NNSVS demos
Singing voice synthesis (SVS) demo using nnsvs. All the models were trained using https://github.com/nnsvs/nnsvs. Recipes to reproduce experiments are included in the repository.
NOTE: Due to the license issues, pre-trained models are not provided except for the ones trained on NIT-song070 database. The use of pre-traiend models are only permitted for research purpose.
Preparation
[1]:
%%capture
try:
import nnsvs
except ImportError:
! pip install git+https://github.com/nnsvs/nnsvs
[2]:
%pylab inline
%load_ext autoreload
%autoreload
import IPython
from IPython.display import Audio
import pysinsy
import librosa
from nnmnkwii.io import hts
%pylab is deprecated, use %matplotlib inline and import the required libraries.
Populating the interactive namespace from numpy and matplotlib
[3]:
from nnsvs.pretrained import create_svs_engine
import nnsvs
Sample 1
[5]:
engine = create_svs_engine("r9y9/yoko_latest")
contexts = pysinsy.extract_fullcontext(nnsvs.util.example_xml_file("song070_f00001_063"))
labels = hts.HTSLabelFile.create_from_contexts(contexts)
wav, sr = engine.svs(labels)
Audio(wav, rate=sr)
[5]: