa_onnx.OnlineRecognizer:
|
152 |
+
d = os.path.join(
|
153 |
+
args.models_root, 'sherpa-onnx-paraformer-trilingual-zh-cantonese-en')
|
154 |
+
if not os.path.exists(d):
|
155 |
+
raise ValueError(f"asr: model not found {d}")
|
156 |
+
|
157 |
+
recognizer = sherpa_onnx.OfflineRecognizer.from_paraformer(
|
158 |
+
paraformer=os.path.join(d, 'model.onnx'),
|
159 |
+
tokens=os.path.join(d, 'tokens.txt'),
|
160 |
+
num_threads=args.threads,
|
161 |
+
sample_rate=samplerate,
|
162 |
+
debug=0,
|
163 |
+
provider=args.asr_provider,
|
164 |
+
)
|
165 |
+
return recognizer
|
166 |
+
|
167 |
+
|
168 |
+
def create_paraformer_en(samplerate: int, args) -> sherpa_onnx.OnlineRecognizer:
|
169 |
+
d = os.path.join(
|
170 |
+
args.models_root, 'sherpa-onnx-paraformer-en')
|
171 |
+
if not os.path.exists(d):
|
172 |
+
raise ValueError(f"asr: model not found {d}")
|
173 |
+
|
174 |
+
recognizer = sherpa_onnx.OfflineRecognizer.from_paraformer(
|
175 |
+
paraformer=os.path.join(d, 'model.onnx'),
|
176 |
+
tokens=os.path.join(d, 'tokens.txt'),
|
177 |
+
num_threads=args.threads,
|
178 |
+
sample_rate=samplerate,
|
179 |
+
use_itn=True,
|
180 |
+
debug=0,
|
181 |
+
provider=args.asr_provider,
|
182 |
+
)
|
183 |
+
return recognizer
|
184 |
+
|
185 |
+
|
186 |
+
def load_asr_engine(samplerate: int, args) -> sherpa_onnx.OnlineRecognizer:
|
187 |
+
cache_engine = _asr_engines.get(args.asr_model)
|
188 |
+
if cache_engine:
|
189 |
+
return cache_engine
|
190 |
+
st = time.time()
|
191 |
+
if args.asr_model == 'zipformer-bilingual':
|
192 |
+
cache_engine = create_zipformer(samplerate, args)
|
193 |
+
elif args.asr_model == 'sensevoice':
|
194 |
+
cache_engine = create_sensevoice(samplerate, args)
|
195 |
+
_asr_engines['vad'] = load_vad_engine(samplerate, args)
|
196 |
+
elif args.asr_model == 'paraformer-trilingual':
|
197 |
+
cache_engine = create_paraformer_trilingual(samplerate, args)
|
198 |
+
_asr_engines['vad'] = load_vad_engine(samplerate, args)
|
199 |
+
elif args.asr_model == 'paraformer-en':
|
200 |
+
cache_engine = create_paraformer_en(samplerate, args)
|
201 |
+
_asr_engines['vad'] = load_vad_engine(samplerate, args)
|
202 |
+
else:
|
203 |
+
raise ValueError(f"asr: unknown model {args.asr_model}")
|
204 |
+
_asr_engines[args.asr_model] = cache_engine
|
205 |
+
logger.info(f"asr: engine loaded in {time.time() - st:.2f}s")
|
206 |
+
return cache_engine
|
207 |
+
|
208 |
+
|
209 |
+
def load_vad_engine(samplerate: int, args, min_silence_duration: float = 0.25, buffer_size_in_seconds: int = 100) -> sherpa_onnx.VoiceActivityDetector:
|
210 |
+
config = sherpa_onnx.VadModelConfig()
|
211 |
+
d = os.path.join(args.models_root, 'silero_vad')
|
212 |
+
if not os.path.exists(d):
|
213 |
+
raise ValueError(f"vad: model not found {d}")
|
214 |
+
|
215 |
+
config.silero_vad.model = os.path.join(d, 'silero_vad.onnx')
|
216 |
+
