Cool_Tools/ultimatevocalremovergui
1
0
Fork 0
mirror of https://github.com/Anjok07/ultimatevocalremovergui.git synced 2024-11-24 07:20:10 +01:00
Code Issues Packages Projects Releases Wiki Activity

Add files via upload

Browse Source
This commit is contained in:
Anjok07 2023-01-11 03:57:27 -06:00 committed by GitHub
parent de582efd02
commit 95ad38be11
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 20 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
separate.py
View File

@ -12,6 +12,7 @@ from lib_v5.vr_network import nets_new
#from lib_v5.vr_network.model_param_init import ModelParameters #from lib_v5.vr_network.model_param_init import ModelParameters
from pathlib import Path from pathlib import Path
from gui_data.constants import * from gui_data.constants import *
import audioread
import gzip import gzip
import librosa import librosa
import math import math
@ -672,9 +673,11 @@ class SeperateVR(SeperateAttributes):
wav_resolution = bp['res_type'] wav_resolution = bp['res_type']
if d == bands_n: # high-end band if d == bands_n: # high-end band
X_wave[d], _ = librosa.load( X_wave[d], _ = librosa.load(self.audio_file, bp['sr'], False, dtype=np.float32, res_type=wav_resolution)
self.audio_file, bp['sr'], False, dtype=np.float32, res_type=wav_resolution)
if not X_wave[d] and self.audio_file.endswith('.mp3'):
X_wave[d] = rerun_mp3(self.audio_file, bp['sr'])
if X_wave[d].ndim == 1: if X_wave[d].ndim == 1:
X_wave[d] = np.asarray([X_wave[d], X_wave[d]]) X_wave[d] = np.asarray([X_wave[d], X_wave[d]])
else: # lower bands else: # lower bands
@ -865,13 +868,20 @@ def gather_sources(primary_stem_name, secondary_stem_name, secondary_sources: di
def prepare_mix(mix, chunk_set, margin_set, mdx_net_cut=False, is_missing_mix=False): def prepare_mix(mix, chunk_set, margin_set, mdx_net_cut=False, is_missing_mix=False):
audio_path = mix
samplerate = 44100 samplerate = 44100
print('mix first: ', mix)
if not isinstance(mix, np.ndarray): if not isinstance(mix, np.ndarray):
mix, samplerate = librosa.load(mix, mono=False, sr=44100) mix, samplerate = librosa.load(mix, mono=False, sr=44100)
else: else:
mix = mix.T mix = mix.T
print('mix: ', mix)
if not mix and audio_path.endswith('.mp3'):
mix = rerun_mp3(audio_path)
print('mix after fix: ', mix)
if mix.ndim == 1: if mix.ndim == 1:
mix = np.asfortranarray([mix,mix]) mix = np.asfortranarray([mix,mix])
@ -906,6 +916,13 @@ def prepare_mix(mix, chunk_set, margin_set, mdx_net_cut=False, is_missing_mix=Fa
raw_mix = get_segmented_mix(chunk_set=0) if mdx_net_cut else mix raw_mix = get_segmented_mix(chunk_set=0) if mdx_net_cut else mix
return segmented_mix, raw_mix, samplerate return segmented_mix, raw_mix, samplerate
def rerun_mp3(audio_file, sample_rate=44100):
with audioread.audio_open(audio_file) as f:
track_length = int(f.duration)
return librosa.load(audio_file, duration=track_length, mono=False, sr=sample_rate)[0]
def save_format(audio_path, save_format, mp3_bit_set): def save_format(audio_path, save_format, mp3_bit_set):
if not save_format == WAV: if not save_format == WAV: