Skip to content

Instantly share code, notes, and snippets.

View r9y9's full-sized avatar
:shipit:
( ˘ω˘ ) zzz

Ryuichi Yamamoto r9y9

:shipit:
( ˘ω˘ ) zzz
View GitHub Profile
#!/usr/bin/env bash
JVSDIR=~/data/jvs_ver1
OUTDIR=~/data/jvs_r9y9_ver1
mkdir -p $OUTDIR
# link all wave file
for speaker in jvs{001..100}; do
JVSDIR=~/data/ita
OUTDIR=~/data/ita_r9y9
mkdir -p $OUTDIR
# link all wave file
for speaker in itako zundamon methane; do
echo $speaker
for corpus in recitation; do
mkdir -p $OUTDIR/$speaker/$corpus
#!/bin/bash
set -e
# 1. ita ディレクトリ (例. $HOME/data/ita) を作成し、このスクリプトと
# https://zunko.jp/multimodal_dev/login.php からダウンロードできる
# 音声データ(3話者分)を配置する
# 2. このスクリプトを実行する
for spk in itako methane zundamon; do
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@r9y9
r9y9 / model.py
Created November 15, 2020 05:22
TADE
# coding: utf-8
import torch
from torch import nn
from torch.nn import functional as F
from torch.nn.utils.rnn import pack_padded_sequence, pad_packed_sequence
from torch.nn.utils import weight_norm
from nnsvs.base import BaseModel, PredictionType
@r9y9
r9y9 / Musical context features-v2.ipynb
Last active November 2, 2020 08:40
Musical context features-v2
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
if self.use_harvest:
import amfm_decompy.pYAAPT as pYAAPT
import amfm_decompy.basic_tools as basic
signal = basic.SignalObj(wav_path)
print(min_f0, max_f0)
min_f0 = min(150, min_f0) # TODO: Fix this property
pitch = pYAAPT.yaapt(signal, f0_min=min_f0, f0_max=max_f0, frame_length=25, frame_space=5)
f0 = pitch.samp_values.astype(np.float64)
timeaxis = np.linspace(0, (pitch.samp_values.shape[0]-1) * 0.005, len(pitch.samp_values))
#f0, timeaxis = pyworld.harvest(x, fs, frame_period=self.frame_period,
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@r9y9
r9y9 / Kiritan singing voice synthesis demo.ipynb
Created May 3, 2020 01:43
Neural_network_based_singing_voice_synthesis_demo_using_kiritan_singing_database_(Japanese)
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
#!/bin/bash
NEUTRINO_DIR=~/sp/NEUTRINO
dst_dir=./sinsy_lab
mkdir -p $dst_dir/full
mkdir -p $dst_dir/mono
for f in musicxml/*.xml
do
name=$(basename $f)