惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

雷峰网
雷峰网
G
Google Developers Blog
D
Docker
The GitHub Blog
The GitHub Blog
H
Help Net Security
WordPress大学
WordPress大学
博客园_首页
Recent Announcements
Recent Announcements
P
Proofpoint News Feed
罗磊的独立博客
I
InfoQ
U
Unit 42
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
V
Visual Studio Blog
Jina AI
Jina AI
J
Java Code Geeks
博客园 - Franky
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
大猫的无限游戏
大猫的无限游戏
小众软件
小众软件
Last Week in AI
Last Week in AI
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
D
DataBreaches.Net
Google DeepMind News
Google DeepMind News

eess.AS updates on arXiv.org

Dependence on Early and Late Reverberation of Single-Channel Speaker Distance Estimation MIST: Multimodal Interactive Speech-based Tool-calling Conversational Assistants for Smart Homes LiVeAction: a Lightweight, Versatile, and Asymmetric Neural Codec Design for Real-time Operation Weight-Decay Turns Transformer Loss Landscapes Villani: Functional-Analytic Foundations for Optimization and Generalization PairAlign: A Framework for Sequence Tokenization via Self-Alignment with Applications to Audio Tokenization WavCube: Unifying Speech Representation for Understanding and Generation via Semantic-Acoustic Joint Modeling Predictive-Generative Drift Decomposition for Speech Enhancement and Separation Minimizing Modality Gap from the Input Side: Your Speech LLM Can Be a Prosody-Aware Text LLM X-Voice: Enabling Everyone to Speak 30 Languages via Zero-Shot Cross-Lingual Voice Cloning JASTIN: Aligning LLMs for Zero-Shot Audio and Speech Evaluation via Natural Language Instructions Phoneme-Level Deepfake Detection Across Emotional Conditions Using Self-Supervised Embeddings When Audio-Language Models Fail to Leverage Multimodal Context for Dysarthric Speech Recognition Dimensionality-Aware Anomaly Detection in Learned Representations of Self-Supervised Speech Models Mitigating Multimodal LLMs Hallucinations via Relevance Propagation at Inference Time Virtual Speech Therapist: A Clinician-in-the-Loop AI Speech Therapy Agent for Personalized and Supervised Therapy LASE: Language-Adversarial Speaker Encoding for Indic Cross-Script Identity Preservation Towards Improving Speaker Distance Estimation through Generative Impulse Response Augmentation Beyond Decodability: Reconstructing Language Model Representations with an Encoding Probe MMAudioReverbs: Video-Guided Acoustic Modeling for Dereverberation and Room Impulse Response Estimation Alethia: A Foundational Encoder for Voice Deepfakes From Birdsong to Rumbles: Classifying Elephant Calls with Out-of-Species Embeddings Beyond the Baseband: Adaptive Multi-Band Encoding for Full-Spectrum Bioacoustics Classification Predicting Upcoming Stuttering Events from Three-Second Audio: Stratified Evaluation Reveals Severity-Selective Precursors, and the Model Deploys Fully On-Device The False Resonance: A Critical Examination of Emotion Embedding Similarity for Speech Generation Evaluation DiffAnon: Diffusion-based Prosody Control for Voice Anonymization Recurrence-Based Nonlinear Vocal Dynamics as Digital Biomarkers for Depression Detection from Conversational Speech One Voice, Many Tongues: Cross-Lingual Voice Cloning for Scientific Speech Similarity Choice and Negative Scaling in Supervised Contrastive Learning for Deepfake Audio Detection Walking Through Uncertainty: An Empirical Study of Uncertainty Estimation for Audio-Aware Large Language Models Praxy Voice: Voice-Prompt Recovery + BUPS for Commercial-Class Indic TTS from a Frozen Non-Indic Base at Zero Commercial-Training-Data Cost
T-Mimi: A Transformer-based Mimi Decoder for Real-Time On...
Haibin Wu, Bach Viet Do, Naveen Suda, Julian Chan, Madhavan C R, · 2026-01-28 · via eess.AS updates on arXiv.org

Neural audio codecs provide promising acoustic features for speech synthesis, with representative streaming codecs like Mimi providing high-quality acoustic features for real-time Text-to-Speech (TTS) applications. However, Mimi's decoder, which employs a hybrid transformer and convolution architecture, introduces significant latency bottlenecks on edge devices due to the the compute intensive nature of deconvolution layers which are not friendly for mobile-CPUs, such as the most representative framework XNNPACK. This paper introduces T-Mimi, a novel modification of the Mimi codec decoder that replaces its convolutional components with a purely transformer-based decoder, inspired by the TS3-Codec architecture. This change dramatically reduces on-device TTS latency from 42.1ms to just 4.4ms. Furthermore, we conduct quantization aware training and derive a crucial finding: the final two transformer layers and the concluding linear layers of the decoder, which are close to the waveform, are highly sensitive to quantization and must be preserved at full precision to maintain audio quality.