Fish Audio releases Live 3 full-duplex voice model
The listing cites a 186 ms reaction time, but the fetched page is only a cookie notice.
Stories per day · 30 days
5 RSS
Relatedhuggingface3granite2ibm2multimodal2artificial-analysis1benchmark1black-box1evasion1fish-audio1flow-matching1full-duplex1granite-switch1grok1lfm31liquid-ai1live-31on-device1on-device-ai1
42 stories, newest first · one card per story
Stories tagged #speech
Phonon TTS launches high-quality on-device speech model
Phonon TTS is a new on-device speech synthesis model that tops benchmarks for quality and efficiency.
ScrambleMark evades speech watermarking via harmonic scrambling
ScrambleMark achieves 97% evasion on commercial platforms while preserving speech naturalness at 4.2/5.
IBM releases Granite Speech 5.0 2B for transcription and translation
The captured page is a cookie notice, so size, languages, license, and availability are not confirmed beyond the title.
Liquid AI releases on-device omni-modal model LFM3-Omni-2B
The 2.17B checkpoint accepts text, images, video and 16 kHz audio, and can reply with text, 24 kHz speech, or both.
StyleStream 2.0 cuts streaming voice conversion latency to 520 ms
A text-conditioned style embedding supports prompt-based voice design and instruction-based editing of one attribute.
Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint
IBM published Granite Switch 4.1, offering chat, decisions, and speech through a single endpoint.
Meet Grok Voice 3.0: SpaceXAI Voice Agent Model Ranks #1 on Artificial Analysis
SpaceXAI’s Grok Voice 3.0 voice-agent model is reported to rank first on Artificial Analysis.
9
OpenPA: Forced-alignment-free Multi-granular Pronunciation Assessment with Mixed-transcript Adaptation
OpenPA uses Whisper cross-attention to score pronunciation at word and utterance level without forced alignment.
Revisiting Accent Classification with DistilHuBERT: Speaker-Disjoint Evaluation and Cross-Corpus Generalization
DistilHuBERT accent classifiers look strong in-corpus but fail speaker-disjoint and cross-corpus tests.
Speech event detector: Timestamped detection of non-verbal vocalizations and background sounds for rich speech transcription
A 5.8M model timestamps 54 non-speech sounds and merges them into Whisper transcripts.
SepGen: Multi-Stem Audio-Video Separation and Generation in a Single Model
SepGen jointly generates video, a mixed soundtrack, and per-source audio stems, and can also separate mixes from captions.
Itgan at NADI 2026 shared task: Parameter-Efficient Whisper Adaptation for Robust, Mixed-Dialect and Code-Switched Arabic ASR
LoRA-adapted Whisper systems for NADI 2026 Arabic ASR reach 57.1%, 46.7%, and 14.49% WER on three subtasks.
Exploiting Acoustic and Content-Oriented Speaker Verification Attacks Against Multilingual Voice Anonymization
Acoustic ASV attackers generally beat content-oriented ones against multilingual voice anonymization.
Language Carries the Expert's Impression: Instrument-Anchored LLM Judges Transfer Counseling-Quality Assessment and Beat In-Domain Training
Instrument-anchored LLM judges transfer counseling-quality ratings across domains and beat in-domain training.
HiPLEX: Hierarchical Policy Factorization for Full Duplex Speech Language Models
HiPLEX factorizes full-duplex speech policies so reinforcement learning can improve timing and content separately.
Microsoft AI releases new transcription and text-to-speech models for voice agents
Microsoft released MAI-Transcribe-2-Streaming and two MAI-Voice text-to-speech models for low-latency voice agents.
UniAE-MoE: A Unified Audio Encoder via Mixture of Experts
UniAE-MoE fuses Qwen2-Audio and Audio-Flamingo 3 encoders, scoring 0.802 on XARES-LLM.
Training-Free Affinity Fusion of Neural and Embedding-Based Speaker Diarization
Training-free affinity fusion combines neural and embedding diarization, reducing error on AMI and CALLHOME.
Who Said What, and Will It Be Remembered? Evaluating Persistent Speaker Attribution Across Meetings
SI-cpWER measures whether the same speaker keeps one identity across meetings, and ThyVoice leads commercial cascades.
Alibaba Qwen Releases Qwen-Audio-3.1-Realtime: A Full-Duplex Voice Model Trained to Think, Act, and Decide When to Speak
Alibaba released API-only Qwen-Audio-3.1-Realtime, a full-duplex voice model for tool-using agents.
Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning
Tacit-TTS distills IndexTTS2 into transcript-free masked voice cloning that is over 10x faster.
DeMark: A Query-Free Black-Box Attack for Quality-Preserving Audio Watermark Removal
DeMark removes learning-based audio watermarks query-free, reaching 0.92–0.96 success while preserving speech quality.
Tokens Change, Structure Endures: Spectral Watermarking for Generated Speech
Redwing keeps generated-speech watermarks intact through retokenization using a spectral substitution basis.