1
0
mirror of synced 2024-12-18 10:26:00 +01:00
Retrieval-based-Voice-Conve.../infer/modules/vc/pipeline.py

458 lines
16 KiB
Python
Raw Normal View History

2023-08-28 09:08:31 +02:00
import os
2023-08-19 12:56:43 +02:00
import sys
2023-08-28 09:08:31 +02:00
import traceback
import logging
logger = logging.getLogger(__name__)
2023-08-28 09:08:31 +02:00
from functools import lru_cache
2023-08-19 12:56:43 +02:00
from time import time as ttime
2023-08-28 09:08:31 +02:00
import faiss
import librosa
2023-08-19 12:56:43 +02:00
import numpy as np
import parselmouth
2023-08-28 09:08:31 +02:00
import pyworld
2023-08-19 12:56:43 +02:00
import torch
import torch.nn.functional as F
2023-08-28 09:08:31 +02:00
import torchcrepe
2023-08-19 12:56:43 +02:00
from scipy import signal
now_dir = os.getcwd()
sys.path.append(now_dir)
2023-08-19 12:57:09 +02:00
2023-08-19 12:56:43 +02:00
bh, ah = signal.butter(N=5, Wn=48, btype="high", fs=16000)
input_audio_path2wav = {}
@lru_cache
def cache_harvest_f0(input_audio_path, fs, f0max, f0min, frame_period):
audio = input_audio_path2wav[input_audio_path]
f0, t = pyworld.harvest(
audio,
fs=fs,
f0_ceil=f0max,
f0_floor=f0min,
frame_period=frame_period,
)
f0 = pyworld.stonemask(audio, f0, t, fs)
return f0
def change_rms(data1, sr1, data2, sr2, rate): # 1是输入音频2是输出音频,rate是2的占比
# print(data1.max(),data2.max())
rms1 = librosa.feature.rms(
y=data1, frame_length=sr1 // 2 * 2, hop_length=sr1 // 2
) # 每半秒一个点
rms2 = librosa.feature.rms(y=data2, frame_length=sr2 // 2 * 2, hop_length=sr2 // 2)
rms1 = torch.from_numpy(rms1)
rms1 = F.interpolate(
rms1.unsqueeze(0), size=data2.shape[0], mode="linear"
).squeeze()
rms2 = torch.from_numpy(rms2)
rms2 = F.interpolate(
rms2.unsqueeze(0), size=data2.shape[0], mode="linear"
).squeeze()
rms2 = torch.max(rms2, torch.zeros_like(rms2) + 1e-6)
data2 *= (
torch.pow(rms1, torch.tensor(1 - rate))
* torch.pow(rms2, torch.tensor(rate - 1))
).numpy()
return data2
class Pipeline(object):
def __init__(self, tgt_sr, config):
self.x_pad, self.x_query, self.x_center, self.x_max, self.is_half = (
config.x_pad,
config.x_query,
config.x_center,
config.x_max,
config.is_half,
)
self.sr = 16000 # hubert输入采样率
self.window = 160 # 每帧点数
self.t_pad = self.sr * self.x_pad # 每条前后pad时间
self.t_pad_tgt = tgt_sr * self.x_pad
self.t_pad2 = self.t_pad * 2
self.t_query = self.sr * self.x_query # 查询切点前后查询时间
self.t_center = self.sr * self.x_center # 查询切点位置
self.t_max = self.sr * self.x_max # 免查询时长阈值
self.device = config.device
2023-08-19 12:57:09 +02:00
2023-08-19 12:56:43 +02:00
def get_f0(
2023-08-19 12:57:09 +02:00
self,
input_audio_path,
x,
p_len,
f0_up_key,
f0_method,
filter_radius,
inp_f0=None,
2023-08-19 12:56:43 +02:00
):
global input_audio_path2wav
time_step = self.window / self.sr * 1000
f0_min = 50
f0_max = 1100
f0_mel_min = 1127 * np.log(1 + f0_min / 700)
f0_mel_max = 1127 * np.log(1 + f0_max / 700)
if f0_method == "pm":
f0 = (
parselmouth.Sound(x, self.sr)
.to_pitch_ac(
time_step=time_step / 1000,
voicing_threshold=0.6,
pitch_floor=f0_min,
pitch_ceiling=f0_max,
)
