WavTTS is the first raw-waveform diffusion TTS model using DiT flow matching and multi-scale mel supervision that approaches SOTA latent zero-shot performance while beating prior end-to-end models.
hub
Vall-e 2: Neural codec language models are human parity zero-shot text to speech synthesizers
25 Pith papers cite this work. Polarity classification is still indexing.
25
Pith papers citing it
abstract
This paper introduces VALL-E 2, the latest advancement in neural codec language models that marks a milestone in zero-shot text-to-speech synthesis (TTS), achieving human parity for the first time. Based on its predecessor, VALL-E, the new iteration introduces two significant enhancements: Repetition Aware Sampling refines the original nucleus sampling process by accounting for token repetition in the decoding history. It not only stabilizes the decoding but also circumvents the infinite loop issue. Grouped Code Modeling organizes codec codes into groups to effectively shorten the sequence length, which not only boosts inference speed but also addresses the challenges of long sequence modeling. Our experiments on the LibriSpeech and VCTK datasets show that VALL-E 2 surpasses previous systems in speech robustness, naturalness, and speaker similarity. It is the first of its kind to reach human parity on these benchmarks. Moreover, VALL-E 2 consistently synthesizes high-quality speech, even for sentences that are traditionally challenging due to their complexity or repetitive phrases. The advantages of this work could contribute to valuable endeavors, such as generating speech for individuals with aphasia or people with amyotrophic lateral sclerosis. See https://aka.ms/valle2 for demos of VALL-E 2.
hub tools
citation-role summary
background 3
baseline 1
citation-polarity summary
representative citing papers
Semantic priors from HuBERT and Whisper improve speech codec intelligibility up to 6 kbps but show diminishing returns beyond that, with a bitrate-aware regulation strategy balancing semantic consistency and naturalness.
CapTalk unifies single-utterance and dialogue voice design via utterance- and speaker-level captions plus a hierarchical variational module for stable timbre with adaptive expression.
MSpoof-TTS improves zero-shot discrete speech synthesis by integrating multi-resolution token-based spoof detection into a hierarchical decoding process that prunes low-quality candidates.
Audex unifies audio understanding and generation on a strong text MoE backbone with multi-stage SFT plus text-only Cascade RL, matching open SOTA audio scores while mostly retaining text capability.
RTFree-F5 replaces reference transcripts with mapped self-supervised speech representations in F5-TTS, cutting WER on dysarthric speech from 24.6% to 10.4% without any transcript at inference.
PASQA predicts pitch-accent quality using synthetic accent-error data, mora-conditioned fusion, ranking loss, and auxiliary localization, outperforming conventional MOS models in ordering by error severity and human agreement.
TLDR groups codec tokens into patches for patch-level autoregressive modeling in pretrained TTS systems, yielding 1.8x speedup and 75% KV-cache reduction at patch size 4.
UniVocal presents a text-context-only framework for speech-singing code-switching synthesis via two-stage curriculum learning and a synthetic data pipeline, claiming SOTA on a new benchmark.
EventSpeech is a text-conditioned neural framework that uses neuromorphic event cameras and a new EVT-SPK benchmark to generate expressive speech, claiming to outperform RGB baselines by preserving fine-grained emotions without motion blur.
FMelCodec is a three-stage mel-spectrogram codec using 640x VQ compression, conditional flow matching refinement, and HiFi-GAN reconstruction that reports higher quality than prior methods at 250 bps for 16 kHz speech.
Large-scale listening study of 35,532 judgments finds human accuracy on real audio fell from 72.7% to 64.1% since 2021 while fake detection remained stable, indicating a skepticism shift toward genuine speech.
AST performs training-free text-based speech editing by stitching inverted source latents with synthesized targets and adaptively guiding the flow-matching decoder, achieving state-of-the-art temporal fidelity and speaker preservation.
ASPIRin decouples speaking timing from token content via binary action space projection and applies GRPO with rule-based rewards to optimize interactivity in SLMs without semantic collapse or repetition.
CoMelSinger introduces a discrete token-based zero-shot SVS framework on MaskGCT with coarse-to-fine contrastive learning and an SVT module to improve melody control and reduce prosody leakage.
