1berry/wokada-voice-changer
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

MMVC minimum input length

Browse Source
This commit is contained in:
wataru 2023-05-26 23:04:56 +09:00
parent 18481feb7e
commit 39cd234268
3 changed files with 24 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
server/voice_changer/MMVCv13/MMVCv13.py
View File

@ -200,8 +200,8 @@ class MMVCv13:
convertSize = inputSize + crossfadeSize + solaSearchFrame
if convertSize < 8192:
convertSize = 8192
# if convertSize < 8192:
# convertSize = 8192
if convertSize % self.hps.data.hop_length != 0: # モデルの出力のホップサイズで切り捨てが発生するので補う。
convertSize = convertSize + (
self.hps.data.hop_length - (convertSize % self.hps.data.hop_length)

2
server/voice_changer/MMVCv15/MMVCv15.py
View File

@ -103,12 +103,14 @@ class MMVCv15:
requires_grad_dec=self.hps.requires_grad.dec,
)
if self.settings.pyTorchModelFile is not None:
self.settings.framework = "PyTorch"
self.net_g.eval()
load_checkpoint(self.settings.pyTorchModelFile, self.net_g, None)
# ONNXモデル生成
self.onxx_input_length = 8192
if self.settings.onnxModelFile is not None:
self.settings.framework = "ONNX"
providers, options = self.getOnnxExecutionProvider()
self.onnx_session = onnxruntime.InferenceSession(
self.settings.onnxModelFile,

38
server/voice_changer/VoiceChanger.py
View File

@ -134,8 +134,7 @@ class VoiceChanger:
vc: VoiceChanger = _vc
currentInputDeviceId = -1
currentInputSampleRate = -1
model_sampling_rate = -1
currentModelSamplingRate = -1
currentOutputDeviceId = -1
currentInputChunkNum = -1
while True:
@ -176,34 +175,43 @@ class VoiceChanger:
block_frame = currentInputChunkNum * 128
# sample rate precheck(alsa cannot use 40000?)
model_sampling_rate = self.voiceChanger.get_processing_sampling_rate()
try:
currentModelSamplingRate = (
self.voiceChanger.get_processing_sampling_rate()
)
except Exception as e:
print("[Voice Changer] ex: get_processing_sampling_rate", e)
continue
try:
with sd.Stream(
callback=self.audio_callback,
blocksize=block_frame,
samplerate=model_sampling_rate,
samplerate=currentModelSamplingRate,
dtype="float32",
channels=[currentInputChannelNum, currentOutputChannelNum],
):
pass
currentInputSampleRate = model_sampling_rate
vc.settings.serverInputAudioSampleRate = model_sampling_rate
print(f"[Voice Changer] sample rate {model_sampling_rate}")
vc.settings.serverInputAudioSampleRate = currentModelSamplingRate
vc.settings.inputSampleRate = currentModelSamplingRate
print(
f"[Voice Changer] sample rate {vc.settings.serverInputAudioSampleRate}"
)
except Exception as e:
print(
"[Voice Changer] ex: fallback to device default samplerate", e
"[Voice Changer] ex: fallback to device default samplerate",
e,
)
currentInputSampleRate = serverInputAudioDevice.default_samplerate
vc.settings.serverInputAudioSampleRate = (
serverInputAudioDevice.default_samplerate
)
vc.settings.inputSampleRate = vc.settings.serverInputAudioSampleRate
# main loop
try:
with sd.Stream(
callback=self.audio_callback,
blocksize=block_frame,
samplerate=currentInputSampleRate,
samplerate=vc.settings.serverInputAudioSampleRate,
dtype="float32",
channels=[currentInputChannelNum, currentOutputChannelNum],
):
@ -212,16 +220,10 @@ class VoiceChanger:
and currentInputDeviceId == vc.settings.serverInputDeviceId
and currentOutputDeviceId
== vc.settings.serverOutputDeviceId
and model_sampling_rate
and currentModelSamplingRate
== self.voiceChanger.get_processing_sampling_rate()
and currentInputChunkNum == vc.settings.serverReadChunkSize
):
vc.settings.serverInputAudioSampleRate = (
self.voiceChanger.get_processing_sampling_rate()
)
vc.settings.inputSampleRate = (
vc.settings.serverInputAudioSampleRate
)
time.sleep(2)
print(
"[Voice Changer] server audio",
@ -232,7 +234,7 @@ class VoiceChanger:
vc.settings.serverAudioStated,
currentInputDeviceId,
currentOutputDeviceId,
currentInputSampleRate,
vc.settings.serverInputAudioSampleRate,
currentInputChunkNum,
)