.selected_array["frequency"]
)
pad_size = (p_len - len(f0) + 1) // 2
if pad_size > 0 or p_len - len(f0) - pad_size > 0:
f0 = np.pad(
f0, [[pad_size, p_len - len(f0) - pad_size]], mode="constant"
)
elif f0_method == "harvest":
input_audio_path2wav[input_audio_path] = x.astype(np.double)
f0 = cache_harvest_f0(input_audio_path, self.sr, f0_max, f0_min, 10)
if filter_radius > 2:
f0 = signal.medfilt(f0, 3)
elif f0_method == "crepe":
model = "full"
# Pick a batch size that doesn't cause memory errors on your gpu
batch_size = 512
# Compute pitch using first gpu
audio = torch.tensor(np.copy(x))[None].float()
f0, pd = torchcrepe.predict(
audio,
self.sr,
self.window,
f0_min,
f0_max,
model,
batch_size=batch_size,
device=self.device,
return_periodicity=True,
)
pd = torchcrepe.filter.median(pd, 3)
f0 = torchcrepe.filter.mean(f0, 3)
f0[pd < 0.1] = 0
f0 = f0[0].cpu().numpy()
elif f0_method == "rmvpe":
if not hasattr(self, "model_rmvpe"):
from infer.lib.rmvpe import RMVPE
logger.info(
2023-09-01 08:11:55 +02:00
"Loading rmvpe model,%s" % "%s/rmvpe.pt" % os.environ["rmvpe_root"]
)
2023-08-19 12:56:43 +02:00
self.model_rmvpe = RMVPE(
"%s/rmvpe.pt" % os.environ["rmvpe_root"],
is_half=self.is_half,
device=self.device,
2023-08-19 12:56:43 +02:00
)
f0 = self.model_rmvpe.infer_from_audio(x, thred=0.03)
2023-08-19 12:57:09 +02:00
2023-09-17 00:37:15 +02:00
if "privateuseone" in str(self.device): # clean ortruntime memory
del self.model_rmvpe.model
del self.model_rmvpe
logger.info("Cleaning ortruntime memory")
2023-08-19 12:57:09 +02:00
2023-08-19 12:56:43 +02:00
f0 *= pow(2, f0_up_key / 12)
# with open("test.txt","w")as f:f.write("\n".join([str(i)for i in f0.tolist()]))
tf0 = self.sr // self.window # 每秒f0点数
if inp_f0 is not None:
delta_t = np.round(
(inp_f0[:, 0].max() - inp_f0[:, 0].min()) * tf0 + 1
).astype("int16")
replace_f0 = np.interp(
list(range(delta_t)), inp_f0[:, 0] * 100, inp_f0[:, 1]
)
2023-08-19 12:57:09 +02:00
shape = f0[self.x_pad * tf0 : self.x_pad * tf0 + len(replace_f0)].shape[0]
f0[self.x_pad * tf0 : self.x_pad * tf0 + len(replace_f0)] = replace_f0[
2023-08-19 12:56:43 +02:00
:shape
]
# with open("test_opt.txt","w")as f:f.write("\n".join([str(i)for i in f0.tolist()]))
f0bak = f0.copy()
f0_mel = 1127 * np.log(1 + f0 / 700)
f0_mel[f0_mel > 0] = (f0_mel[f0_mel > 0] - f0_mel_min) * 254 / (
f0_mel_max - f0_mel_min
) + 1
f0_mel[f0_mel <= 1] = 1
f0_mel[f0_mel > 255] = 255
f0_coarse = np.rint(f0_mel).astype(np.int32)
return f0_coarse, f0bak # 1-0
def vc(
2023-08-19 12:57:09 +02:00
self,
model,
net_g,
sid,
audio0,
pitch,
pitchf,
times,
index,
big_npy,
index_rate,
version,
protect,
2023-08-19 12:56:43 +02:00
): # ,file_index,file_big_npy
feats = torch.from_numpy(audio0)
if self.is_half:
feats = feats.half()
else:
feats = feats.float()
if feats.dim() == 2: # double channels
feats = feats.mean(-1)
assert feats.dim() == 1, feats.dim()
feats = feats.view(1, -1)
padding_mask = torch.BoolTensor(feats.shape).to(self.device).fill_(False)
inputs = {
"source": feats.to(self.device),
"padding_mask": padding_mask,
"output_layer": 9 if version == "v1" else 12,
