CodecAttack perturbs audio in codec latent space with multi-bitrate EoT to achieve 85.5% average ASR on Opus-compressed Audio LLMs versus under 26% for waveform baselines, with transfer to MP3 and AAC.
hub Canonical reference
Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers
Canonical reference. 89% of citing Pith papers cite this work as background.
55
Pith papers citing it
Background
89% of classified citations
abstract
We introduce a language modeling approach for text to speech synthesis (TTS). Specifically, we train a neural codec language model (called Vall-E) using discrete codes derived from an off-the-shelf neural audio codec model, and regard TTS as a conditional language modeling task rather than continuous signal regression as in previous work. During the pre-training stage, we scale up the TTS training data to 60K hours of English speech which is hundreds of times larger than existing systems. Vall-E emerges in-context learning capabilities and can be used to synthesize high-quality personalized speech with only a 3-second enrolled recording of an unseen speaker as an acoustic prompt. Experiment results show that Vall-E significantly outperforms the state-of-the-art zero-shot TTS system in terms of speech naturalness and speaker similarity. In addition, we find Vall-E could preserve the speaker's emotion and acoustic environment of the acoustic prompt in synthesis. See https://aka.ms/valle for demos of our work.
hub tools
citation-role summary
background 9
citation-polarity summary
roles
background 9representative citing papers
AffectCodec is an emotion-guided neural speech codec that preserves emotional cues during quantization while maintaining semantic fidelity and prosodic naturalness.
GibbsTTS combines a training-free kinetic-optimal scheduler with finite-step moment correction in MI-DFM to deliver top naturalness and strong speaker similarity in zero-shot TTS.
Large-model adaptation with Tibetan text handling produces natural speech from limited data, outperforming commercial systems.
MelShield adds keyed low-energy spread-spectrum perturbations to Mel-spectrograms inside TTS pipelines before vocoding to enable robust extraction of user-specific attribution signals even after compression or noise.
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.
V.O.I.C.E is a new taxonomy that organizes synthetic voice risks into five categories and shows how they interact with exposure, visibility, and legal context using empirical incident data.
PhySE combines VLM pre-training for fast social context profiling with a dynamic psychological agent to overcome delays and static tactics in AR-LLM social engineering attacks, tested in a 60-person user study.
Introduces the Indic-CodecFake dataset for Indic codec deepfakes and SATYAM, a novel hyperbolic ALM that outperforms baselines through dual-stage semantic-prosodic fusion using Bhattacharya distance.
X-VC achieves zero-shot streaming voice conversion via one-step codec-space conversion with dual-conditioning acoustic converter and role-assignment training on generated paired data.
Moshi is the first real-time full-duplex spoken large language model that casts dialogue as speech-to-speech generation using parallel audio streams and an inner monologue of time-aligned text tokens.
DASB is a new benchmark for discrete audio tokens showing semantic tokens outperform acoustic ones but discrete representations remain less robust than continuous features across domains.
Kosmos-1 shows strong zero-shot and few-shot results on language tasks, image captioning, visual QA, OCR-free document understanding, and image recognition guided by text instructions.
CleanCodec reframes audio tokenization as a selective information bottleneck to encode only perceptually important features at 12.5 tokens per second, outperforming prior codecs in efficiency, speaker similarity, and intelligibility.
The paper introduces target-KL regularization to train audio VAEs at specific bitrates, enabling rate-distortion curves and comparison to discrete audio codecs for improved text-to-sound generation.
Break-the-Beat! renders drum MIDI audio that matches the timbre of a reference clip by fine-tuning a text-to-audio model with a content encoder and hybrid conditioning on a new paired dataset.
LATTE creates a compact latent token bottleneck in audio tokenizers that aggregates global information and enables unsupervised editing of attributes like speaker identity via token swapping.
CASCADE formalizes semantic interchangeability and convergence in target model representations to enable context-aware acceptance relaxation in tree-based speculative decoding, delivering up to 3.6x speedup on text-to-image models without quality loss.
MiniMind-O delivers a working 0.1B-scale open omni model with speech-native output, Thinker-Talker split, frozen encoders, and full release of code, checkpoints, and training data.
Chain-of-Details (CoD) is a cascaded TTS method that explicitly models temporal coarse-to-fine dynamics with a shared decoder, achieving competitive performance using significantly fewer parameters.
