Skip to content
This repository has been archived by the owner on Oct 19, 2024. It is now read-only.

feat: Add force slice option for cmd inference #150

Open
wants to merge 2 commits into
base: 4.0
Choose a base branch
from
Open

feat: Add force slice option for cmd inference #150

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3e1c3a2
feat: Add force slice option
MistEO Feb 27, 2023
39b8527
fix: fix force slice
MistEO Feb 27, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
56 changes: 38 additions & 18 deletions inference_main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,7 @@ def main():

# 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')
parser.add_argument('-sf', '--slice_forced', type=int, default=0, help='强制切片，会在db切片的基础上再切一次，填入切片长度（秒），0则不切片')
parser.add_argument('-d', '--device', type=str, default=None, help='推理设备，None则为自动选择cpu和gpu')
parser.add_argument('-ns', '--noice_scale', type=float, default=0.4, help='噪音级别，会影响咬字和音质，较为玄学')
parser.add_argument('-p', '--pad_seconds', type=float, default=0.5, help='推理音频pad秒数，由于未知原因开头结尾会有异响，pad一小段静音段后就不会出现')
Expand Down Expand Up @@ -69,29 +70,48 @@ def main():
for spk in spk_list:
audio = []
for (slice_tag, data) in audio_data:
print(f'#=====segment start, {round(len(data) / audio_sr, 3)}s======')
def len_in_secs(length):
return round(length / audio_sr, 3)

def get_length(data):
return int(np.ceil(len(data) / audio_sr * svc_model.target_sample))

data_secs = len_in_secs(len(data))
length = get_length(data)
print(f'#=====segment start, {data_secs}s======')

length = int(np.ceil(len(data) / audio_sr * svc_model.target_sample))
if slice_tag:
print('jump empty segment')
_audio = np.zeros(length)
audio.extend(list(infer_tool.pad_array(_audio, length)))
else:
# padd
pad_len = int(audio_sr * pad_seconds)
data = np.concatenate([np.zeros([pad_len]), data, np.zeros([pad_len])])
raw_path = io.BytesIO()
soundfile.write(raw_path, data, audio_sr, format="wav")
raw_path.seek(0)
out_audio, out_sr = svc_model.infer(spk, tran, raw_path,
cluster_infer_ratio=cluster_infer_ratio,
auto_predict_f0=auto_predict_f0,
noice_scale=noice_scale
)
_audio = out_audio.cpu().numpy()
pad_len = int(svc_model.target_sample * pad_seconds)
_audio = _audio[pad_len:-pad_len]

audio.extend(list(infer_tool.pad_array(_audio, length)))
def pad_and_infer(data):
# padd
pad_len = int(audio_sr * pad_seconds)
data = np.concatenate([np.zeros([pad_len]), data, np.zeros([pad_len])])
data_with_pad = io.BytesIO()
soundfile.write(data_with_pad, data, audio_sr, format="wav")
data_with_pad.seek(0)
out_audio, out_sr = svc_model.infer(spk, tran, data_with_pad,
cluster_infer_ratio=cluster_infer_ratio,
auto_predict_f0=auto_predict_f0,
noice_scale=noice_scale)
_audio = out_audio.cpu().numpy()
pad_len = int(svc_model.target_sample * pad_seconds)
_audio = _audio[pad_len:-pad_len]
return _audio

if args.slice_forced > 0 and data_secs > args.slice_forced:
step = int(args.slice_forced * audio_sr)
for i in range(0, len(data), step):
ends = i + step if i + step < len(data) else len(data)
print(f'forced slice {len_in_secs(i)}~{len_in_secs(ends)}')
_audio = pad_and_infer(data[i:ends])
audio.extend(list(infer_tool.pad_array(_audio, get_length(data[i:ends]))))
else:
_audio = pad_and_infer(data)
audio.extend(list(infer_tool.pad_array(_audio, length)))

key = "auto" if auto_predict_f0 else f"{tran}key"
cluster_name = "" if cluster_infer_ratio == 0 else f"_{cluster_infer_ratio}"
res_path = f'./results/{clean_name}_{key}_{spk}{cluster_name}.{wav_format}'
Expand Down