}
t0 = ttime()
with torch.no_grad():
logits = model.extract_features(**inputs)
feats = model.final_proj(logits[0]) if version == "v1" else logits[0]
if protect < 0.5 and pitch is not None and pitchf is not None:
feats0 = feats.clone()
if (
not isinstance(index, type(None))
and not isinstance(big_npy, type(None))
and index_rate != 0
):
npy = feats[0].cpu().numpy()
if self.is_half:
npy = npy.astype("float32")
# _, I = index.search(npy, 1)
# npy = big_npy[I.squeeze()]
score, ix = index.search(npy, k=8)
weight = np.square(1 / score)
weight /= weight.sum(axis=1, keepdims=True)
npy = np.sum(big_npy[ix] * np.expand_dims(weight, axis=2), axis=1)
if self.is_half:
npy = npy.astype("float16")
feats = (
torch.from_numpy(npy).unsqueeze(0).to(self.device) * index_rate
+ (1 - index_rate) * feats
)
feats = F.interpolate(feats.permute(0, 2, 1), scale_factor=2).permute(0, 2, 1)
if protect < 0.5 and pitch is not None and pitchf is not None:
feats0 = F.interpolate(feats0.permute(0, 2, 1), scale_factor=2).permute(
0, 2, 1
)
t1 = ttime()
p_len = audio0.shape[0] // self.window
if feats.shape[1] < p_len:
p_len = feats.shape[1]
if pitch is not None and pitchf is not None:
pitch = pitch[:, :p_len]
pitchf = pitchf[:, :p_len]
if protect < 0.5 and pitch is not None and pitchf is not None:
pitchff = pitchf.clone()
pitchff[pitchf > 0] = 1
pitchff[pitchf < 1] = protect
pitchff = pitchff.unsqueeze(-1)
feats = feats * pitchff + feats0 * (1 - pitchff)
feats = feats.to(feats0.dtype)
p_len = torch.tensor([p_len], device=self.device).long()
with torch.no_grad():
2023-09-01 08:11:55 +02:00
hasp = pitch is not None and pitchf is not None
arg = (feats, p_len, pitch, pitchf, sid) if hasp else (feats, p_len, sid)
audio1 = (net_g.infer(*arg)[0][0, 0]).data.cpu().float().numpy()
2023-09-01 08:11:55 +02:00
del hasp, arg
2023-08-19 12:56:43 +02:00
del feats, p_len, padding_mask
if torch.cuda.is_available():
torch.cuda.empty_cache()
t2 = ttime()
times[0] += t1 - t0
times[2] += t2 - t1
return audio1
def pipeline(
2023-08-19 12:57:09 +02:00
self,
model,
net_g,
sid,
audio,
input_audio_path,
times,
f0_up_key,
f0_method,
file_index,
index_rate,
if_f0,
filter_radius,
tgt_sr,
resample_sr,
rms_mix_rate,
version,
protect,
f0_file=None,
2023-08-19 12:56:43 +02:00
):
if (
file_index != ""
# and file_big_npy != ""
# and os.path.exists(file_big_npy) == True
and os.path.exists(file_index)
and index_rate != 0
):
try:
index = faiss.read_index(file_index)
# big_npy = np.load(file_big_npy)
big_npy = index.reconstruct_n(0, index.ntotal)
except:
traceback.print_exc()
index = big_npy = None
else:
index = big_npy = None
audio = signal.filtfilt(bh, ah, audio)
audio_pad = np.pad(audio, (self.window // 2, self.window // 2), mode="reflect")
opt_ts = []
if audio_pad.shape[0] > self.t_max:
audio_sum = np.zeros_like(audio)
for i in range(self.window):
audio_sum += np.abs(audio_pad[i : i - self.window])
2023-08-19 12:56:43 +02:00
for t in range(self.t_center, audio.shape[0], self.t_center):
opt_ts.append(
t
- self.t_query
+ np.where(