config.silero_vad.min_silence_duration = min_silence_duration
|
217 |
+
config.sample_rate = samplerate
|
218 |
+
config.provider = args.asr_provider
|
219 |
+
config.num_threads = args.threads
|
220 |
+
|
221 |
+
vad = sherpa_onnx.VoiceActivityDetector(
|
222 |
+
config,
|
223 |
+
buffer_size_in_seconds=buffer_size_in_seconds)
|
224 |
+
return vad
|
225 |
+
|
226 |
+
|
227 |
+
async def start_asr_stream(samplerate: int, args) -> ASRStream:
|
228 |
+
"""
|
229 |
+
Start a ASR stream
|
230 |
+
"""
|
231 |
+
stream = ASRStream(load_asr_engine(samplerate, args), samplerate)
|
232 |
+
await stream.start()
|
233 |
+
return stream
|
tts.py
ADDED
@@ -0,0 +1,216 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from typing import *
|
2 |
+
import os
|
3 |
+
import time
|
4 |
+
import sherpa_onnx
|
5 |
+
import logging
|
6 |
+
import numpy as np
|
7 |
+
import asyncio
|
8 |
+
import time
|
9 |
+
import soundfile
|
10 |
+
from scipy.signal import resample
|
11 |
+
import io
|
12 |
+
import re
|
13 |
+
|
14 |
+
logger = logging.getLogger(__file__)
|
15 |
+
|
16 |
+
splitter = re.compile(r'[,,。.!?!?;;、\n]')
|
17 |
+
_tts_engines = {}
|
18 |
+
|
19 |
+
tts_configs = {
|
20 |
+
'vits-zh-hf-theresa': {
|
21 |
+
'model': 'theresa.onnx',
|
22 |
+
'lexicon': 'lexicon.txt',
|
23 |
+
'dict_dir': 'dict',
|
24 |
+
'tokens': 'tokens.txt',
|
25 |
+
'sample_rate': 22050,
|
26 |
+
# 'rule_fsts': ['phone.fst', 'date.fst', 'number.fst'],
|
27 |
+
},
|
28 |
+
'vits-melo-tts-zh_en': {
|
29 |
+
'model': 'model.onnx',
|
30 |
+
'lexicon': 'lexicon.txt',
|
31 |
+
'dict_dir': 'dict',
|
32 |
+
'tokens': 'tokens.txt',
|
33 |
+
'sample_rate': 44100,
|
34 |
+
'rule_fsts': ['phone.fst', 'date.fst', 'number.fst'],
|
35 |
+
},
|
36 |
+
}
|
37 |
+
|
38 |
+
|
39 |
+
def load_tts_model(name: str, model_root: str, provider: str, num_threads: int = 1, max_num_sentences: int = 20) -> sherpa_onnx.OfflineTtsConfig:
|
40 |
+
cfg = tts_configs[name]
|
41 |
+
fsts = []
|
42 |
+
model_dir = os.path.join(model_root, name)
|
43 |
+
for f in cfg.get('rule_fsts', ''):
|
44 |
+
fsts.append(os.path.join(model_dir, f))
|
45 |
+
tts_rule_fsts = ','.join(fsts) if fsts else ''
|
46 |
+
|
47 |
+
model_config = sherpa_onnx.OfflineTtsModelConfig(
|
48 |
+
vits=sherpa_onnx.OfflineTtsVitsModelConfig(
|
49 |
+
model=os.path.join(model_dir, cfg['model']),
|
50 |
+
lexicon=os.path.join(model_dir, cfg['lexicon']),
|
51 |
+
dict_dir=os.path.join(model_dir, cfg['dict_dir']),
|
52 |
+
tokens=os.path.join(model_dir, cfg['tokens']),
|
53 |
+
),
|
54 |
+
provider=provider,
|
55 |
+
debug=0,
|
56 |
+
num_threads=num_threads,
|
57 |
+
)
|
58 |
+
tts_config = sherpa_onnx.OfflineTtsConfig(
|
59 |
+
model=model_config,
|
60 |
+
rule_fsts=tts_rule_fsts,
|
61 |
+
max_num_sentences=max_num_sentences)
|
62 |
+
|
63 |
+
if not tts_config.validate():
|
64 |
+
raise ValueError("tts: invalid config")
|
65 |
+
|
66 |
+
return tts_config
|
67 |
+
|
68 |
