SOTAVerified

Text-To-Speech Synthesis

Text-To-Speech Synthesis is a machine learning task that involves converting written text into spoken words. The goal is to generate synthetic speech that sounds natural and resembles human speech as closely as possible.

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 201–225 of 332 papers

TitleStatusHype
Voice Conversion by Cascading Automatic Speech Recognition and Text-to-Speech Synthesis with Prosody Transfer—0
Multi-speaker Multi-style Text-to-speech Synthesis With Single-speaker Single-style Training Data Scenarios—0
Multi-speaker Text-to-speech Synthesis Using Deep Gaussian Processes—0
Multi-Stage Deep Transfer Learning for EmIoT-enabled Human-Computer Interaction—0
Multi-step Natural Language Understanding—0
Grad-StyleSpeech: Any-speaker Adaptive Text-to-Speech Synthesis with Diffusion Models—0
An Overview of Affective Speech Synthesis and Conversion in the Deep Learning Era—0
Neural Harmonic-plus-Noise Waveform Model with Trainable Maximum Voice Frequency for Text-to-Speech Synthesis—0
Neural Models of Text Normalization for Speech Applications—0
Neural Speech Synthesis in German—0
A Novel Data Augmentation Approach for Automatic Speaking Assessment on Opinion Expressions—0
Neural Text Normalization with Subword Units—0
Neural Text-to-Speech Synthesis for an Under-Resourced Language in a Diglossic Environment: the Case of Gascon Occitan—0
Noise-robust zero-shot text-to-speech synthesis conditioned on self-supervised speech-representation model with adapters—0
Normalization of Lithuanian Text Using Regular Expressions—0
Normalization of Non-Standard Words in Croatian Texts—0
Normalizing Text using Language Modelling based on Phonetics and String Similarity—0
Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing—0
VQTTS: High-Fidelity Text-to-Speech Synthesis with Self-Supervised VQ Acoustic Feature—0
An objective evaluation of the effects of recording conditions and speaker characteristics in multi-speaker deep neural speech synthesis—0
An In-depth Analysis of the Effect of Text Normalization in Social Media—0
Parallel WaveNet conditioned on VAE latent vectors—0
ParrotTTS: Text-to-Speech synthesis by exploiting self-supervised representations—0
Phonetic Enhanced Language Modeling for Text-to-Speech Synthesis—0
PnG BERT: Augmented BERT on Phonemes and Graphemes for Neural TTS—0
Show:102550
← PrevPage 9 of 14Next →
All datasetsLJSpeech20000 utterancesCMUDict 0.7bHUI speech corpusThorsten voice 21.02 neutralTrinity Speech-Gesture Dataset

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1NaturalSpeechAudio Quality MOS4.56—Unverified
2VITSAudio Quality MOS4.43—Unverified
3Grad-TTS + HiFiGAN (1000 steps)Audio Quality MOS4.37—Unverified
4FastSpeech 2 + HiFiGANAudio Quality MOS4.34—Unverified
5Glow-TTS + HiFiGANAudio Quality MOS4.34—Unverified
6FastSpeech 2 + HiFiGANAudio Quality MOS4.32—Unverified
7FastDiff (4 steps)Audio Quality MOS4.28—Unverified
8FastDiff-TTSAudio Quality MOS4.03—Unverified
9Transformer TTS (Mel + WaveGlow)Audio Quality MOS3.88—Unverified
10FastSpeech (Mel + WaveGlow)Audio Quality MOS3.84—Unverified
#ModelMetricClaimedVerifiedStatus
1Mia10-keyword Speech Commands dataset16—Unverified
#ModelMetricClaimedVerifiedStatus
1Token-Level Ensemble DistillationPhoneme Error Rate4.6—Unverified
#ModelMetricClaimedVerifiedStatus
1Tacotron 2Mean Opinion Score3.74—Unverified
#ModelMetricClaimedVerifiedStatus
1Tacotron 2Mean Opinion Score3.49—Unverified
#ModelMetricClaimedVerifiedStatus
1Match-TTSGMOS3.7—Unverified