audio_sum[t - self.t_query : t + self.t_query]
== audio_sum[t - self.t_query : t + self.t_query].min()
2023-08-19 12:56:43 +02:00
)[0][0]
)
s = 0
audio_opt = []
t = None
t1 = ttime()
audio_pad = np.pad(audio, (self.t_pad, self.t_pad), mode="reflect")
p_len = audio_pad.shape[0] // self.window
inp_f0 = None
if hasattr(f0_file, "name"):
try:
with open(f0_file.name, "r") as f:
lines = f.read().strip("\n").split("\n")
inp_f0 = []
for line in lines:
inp_f0.append([float(i) for i in line.split(",")])
inp_f0 = np.array(inp_f0, dtype="float32")
except:
traceback.print_exc()
sid = torch.tensor(sid, device=self.device).unsqueeze(0).long()
pitch, pitchf = None, None
if if_f0 == 1:
pitch, pitchf = self.get_f0(
input_audio_path,
audio_pad,
p_len,
f0_up_key,
f0_method,
filter_radius,
inp_f0,
)
pitch = pitch[:p_len]
pitchf = pitchf[:p_len]
2023-09-17 00:37:15 +02:00
if "mps" not in str(self.device) or "xpu" not in str(self.device):
2023-08-19 12:56:43 +02:00
pitchf = pitchf.astype(np.float32)
pitch = torch.tensor(pitch, device=self.device).unsqueeze(0).long()
pitchf = torch.tensor(pitchf, device=self.device).unsqueeze(0).float()
t2 = ttime()
times[1] += t2 - t1
for t in opt_ts:
t = t // self.window * self.window
if if_f0 == 1:
audio_opt.append(
self.vc(
model,
net_g,
sid,
2023-08-19 12:57:09 +02:00
audio_pad[s : t + self.t_pad2 + self.window],
pitch[:, s // self.window : (t + self.t_pad2) // self.window],
pitchf[:, s // self.window : (t + self.t_pad2) // self.window],
2023-08-19 12:56:43 +02:00
times,
index,
big_npy,
index_rate,
version,
protect,
2023-08-19 12:57:09 +02:00
)[self.t_pad_tgt : -self.t_pad_tgt]
2023-08-19 12:56:43 +02:00
)
else:
audio_opt.append(
self.vc(
model,
net_g,
sid,
2023-08-19 12:57:09 +02:00
audio_pad[s : t + self.t_pad2 + self.window],
2023-08-19 12:56:43 +02:00
None,
None,
times,
index,
big_npy,
index_rate,
version,
protect,
2023-08-19 12:57:09 +02:00
)[self.t_pad_tgt : -self.t_pad_tgt]
2023-08-19 12:56:43 +02:00
)
s = t
if if_f0 == 1:
audio_opt.append(
self.vc(
model,
net_g,
sid,
audio_pad[t:],
2023-08-19 12:57:09 +02:00
pitch[:, t // self.window :] if t is not None else pitch,
pitchf[:, t // self.window :] if t is not None else pitchf,
2023-08-19 12:56:43 +02:00
times,
index,
big_npy,
index_rate,
version,
protect,
2023-08-19 12:57:09 +02:00
)[self.t_pad_tgt : -self.t_pad_tgt]
2023-08-19 12:56:43 +02:00
)
else:
audio_opt.append(
self.vc(
model,
net_g,
sid,
audio_pad[t:],
None,
None,
times,
index,
big_npy,
index_rate,
version,
protect,
2023-08-19 12:57:09 +02:00
)[self.t_pad_tgt : -self.t_pad_tgt]
2023-08-19 12:56:43 +02:00
)
audio_opt = np.concatenate(audio_opt)
if rms_mix_rate != 1:
audio_opt = change_rms(audio, 16000, audio_opt, tgt_sr, rms_mix_rate)
if tgt_sr != resample_sr >= 16000:
audio_opt = librosa.resample(
audio_opt, orig_sr=tgt_sr, target_sr=resample_sr
)
audio_max = np.abs(audio_opt).max() / 0.99
max_int16 = 32768
if audio_max > 1:
max_int16 /= audio_max
audio_opt = (audio_opt * max_int16).astype(np.int16)
del pitch, pitchf, sid
if torch.cuda.is_available():
torch.cuda.empty_cache()
return audio_opt