CosyVoice 3 achieves better content consistency, speaker similarity, and prosody naturalness in zero-shot multilingual speech synthesis by scaling data to one million hours, model size to 1.5 billion parameters, and introducing a supervised multi-task speech tokenizer plus a differentiable reward模型.
CrossAccent-TTS adds an Accent Intensity Controller to disentangled representations for controllable accent strength in cross-lingual TTS on Indic and L2 datasets.
MagpieTTS-LF enables coherent long-form TTS via three inference-time innovations without any retraining on long-form data.
UniVoice is a conditional flow matching model with a Diffusion Transformer backbone that unifies TTS and SVS via modality-specific encoders and a null melody token for speech, achieving 5.26% speech PER and 16.22% singing PER.
ActorMind is a four-agent chain-of-thought framework that emulates human actors to produce spontaneous, emotion-infused speech responses for role-playing scenarios.
CosyVoice 2 delivers human-parity naturalness and near-lossless streaming speech synthesis by combining finite-scalar quantization, a streamlined pre-trained LLM, and chunk-aware causal flow matching on large multilingual data.
F5-TTS generates natural speech from text via flow matching on DiT with simple text padding, ConvNeXt refinement, and sway sampling, trained on 100K hours multilingual data.
Zero-shot voice cloning augments dysarthric ASR training data to achieve 26.00% WER on TORGO (vs. 24.44% real, 25.12% hybrid) with best cross-corpus results on SAP-1102.
MELD jointly optimizes a discrete latent variable encoder on mel-spectrograms with an autoregressive speech LM, claiming gains over codec and mel baselines on zero-shot TTS/STT plus fewer autoregressive artifacts.
citing papers explorer
-
WavTTS: Towards High-Quality Zero-Shot TTS via Direct Raw Waveform Modeling
WavTTS is the first raw-waveform diffusion TTS model using DiT flow matching and multi-scale mel supervision that approaches SOTA latent zero-shot performance while beating prior end-to-end models.
-
SPG-Codec: Exploring the Role and Boundaries of Semantic Priors in Ultra-Low-Bitrate Neural Speech Coding
Semantic priors from HuBERT and Whisper improve speech codec intelligibility up to 6 kbps but show diminishing returns beyond that, with a bitrate-aware regulation strategy balancing semantic consistency and naturalness.
-
CapTalk: Unified Voice Design for Single-Utterance and Dialogue Speech Generation
CapTalk unifies single-utterance and dialogue voice design via utterance- and speaker-level captions plus a hierarchical variational module for stable timbre with adaptive expression.
-
Hierarchical Decoding for Discrete Speech Synthesis with Multi-Resolution Spoof Detection
MSpoof-TTS improves zero-shot discrete speech synthesis by integrating multi-resolution token-based spoof detection into a hierarchical decoding process that prunes low-quality candidates.
-
Unified Audio Intelligence Without Regressing on Text Intelligence
Audex unifies audio understanding and generation on a strong text MoE backbone with multi-stage SFT plus text-only Cascade RL, matching open SOTA audio scores while mostly retaining text capability.
-
Transcript-Free Flow-Matching Text-to-Speech via Speech Feature Conditioning
RTFree-F5 replaces reference transcripts with mapped self-supervised speech representations in F5-TTS, cutting WER on dysarthric speech from 24.6% to 10.4% without any transcript at inference.
-
PASQA: Pitch-Accent-Focused Speech Quality Assessment Model Trained on Synthetic Speech with Accent Errors
PASQA predicts pitch-accent quality using synthetic accent-error data, mora-conditioned fusion, ranking loss, and auxiliary localization, outperforming conventional MOS models in ordering by error severity and human agreement.
-
TLDR: Compressing Audio Tokens for Efficient Autoregressive Text-to-Speech
TLDR groups codec tokens into patches for patch-level autoregressive modeling in pretrained TTS systems, yielding 1.8x speedup and 75% KV-cache reduction at patch size 4.
-
UniVocal: Unified Speech-Singing Code-Switching Synthesis
UniVocal presents a text-context-only framework for speech-singing code-switching synthesis via two-stage curriculum learning and a synthetic data pipeline, claiming SOTA on a new benchmark.