+
|
69 |
+
def get_tts_engine(args) -> Tuple[sherpa_onnx.OfflineTts, int]:
|
70 |
+
sample_rate = tts_configs[args.tts_model]['sample_rate']
|
71 |
+
cache_engine = _tts_engines.get(args.tts_model)
|
72 |
+
if cache_engine:
|
73 |
+
return cache_engine, sample_rate
|
74 |
+
st = time.time()
|
75 |
+
tts_config = load_tts_model(
|
76 |
+
args.tts_model, args.models_root, args.tts_provider)
|
77 |
+
|
78 |
+
cache_engine = sherpa_onnx.OfflineTts(tts_config)
|
79 |
+
elapsed = time.time() - st
|
80 |
+
logger.info(f"tts: loaded {args.tts_model} in {elapsed:.2f}s")
|
81 |
+
_tts_engines[args.tts_model] = cache_engine
|
82 |
+
|
83 |
+
return cache_engine, sample_rate
|
84 |
+
|
85 |
+
|
86 |
+
class TTSResult:
|
87 |
+
def __init__(self, pcm_bytes: bytes, finished: bool):
|
88 |
+
self.pcm_bytes = pcm_bytes
|
89 |
+
self.finished = finished
|
90 |
+
self.progress: float = 0.0
|
91 |
+
self.elapsed: float = 0.0
|
92 |
+
self.audio_duration: float = 0.0
|
93 |
+
self.audio_size: int = 0
|
94 |
+
|
95 |
+
def to_dict(self):
|
96 |
+
return {
|
97 |
+
"progress": self.progress,
|
98 |
+
"elapsed": f'{int(self.elapsed * 1000)}ms',
|
99 |
+
"duration": f'{self.audio_duration:.2f}s',
|
100 |
+
"size": self.audio_size
|
101 |
+
}
|
102 |
+
|
103 |
+
|
104 |
+
class TTSStream:
|
105 |
+
def __init__(self, engine, sid: int, speed: float = 1.0, sample_rate: int = 16000, original_sample_rate: int = 16000):
|
106 |
+
self.engine = engine
|
107 |
+
self.sid = sid
|
108 |
+
self.speed = speed
|
109 |
+
self.outbuf: asyncio.Queue[TTSResult | None] = asyncio.Queue()
|
110 |
+
self.is_closed = False
|
111 |
+
self.target_sample_rate = sample_rate
|
112 |
+
self.original_sample_rate = original_sample_rate
|
113 |
+
|
114 |
+
def on_process(self, chunk: np.ndarray, progress: float):
|
115 |
+
if self.is_closed:
|
116 |
+
return 0
|
117 |
+
|
118 |
+
# resample to target sample rate
|
119 |
+
if self.target_sample_rate != self.original_sample_rate:
|
120 |
+
num_samples = int(
|
121 |
+
len(chunk) * self.target_sample_rate / self.original_sample_rate)
|
122 |
+
resampled_chunk = resample(chunk, num_samples)
|
123 |
+
chunk = resampled_chunk.astype(np.float32)
|
124 |
+
|
125 |
+
scaled_chunk = chunk * 32768.0
|
126 |
+
clipped_chunk = np.clip(scaled_chunk, -32768, 32767)
|
127 |
+
int16_chunk = clipped_chunk.astype(np.int16)
|
128 |
+
samples = int16_chunk.tobytes()
|
129 |
+
self.outbuf.put_nowait(TTSResult(samples, False))
|
130 |
+
return self.is_closed and 0 or 1
|
131 |
+
|
132 |
+
async def write(self, text: str, split: bool, pause: float = 0.2):
|
133 |
+
start = time.time()
|
134 |
+
if split:
|
135 |
+
texts = re.split(splitter, text)
|
136 |
+
else:
|
137 |
+
texts = [text]
|
138 |
+
|
139 |
+
audio_duration = 0.0
|
140 |
+
audio_size = 0
|
141 |
+
|
142 |
+
for idx, text in enumerate(texts):
|
143 |
+
text = text.strip()
|
144 |
+
if not text:
|
