Nofated095
/
so-vits-svc
mirror of https://github.com/svc-develop-team/so-vits-svc.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Updata F0 mean pooling

This commit is contained in:
ylzz1997 2023-04-05 20:02:32 +08:00
parent 18f747d131
commit f25e5667d7
2 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
inference_main.py
View File

@ -35,7 +35,7 @@ def main():
parser.add_argument('-cm', '--cluster_model_path', type=str, default="logs/44k/kmeans_10000.pt", help='聚类模型路径，如果没有训练聚类则随便填')
parser.add_argument('-cr', '--cluster_infer_ratio', type=float, default=0, help='聚类方案占比，范围0-1，若没有训练聚类模型则默认0即可')
parser.add_argument('-lg', '--linear_gradient', type=float, default=0, help='两段音频切片的交叉淡入长度，如果强制切片后出现人声不连贯可调整该数值，如果连贯建议采用默认值0，单位为秒')
parser.add_argument('-fmp', '--f0_mean_pooling', type=bool, default=False, help='是否对F0使用均值滤波器(池化)，对部分哑音可能有改善。注意，启动该选项会导致推理速度下降，默认关闭')
parser.add_argument('-fmp', '--f0_mean_pooling', type=bool, default=False, help='是否对F0使用均值滤波器(池化)，对部分哑音有改善。注意，启动该选项会导致推理速度下降，默认关闭')
# 不用动的部分
parser.add_argument('-sd', '--slice_db', type=int, default=-40, help='默认-40，嘈杂的音频可以-30，干声保留呼吸可以-50')

13
webUI.py
View File

@ -31,7 +31,7 @@ if torch.cuda.is_available():
for i in range(torch.cuda.device_count()):
cuda.append("cuda:{}".format(i))
def vc_fn(sid, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num):
def vc_fn(sid, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,F0_mean_pooling):
global model
try:
if input_audio is None:
@ -45,7 +45,7 @@ def vc_fn(sid, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise
audio = librosa.to_mono(audio.transpose(1, 0))
temp_path = "temp.wav"
soundfile.write(temp_path, audio, sampling_rate, format="wav")
_audio = model.slice_inference(temp_path, sid, vc_transform, slice_db, cluster_ratio, auto_f0, noise_scale,pad_seconds,cl_num,lg_num,lgr_num)
_audio = model.slice_inference(temp_path, sid, vc_transform, slice_db, cluster_ratio, auto_f0, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,F0_mean_pooling)
model.clear_empty()
os.remove(temp_path)
#构建保存文件的路径，并保存到results文件夹内
@ -80,7 +80,7 @@ def tts_func(_text,_rate):
p.wait()
return output_file
def vc_fn2(sid, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,text2tts,tts_rate):
def vc_fn2(sid, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,text2tts,tts_rate,F0_mean_pooling):
#使用edge-tts把文字转成音频
output_file=tts_func(text2tts,tts_rate)
@ -97,7 +97,7 @@ def vc_fn2(sid, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, nois
sample_rate, data=gr_pu.audio_from_file(save_path2)
vc_input=(sample_rate, data)
a,b=vc_fn(sid, vc_input, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num)
a,b=vc_fn(sid, vc_input, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,F0_mean_pooling)
os.remove(output_file)
os.remove(save_path2)
return a,b
@ -137,6 +137,7 @@ with app:
vc_transform = gr.Number(label="变调（整数，可以正负，半音数量，升高八度就是12）", value=0)
cluster_ratio = gr.Number(label="聚类模型混合比例，0-1之间，默认为0不启用聚类，能提升音色相似度，但会导致咬字下降（如果使用建议0.5左右）", value=0)
auto_f0 = gr.Checkbox(label="自动f0预测，配合聚类模型f0预测效果更好,会导致变调功能失效（仅限转换语音，歌声不要勾选此项会究极跑调）", value=False)
F0_mean_pooling = gr.Checkbox(label="是否对F0使用均值滤波器(池化)，对部分哑音有改善。注意，启动该选项会导致推理速度下降，默认关闭", value=False)
slice_db = gr.Number(label="切片阈值", value=-40)
noise_scale = gr.Number(label="noise_scale 建议不要动，会影响音质，玄学参数", value=0.4)
cl_num = gr.Number(label="音频自动切片，0为不切片，单位为秒/s", value=0)
@ -163,8 +164,8 @@ with app:
return sid.update(choices = spks,value=spks[0]),"ok,模型被加载到了设备{}之上".format(device_name)
except Exception as e:
return "","异常信息:"+str(e)+"\n请排障后重试"
vc_submit.click(vc_fn, [sid, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num], [vc_output1, vc_output2])
vc_submit2.click(vc_fn2, [sid, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,text2tts,tts_rate], [vc_output1, vc_output2])
vc_submit.click(vc_fn, [sid, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,F0_mean_pooling], [vc_output1, vc_output2])
vc_submit2.click(vc_fn2, [sid, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale,pad_seconds,cl_num,lg_num,lgr_num,text2tts,tts_rate,F0_mean_pooling], [vc_output1, vc_output2])
model_analysis_button.click(modelAnalysis,[model_path,config_path,cluster_model_path,device],[sid,sid_output])
app.launch()