-
Can We Hear from Events? Generating Speech from Event Camera
EventSpeech is a text-conditioned neural framework that uses neuromorphic event cameras and a new EVT-SPK benchmark to generate expressive speech, claiming to outperform RGB baselines by preserving fine-grained emotions without motion blur.
-
Ultra-Low-Bitrate Mel-Spectrogram-based Neural Speech Coding with Flow-Matching-based Refinement and Vocoding-driven Reconstruction
FMelCodec is a three-stage mel-spectrogram codec using 640x VQ compression, conditional flow matching refinement, and HiFi-GAN reconstruction that reports higher quality than prior methods at 250 bps for 16 kHz speech.
-
Eroding Trust in Real Speech: A Large-Scale Study of Human Audio Deepfake Perception
Large-scale listening study of 35,532 judgments finds human accuracy on real audio fell from 72.7% to 64.1% since 2021 while fake detection remained stable, indicating a skepticism shift toward genuine speech.
-
AST: Adaptive, Seamless, and Training-Free Precise Speech Editing
AST performs training-free text-based speech editing by stitching inverted source latents with synthesized targets and adaptively guiding the flow-matching decoder, achieving state-of-the-art temporal fidelity and speaker preservation.
-
ASPIRin: Action Space Projection for Interactivity-Optimized Reinforcement Learning in Full-Duplex Speech Language Models
ASPIRin decouples speaking timing from token content via binary action space projection and applies GRPO with rule-based rewards to optimize interactivity in SLMs without semantic collapse or repetition.
-
CoMelSinger: Discrete Token-Based Zero-Shot Singing Synthesis With Structured Melody Control and Guidance
CoMelSinger introduces a discrete token-based zero-shot SVS framework on MaskGCT with coarse-to-fine contrastive learning and an SVT module to improve melody control and reduce prosody leakage.
-
CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training
CosyVoice 3 achieves better content consistency, speaker similarity, and prosody naturalness in zero-shot multilingual speech synthesis by scaling data to one million hours, model size to 1.5 billion parameters, and introducing a supervised multi-task speech tokenizer plus a differentiable reward模型.
-
CrossAccent-TTS: Cross-Lingual Accent-Intensity Controllable Text-to-Speech via Disentangled Speaker and Accent Representations
CrossAccent-TTS adds an Accent Intensity Controller to disentangled representations for controllable accent strength in cross-lingual TTS on Indic and L2 datasets.
-
MagpieTTS-LF: Inference-Time Long-Form Speech Generation Without Training on Long-Form data
MagpieTTS-LF enables coherent long-form TTS via three inference-time innovations without any retraining on long-form data.
-
UniVoice: A Unified Model for Speech and Singing Voice Generation
UniVoice is a conditional flow matching model with a Diffusion Transformer backbone that unifies TTS and SVS via modality-specific encoders and a null melody token for speech, achieving 5.26% speech PER and 16.22% singing PER.
-
ActorMind: Emulating Human Actor Reasoning for Speech Role-Playing
ActorMind is a four-agent chain-of-thought framework that emulates human actors to produce spontaneous, emotion-infused speech responses for role-playing scenarios.
-
CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Models
CosyVoice 2 delivers human-parity naturalness and near-lossless streaming speech synthesis by combining finite-scalar quantization, a streamlined pre-trained LLM, and chunk-aware causal flow matching on large multilingual data.
-
F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching
F5-TTS generates natural speech from text via flow matching on DiT with simple text padding, ConvNeXt refinement, and sway sampling, trained on 100K hours multilingual data.
-
Low-Burden Data Augmentation for Dysarthric ASR via Zero-Shot Voice Cloning
Zero-shot voice cloning augments dysarthric ASR training data to achieve 26.00% WER on TORGO (vs. 24.44% real, 25.12% hybrid) with best cross-corpus results on SAP-1102.
-
MELD: Mel-Spectrogram-Based Speech Language Modeling with Discrete Latent Variables
MELD jointly optimizes a discrete latent variable encoder on mel-spectrograms with an autoregressive speech LM, claiming gains over codec and mel baselines on zero-shot TTS/STT plus fewer autoregressive artifacts.
-
From Objectives to Applications: Aligning Architectural Biases in Audio Self-Supervised Learning
A survey that organizes audio SSL into five objective paradigms, relates their demands to architectural biases, and interprets downstream applications as tests of generalization.