145 |
+
continue
|
146 |
+
sub_start = time.time()
|
147 |
+
|
148 |
+
audio = await asyncio.to_thread(self.engine.generate,
|
149 |
+
text, self.sid, self.speed,
|
150 |
+
self.on_process)
|
151 |
+
|
152 |
+
if not audio or not audio.sample_rate or not audio.samples:
|
153 |
+
logger.error(f"tts: failed to generate audio for "
|
154 |
+
f"'{text}' (audio={audio})")
|
155 |
+
continue
|
156 |
+
|
157 |
+
if split and idx < len(texts) - 1: # add a pause between sentences
|
158 |
+
noise = np.zeros(int(audio.sample_rate * pause))
|
159 |
+
self.on_process(noise, 1.0)
|
160 |
+
audio.samples = np.concatenate([audio.samples, noise])
|
161 |
+
|
162 |
+
audio_duration += len(audio.samples) / audio.sample_rate
|
163 |
+
audio_size += len(audio.samples)
|
164 |
+
elapsed_seconds = time.time() - sub_start
|
165 |
+
logger.info(f"tts: generated audio for '{text}', "
|
166 |
+
f"audio duration: {audio_duration:.2f}s, "
|
167 |
+
f"elapsed: {elapsed_seconds:.2f}s")
|
168 |
+
|
169 |
+
elapsed_seconds = time.time() - start
|
170 |
+
logger.info(f"tts: generated audio in {elapsed_seconds:.2f}s, "
|
171 |
+
f"audio duration: {audio_duration:.2f}s")
|
172 |
+
|
173 |
+
r = TTSResult(None, True)
|
174 |
+
r.elapsed = elapsed_seconds
|
175 |
+
r.audio_duration = audio_duration
|
176 |
+
r.progress = 1.0
|
177 |
+
r.finished = True
|
178 |
+
await self.outbuf.put(r)
|
179 |
+
|
180 |
+
async def close(self):
|
181 |
+
self.is_closed = True
|
182 |
+
self.outbuf.put_nowait(None)
|
183 |
+
logger.info("tts: stream closed")
|
184 |
+
|
185 |
+
async def read(self) -> TTSResult:
|
186 |
+
return await self.outbuf.get()
|
187 |
+
|
188 |
+
async def generate(self, text: str) -> io.BytesIO:
|
189 |
+
start = time.time()
|
190 |
+
audio = await asyncio.to_thread(self.engine.generate,
|
191 |
+
text, self.sid, self.speed)
|
192 |
+
elapsed_seconds = time.time() - start
|
193 |
+
audio_duration = len(audio.samples) / audio.sample_rate
|
194 |
+
|
195 |
+
logger.info(f"tts: generated audio in {elapsed_seconds:.2f}s, "
|
196 |
+
f"audio duration: {audio_duration:.2f}s, "
|
197 |
+
f"sample rate: {audio.sample_rate}")
|
198 |
+
|
199 |
+
if self.target_sample_rate != audio.sample_rate:
|
200 |
+
audio.samples = resample(audio.samples,
|
201 |
+
int(len(audio.samples) * self.target_sample_rate / audio.sample_rate))
|
202 |
+
audio.sample_rate = self.target_sample_rate
|
203 |
+
|
204 |
+
output = io.BytesIO()
|
205 |
+
soundfile.write(output,
|
206 |
+
audio.samples,
|
207 |
+
samplerate=audio.sample_rate,
|
208 |
+
subtype="PCM_16",
|
209 |
+
format="WAV")
|
210 |
+
output.seek(0)
|
211 |
+
return output
|
212 |
+
|
213 |
+
|
214 |
+
async def start_tts_stream(sid: int, sample_rate: int, speed: float, args) -> TTSStream:
|
215 |
+
engine, original_sample_rate = get_tts_engine(args)
|
216 |
+
return TTSStream(engine, sid, speed, sample_rate, original_sample_rate)
|