Paper: https://arxiv.org/abs/2203.16937

Inference code: https://github.com/tinkoff-ai/hifi_vc

Many-to-Many

Source

062.wav

062.wav

291.wav

587.wav

786.wav

Target

000.wav

062.wav

291.wav

587.wav

786.wav

NVCNet

000_e6b671437618b326b1e3e6f4ea20ace9.wav

062_154e86dfa4ef0a9484a2e746ded84052.wav

291_29e741e5ee209a1621c7ad8e0e512da6.wav

587_e6b671437618b326b1e3e6f4ea20ace9.wav

786_fea595bec0383ffe372acccc70b18564.wav

PPG-VC

000_e6b671437618b326b1e3e6f4ea20ace9.wav

062_154e86dfa4ef0a9484a2e746ded84052.wav

291_29e741e5ee209a1621c7ad8e0e512da6.wav

587_e6b671437618b326b1e3e6f4ea20ace9.wav

786_fea595bec0383ffe372acccc70b18564.wav

VQMIVC

000_e6b671437618b326b1e3e6f4ea20ace9.wav

062_154e86dfa4ef0a9484a2e746ded84052.wav

291_29e741e5ee209a1621c7ad8e0e512da6.wav

587_e6b671437618b326b1e3e6f4ea20ace9.wav

786_fea595bec0383ffe372acccc70b18564.wav

StarGANv2-VC

000_e6b671437618b326b1e3e6f4ea20ace9.wav

062_154e86dfa4ef0a9484a2e746ded84052.wav

291_29e741e5ee209a1621c7ad8e0e512da6.wav

587_e6b671437618b326b1e3e6f4ea20ace9.wav

786_fea595bec0383ffe372acccc70b18564.wav

HiFi-VC (ours)

000_e6b671437618b326b1e3e6f4ea20ace9.wav

062_154e86dfa4ef0a9484a2e746ded84052.wav

291_29e741e5ee209a1621c7ad8e0e512da6.wav

587_e6b671437618b326b1e3e6f4ea20ace9.wav

786_fea595bec0383ffe372acccc70b18564.wav

Any-to-Any

Source

021.wav

064.wav

111.wav

132.wav

Target

021.wav

064.wav

111.wav

132.wav

NVCNet

021_78914c4dd7092c0d848ed49b3eaeb6e4.wav

064_198daa781faf18138c8407ac8648cda2.wav

111_02109d0312b3c8bcf1c5729fc5b6667c.wav

132_0756a90fbbb1806c72c690e3a879448e.wav

PPG-VC

021_78914c4dd7092c0d848ed49b3eaeb6e4.wav

064_198daa781faf18138c8407ac8648cda2.wav

111_02109d0312b3c8bcf1c5729fc5b6667c.wav

132_0756a90fbbb1806c72c690e3a879448e.wav

VQMIVC

021_78914c4dd7092c0d848ed49b3eaeb6e4.wav

064_198daa781faf18138c8407ac8648cda2.wav

111_02109d0312b3c8bcf1c5729fc5b6667c.wav

132_0756a90fbbb1806c72c690e3a879448e.wav

HiFi-VC (ours)

021_78914c4dd7092c0d848ed49b3eaeb6e4.wav

064_198daa781faf18138c8407ac8648cda2.wav

111_02109d0312b3c8bcf1c5729fc5b6667c.wav

132_0756a90fbbb1806c72c690e3a879448e.wav

Ablations (M2M)

Source

113.wav

333.wav

489.wav

634 .wav

Target

113.wav

333.wav

489.wav

634.wav

no asr

113_4b7f4835f57418d2abe37d8747c617a2.wav

333_705d4758a3c0090390d6f7ebe68df871.wav

489_26258cf34e562aecb7eddd5a1d9acc97.wav

634_beb480faea6975a45eafeb16c36a3eb1.wav

no f0

113_4b7f4835f57418d2abe37d8747c617a2.wav

333_705d4758a3c0090390d6f7ebe68df871.wav

489_26258cf34e562aecb7eddd5a1d9acc97.wav

634_beb480faea6975a45eafeb16c36a3eb1.wav

HiFi-VC (ours)

113_4b7f4835f57418d2abe37d8747c617a2.wav

333_705d4758a3c0090390d6f7ebe68df871.wav