HCFD is a new pathology-aware benchmark and dataset for codec-fake audio detection in healthcare, with PHOENIX-Mamba achieving up to 97% accuracy by modeling fakes as modes in hyperbolic space.
StreamMark trains an Encoder-Distortion-Decoder network to embed semi-fragile watermarks that remain recoverable after benign audio transformations but drop to random accuracy under voice conversion and editing attacks.
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.
Borderless Long Speech Synthesis unifies voice design, multi-speaker TTS, and long-form generation via Global-Sentence-Token annotations, CoT reasoning, and a Structured Semantic Interface for agent-centric control.
citing papers explorer
-
Codec-Robust Attacks on Audio LLMs
CodecAttack perturbs audio in codec latent space with multi-bitrate EoT to achieve 85.5% average ASR on Opus-compressed Audio LLMs versus under 26% for waveform baselines, with transfer to MP3 and AAC.
-
AffectCodec: Emotion-Preserving Neural Speech Codec for Expressive Speech Modeling
AffectCodec is an emotion-guided neural speech codec that preserves emotional cues during quantization while maintaining semantic fidelity and prosodic naturalness.
-
Kinetic-Optimal Scheduling with Moment Correction for Metric-Induced Discrete Flow Matching in Zero-Shot Text-to-Speech
GibbsTTS combines a training-free kinetic-optimal scheduler with finite-step moment correction in MI-DFM to deliver top naturalness and strong speaker similarity in zero-shot TTS.
-
Tibetan-TTS:Low-Resource Tibetan Speech Synthesis with Large Model Adaptation
Large-model adaptation with Tibetan text handling produces natural speech from limited data, outperforming commercial systems.
-
MelShield: Robust Mel-Domain Audio Watermarking for Provenance Attribution of AI Generated Synthesized Speech
MelShield adds keyed low-energy spread-spectrum perturbations to Mel-spectrograms inside TTS pipelines before vocoding to enable robust extraction of user-specific attribution signals even after compression or noise.
-
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.
-
V.O.I.C.E (Voice, Ownership, Identity, Control, Expression): Risk Taxonomy of Synthetic Voice Generation From Empirical Data
V.O.I.C.E is a new taxonomy that organizes synthetic voice risks into five categories and shows how they interact with exposure, visibility, and legal context using empirical incident data.
-
PhySE: A Psychological Framework for Real-Time AR-LLM Social Engineering Attacks
PhySE combines VLM pre-training for fast social context profiling with a dynamic psychological agent to overcome delays and static tactics in AR-LLM social engineering attacks, tested in a 60-person user study.
-
Indic-CodecFake meets SATYAM: Towards Detecting Neural Audio Codec Synthesized Speech Deepfakes in Indic Languages
Introduces the Indic-CodecFake dataset for Indic codec deepfakes and SATYAM, a novel hyperbolic ALM that outperforms baselines through dual-stage semantic-prosodic fusion using Bhattacharya distance.
-
X-VC: Zero-shot Streaming Voice Conversion in Codec Space
X-VC achieves zero-shot streaming voice conversion via one-step codec-space conversion with dual-conditioning acoustic converter and role-assignment training on generated paired data.
-
Moshi: a speech-text foundation model for real-time dialogue
Moshi is the first real-time full-duplex spoken large language model that casts dialogue as speech-to-speech generation using parallel audio streams and an inner monologue of time-aligned text tokens.
-
DASB - Discrete Audio and Speech Benchmark
DASB is a new benchmark for discrete audio tokens showing semantic tokens outperform acoustic ones but discrete representations remain less robust than continuous features across domains.
-
Language Is Not All You Need: Aligning Perception with Language Models
Kosmos-1 shows strong zero-shot and few-shot results on language tasks, image captioning, visual QA, OCR-free document understanding, and image recognition guided by text instructions.
-
CleanCodec: Efficient and Robust Speech Tokenization via Perceptually Guided Encoding
CleanCodec reframes audio tokenization as a selective information bottleneck to encode only perceptually important features at 12.5 tokens per second, outperforming prior codecs in efficiency, speaker similarity, and intelligibility.
-
Taming Audio VAEs via Target-KL Regularization
The paper introduces target-KL regularization to train audio VAEs at specific bitrates, enabling rate-distortion curves and comparison to discrete audio codecs for improved text-to-sound generation.
