Text-To-Speech Synthesis on LJSpeech

16개 결과 · ⬇ CSV · JSON

Audio Quality MOS

Rank	Model	Audio Quality MOS	Pleasantness MOS	Word Error Rate (WER)	MOS	WER (%)	Extra Training Data	Paper	Code	Year
1	NaturalSpeech	4.56	–	–	–	–	✓	NaturalSpeech: End-to-End Text to Speech Synthesis with Human-Level Quality	microsoft/NeuralSpeech · daniilrobnikov/vits2 · heatz123/naturalspeech	2022
2	VITS	4.43	–	–	–	–	✓	NaturalSpeech: End-to-End Text to Speech Synthesis with Human-Level Quality	microsoft/NeuralSpeech · daniilrobnikov/vits2 · heatz123/naturalspeech	2022
3	Grad-TTS + HiFiGAN (1000 steps)	4.37	–	–	–	–	✓	Grad-TTS: A Diffusion Probabilistic Model for Text-to-Speech	huawei-noah/Speech-Backbones · keonlee9420/DiffGAN-TTS · keonlee9420/DiffSinger · +3	2021
4	Glow-TTS + HiFiGAN	4.34	–	–	–	–	✓	Glow-TTS: A Generative Flow for Text-to-Speech via Monotonic Alignment Search	coqui-ai/TTS · jaywalnut310/glow-tts · supertone-inc/super-monotonic-align · +3	2020
4	FastSpeech 2 + HiFiGAN	4.34	–	–	–	–	✓	NaturalSpeech: End-to-End Text to Speech Synthesis with Human-Level Quality	microsoft/NeuralSpeech · daniilrobnikov/vits2 · heatz123/naturalspeech	2022
6	FastSpeech 2 + HiFiGAN	4.32	–	–	–	–	✓	FastSpeech 2: Fast and High-Quality End-to-End Text to Speech	coqui-ai/TTS · PaddlePaddle/PaddleSpeech · TensorSpeech/TensorflowTTS · +34	2020
7	FastDiff (4 steps)	4.28	–	–	–	–	✓	FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis	Rongjiehuang/ProDiff · Rongjiehuang/FastDiff	2022
8	FastDiff-TTS	4.03	–	–	–	–	✓	FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis	Rongjiehuang/ProDiff · Rongjiehuang/FastDiff	2022
9	Transformer TTS (Mel + WaveGlow)	3.88	–	–	–	–	✓	Neural Speech Synthesis with Transformer Network	PaddlePaddle/PaddleSpeech · as-ideas/TransformerTTS · soobinseo/transformer-tts · +3	2018
10	FastSpeech (Mel + WaveGlow)	3.84	–	–	–	–	✓	FastSpeech: Fast, Robust and Controllable Text to Speech	coqui-ai/TTS · PaddlePaddle/PaddleSpeech · ming024/FastSpeech2 · +19	2019
11	OverFlow	3.37	–	2.30	–	–	✓	OverFlow: Putting flows on top of neural transducers for better TTS	coqui-ai/TTS · shivammehta25/OverFlow	2022
12	Merlin	2.4	–	–	–	–	✓	FastSpeech: Fast, Robust and Controllable Text to Speech	coqui-ai/TTS · PaddlePaddle/PaddleSpeech · ming024/FastSpeech2 · +19	2019
13	temp	1.25	–	–	–	–	✓
14	Flowtron	–	3.665	–	–	–	✓	Flowtron: an Autoregressive Flow-based Generative Network for Text-to-Speech Synthesis	NVIDIA/flowtron · NVIDIA/radtts · KathyReid/opensource-voice-tools	2020
15	Tacotron 2	–	3.521	–	–	–	✓	Flowtron: an Autoregressive Flow-based Generative Network for Text-to-Speech Synthesis	NVIDIA/flowtron · NVIDIA/radtts · KathyReid/opensource-voice-tools	2020
16	Matcha-TTS	–	–	–	3.84	2.09		Matcha-TTS: A fast TTS architecture with conditional flow matching	shivammehta25/Matcha-TTS	2023