github.com/r9y9/nnsvs

Test on a Neural source filter (NSF) model
NOTE: Training is still in progress (350k steps out of 400k)
NOTE: Haven't done any optimization yet

ust: www.nicovideo.jp/watch/sm39319737

Statistical parametric SVS (sampling rate: 48000)
Time-lag model: {
"_target_": "nnsvs.model.MDN",
"in_dim": 331,
"out_dim": 1,
"hidden_dim": 32,
"num_layers": 3,
"dropout": 0.5,
"num_gaussians": 1
}
Duration model: {
"_target_": "nnsvs.model.MDN",
"in_dim": 331,
"out_dim": 1,
"hidden_dim": 256,
"num_layers": 3,
"dropout": 0.5,
"num_gaussians": 1
}
Acoustic model: {
"_target_": "nnsvs.model.ResSkipF0FFConvLSTM",
"in_dim": 335,
"out_dim": 206,
"ff_hidden_dim": 1024,
"conv_hidden_dim": 512,
"lstm_hidden_dim": 256,
"dropout": 0.0,
"bidirectional": true,
"in_lf0_idx": 286,
"out_lf0_idx": 180,
"in_lf0_min": 5.162578582763672,
"in_lf0_max": 6.491110801696777,
"out_lf0_mean": 5.885598413281433,
"out_lf0_scale": 0.23067350503371858
}
Vocoder model: {
"generator_type": "HnSincNSF",
"generator_params": {
"aux_context_window": 2,
"cin_channels": 67,
"cnn_kernel_s": 5,
"cnn_num_in_block": 10,
"filter_block_num": 5,
"harmonic_num": 7,
"hidden_dim": 64,
"noise_std": 0.003,
"out_channels": 1,
"out_lf0_idx": 60,
"out_lf0_mean": 5.885598413281433,
"out_lf0_scale": 0.23067350503371858,
"out_vuv_idx": 61,
"sample_rate": 48000,
"sinc_order": 31,
"sine_amp": 0.1,
"upsample_net": "ConvInUpsampleNetwork",
"upsample_params": {
"upsample_scales": [
2,
3,
4,
10
],
"aux_channels": 67,
"aux_context_window": 2
},
"upsample_rate": 240
}
}

2022/04/06 Kiritan test SVS w/ NSF 350k steps Roshin Yuukai

r9y9