-
Break-the-Beat! Controllable MIDI-to-Drum Audio Synthesis
Break-the-Beat! renders drum MIDI audio that matches the timbre of a reference clip by fine-tuning a text-to-audio model with a content encoder and hybrid conditioning on a new paired dataset.
-
Exploring Token-Space Manipulation in Latent Audio Tokenizers
LATTE creates a compact latent token bottleneck in audio tokenizers that aggregates global information and enables unsupervised editing of attributes like speaker identity via token swapping.
-
CASCADE: Context-Aware Relaxation for Speculative Image Decoding
CASCADE formalizes semantic interchangeability and convergence in target model representations to enable context-aware acceptance relaxation in tree-based speculative decoding, delivering up to 3.6x speedup on text-to-image models without quality loss.
-
MiniMind-O Technical Report: An Open Small-Scale Speech-Native Omni Model
MiniMind-O delivers a working 0.1B-scale open omni model with speech-native output, Thinker-Talker split, frozen encoders, and full release of code, checkpoints, and training data.
-
Text-To-Speech with Chain-of-Details: modeling temporal dynamics in speech generation
Chain-of-Details (CoD) is a cascaded TTS method that explicitly models temporal coarse-to-fine dynamics with a shared decoder, achieving competitive performance using significantly fewer parameters.
-
HCFD: A Benchmark for Audio Deepfake Detection in Healthcare
HCFD is a new pathology-aware benchmark and dataset for codec-fake audio detection in healthcare, with PHOENIX-Mamba achieving up to 97% accuracy by modeling fakes as modes in hyperbolic space.
-
StreamMark: A Deep Learning-Based Semi-Fragile Audio Watermarking for Proactive Deepfake Detection
StreamMark trains an Encoder-Distortion-Decoder network to embed semi-fragile watermarks that remain recoverable after benign audio transformations but drop to random accuracy under voice conversion and editing attacks.
-
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.
-
Borderless Long Speech Synthesis
Borderless Long Speech Synthesis unifies voice design, multi-speaker TTS, and long-form generation via Global-Sentence-Token annotations, CoT reasoning, and a Structured Semantic Interface for agent-centric control.
-
Qwen3-TTS Technical Report
Qwen3-TTS delivers state-of-the-art multilingual TTS performance with 3-second voice cloning, description control, and ultra-low-latency streaming via dual tokenizers and a dual-track LM architecture trained on over 5 million hours of data.
-
Aliasing-Free Neural Audio Synthesis
Pupu-Vocoder and Pupu-Codec integrate differentiable anti-aliasing into neural audio models to eliminate aliasing artifacts from non-linear activations and upsampling, yielding better results on music and singing voice.
-
Two-Dimensional Quantization for Geometry-Aware Audio Coding
Q2D2 uses 2D geometric grid projections to quantize feature pairs in neural audio codecs, yielding implicit codebooks that improve efficiency and utilization over RVQ, VQ, and FSQ while maintaining reconstruction quality.
-
Speculative Coupled Decoding for Training-Free Lossless Acceleration of Autoregressive Visual Generation
Speculative Coupled Decoding stabilizes draft sampling in Speculative Jacobi Decoding via an information-theoretic coupling step, delivering up to 4.2x image and 13.6x video speedups with no quality loss or training.
-
Step-Audio 2 Technical Report
Step-Audio 2 integrates a latent audio encoder, reasoning-centric reinforcement learning, and discrete audio token generation into language modeling to deliver state-of-the-art performance on audio understanding and conversational benchmarks.
-
ZipVoice-Dialog: Non-Autoregressive Spoken Dialogue Generation with Flow Matching
ZipVoice-Dialog is a flow-matching non-autoregressive model for zero-shot spoken dialogue generation that uses curriculum learning and speaker-turn embeddings, paired with a new 6.8k-hour OpenDialog dataset, and reports better speed and quality than autoregressive baselines.
-
JAM-Flow: Joint Audio-Motion Synthesis with Flow Matching
JAM-Flow introduces a unified flow-matching model with a Multi-Modal Diffusion Transformer that jointly synthesizes facial motion and speech from text, audio, or motion inputs.
-
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.
