PublicMirrors/voice-changer
1
0
Fork 0
mirror of https://github.com/w-okada/voice-changer.git synced 2025-01-23 21:45:00 +03:00
Code Issues Actions Packages Projects Releases Wiki Activity

delete padding

Browse Source
This commit is contained in:
nadare 2023-07-09 03:34:10 +09:00
parent b1d361aaae
commit 816798e8b5
4 changed files with 1 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
server/voice_changer/RVC/pitchExtractor/CrepeOnnxPitchExtractor.py
View File

@ -54,8 +54,6 @@ class CrepeOnnxPitchExtractor(PitchExtractor):
f0[pd < 0.1] = 0
f0 = f0.squeeze()
f0 = np.pad(f0, (start_frame, n_frames - f0.shape[0] - start_frame), 'constant', constant_values=(0, 0))
f0 *= pow(2, f0_up_key / 12)
pitchf[-f0.shape[0]:] = f0[:pitchf.shape[0]]
f0bak = pitchf.copy()

4
server/voice_changer/RVC/pitchExtractor/CrepePitchExtractor.py
View File

@ -47,10 +47,6 @@ class CrepePitchExtractor(PitchExtractor):
f0[pd < 0.1] = 0
f0 = f0.squeeze()
f0 = torch.nn.functional.pad(
f0, (start_frame, n_frames - f0.shape[0] - start_frame)
)
f0 *= pow(2, f0_up_key / 12)
pitchf[-f0.shape[0]:] = f0.detach().cpu().numpy()[:pitchf.shape[0]]
f0bak = pitchf.copy()

3
server/voice_changer/RVC/pitchExtractor/DioPitchExtractor.py
View File

@ -17,7 +17,7 @@ class DioPitchExtractor(PitchExtractor):
start_frame = int(silence_front * sr / window)
real_silence_front = start_frame * window / sr
silence_front_offset = int(np.round(real_silence_front * sr))
silence_front_offset = max(min(int(np.round(real_silence_front * sr)), len(audio) - 3000), 0)
audio = audio[silence_front_offset:]
f0_min = 50
@ -34,7 +34,6 @@ class DioPitchExtractor(PitchExtractor):
frame_period=10,
)
f0 = pyworld.stonemask(audio.astype(np.double), _f0, t, sr)
# f0 = np.pad(f0.astype("float"), (start_frame, n_frames - len(f0) - start_frame))
f0 *= pow(2, f0_up_key / 12)
pitchf[-f0.shape[0]:] = f0[:pitchf.shape[0]]

2
server/voice_changer/RVC/pitchExtractor/HarvestPitchExtractor.py
View File

@ -35,8 +35,6 @@ class HarvestPitchExtractor(PitchExtractor):
f0 = pyworld.stonemask(audio.astype(np.double), f0, t, sr)
f0 = signal.medfilt(f0, 3)
# f0 = np.pad(f0.astype("float"), (start_frame, n_frames - len(f0) - start_frame))
f0 *= pow(2, f0_up_key / 12)
pitchf[-f0.shape[0]:] = f0[:pitchf.shape[0]]
f0bak = pitchf.copy()