-
Perceptual implications of automatic anonymization in pathological speech
Listeners detect automatic anonymization in pathological speech at 91-93% accuracy with a 30-point perceived quality drop, yet clinical severity ratings stay nearly unchanged for dysarthria, dysglossia, and dysphonia.
-
AudioPaLM: A Large Language Model That Can Speak and Listen
AudioPaLM unifies PaLM-2 and AudioLM to outperform prior systems on speech translation while enabling zero-shot speech-to-text for many unseen language pairs and voice transfer from short prompts.
-
UNISON: A Unified Sound Generation and Editing Framework via Deep LLM Fusion
UNISON introduces a unified latent diffusion framework with layer-wise LLM fusion and channel-mask task encoding for multiple speech and sound generation and editing tasks.
-
Raon-OpenTTS: Open Models and Data for Robust Text-to-Speech
Raon-OpenTTS provides an open 510K-hour curated speech dataset and DiT-based TTS models up to 1B parameters that achieve competitive WER and speaker similarity on benchmarks versus closed models trained on millions of hours.
-
AgentSteerTTS: A Multi-Agent Closed-Loop Framework for Composite-Instruction Text-to-Speech
AgentSteerTTS proposes a multi-agent framework with adversarial disentanglement, dual-stream anchoring via acoustic prototypes, and fast-slow feedback to achieve intent-faithful expressive TTS for composite instructions.
-
Controllable Singing Style Conversion with Boundary-Aware Information Bottleneck
A singing voice conversion system with boundary-aware information bottleneck and high-frequency augmentation achieves the best naturalness in SVCC2025 subjective tests while using less extra data than competitors.
-
Voxtral TTS
Voxtral TTS produces expressive multilingual speech from 3-second reference audio with a hybrid autoregressive-plus-flow-matching architecture and a new VQ-FSQ tokenizer, achieving 68.4% win rate over ElevenLabs in human evaluations.
-
AUHead: Realistic Emotional Talking Head Generation via Action Units Control
AUHead uses audio-language models to generate Action Unit sequences from speech and feeds them into a controllable diffusion model to synthesize realistic emotional talking-head videos.
-
Enhancing Speech Large Language Models through Reinforced Behavior Alignment
Reinforced Behavior Alignment (RBA) uses self-synthesized data from a teacher LLM and reinforcement learning to close the instruction-following gap in SpeechLMs, outperforming distillation and reaching SOTA on spoken QA and speech-to-text translation benchmarks.
-
Kimi-Audio Technical Report
Kimi-Audio is an open-source audio foundation model that achieves state-of-the-art results on speech recognition, audio understanding, question answering, and conversation after pre-training on more than 13 million hours of speech, sound, and music data.
-
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.
-
EntangleCodec: A Unified Discrete Audio Tokenizer via Semantic-Acoustic Entanglement
EntangleCodec unifies semantic and acoustic audio tokenization via caption alignment and flow-matching decoding, reporting competitive reconstruction, +7.4% gains on MMAR understanding, and 0.6B-parameter ALMs surpassing 13B-parameter continuous baselines.
-
ATRIE: Adaptive Tuning for Robust Inference and Emotion in Persona-Driven Speech Synthesis
ATRIE disentangles timbre and prosody in a Persona-Prosody Dual-Track model distilled from a large LLM to achieve strong identity preservation (EER 0.04) and emotional speech synthesis with SOTA results on an extended AnimeTTS-Bench.
-
The Rise and Potential of Large Language Model Based Agents: A Survey
The paper surveys the origins, frameworks, applications, and open challenges of AI agents built on large language models.
-
From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI
The paper analyzes evolving security and safety threats in generative AI from content generation to agentic actions, noting that attack surfaces expand faster than defenses and that many safeguards require institutional coordination not yet in place.
-
Expressive Prompting: Improving Emotion Intensity and Speaker Consistency in Zero-Shot TTS
A two-stage static-then-dynamic prompt selection strategy using prosodic features, LLM coherence scores, and similarity metrics improves emotion intensity and speaker consistency in zero-shot TTS.
-
Multilingual and Multimodal LLMs in the Wild: Building for Low-Resource Languages
A tutorial synthesizing foundations, recent models such as PALO and Maya, and low-cost methods for tri-modal multilingual AI in resource-constrained settings.