Topics / Sound / Oct 12 - 18, 2026
Weekly issue
Top 20 Sound Papers: Oct 12 - 18, 2026
- 1What Do Deepfake Audio Detectors Hear? From Self-Supervised Models to Audio LLMs
When processing these mixtures, it remains unclear whether detectors attend to the speech or the background event, and where their spectral evidence lies. We study this on the CompSpoof V2 database, whose mixtures pair independently real-or-fake speech and event components.
- 2Full-Duplex Interaction with Speech, Expressions and Gestures
We present ELLSA 2, an end-to-end full-duplex dialogue model that jointly optimizes speech interaction and streaming gesture-expression generation.
- 3Spend Where It Matters: Particle Search with MH Refinement for Budget-Constrained Audio-Visual Alignment
We introduce \textbf{S}pend \textbf{W}here \textbf{I}t \textbf{M}atters (\modelname), a two-stage, training-free, reward-agnostic framework that combines particle search with MH refinement under a fixed compute budget.
- 4Soloist: A Generative Model for Blind Voice Separation in The Wild
We introduce Soloist, a generative model for blind voice separation that autoregressively generates individual voice streams as audio-codec sequences conditioned on the mixture, without requiring a target-speaker query or a predefined voice count.
- 5Multimodal Contrastive Representation Learning Based Voice Conversion for Electrolaryngeal Speech Enhancement
These mismatches also make sequence-to-sequence voice conversion struggle to learn robust intermediate representations, hindering EL-speech-to-normal-speech conversion (EL2SP) performance. To address this, we propose a novel multimodal contrastive representation learning framework to 1) achieve richer multimodal-representation learning, and 2) facilitate the transferability of multimodal representations to speech-only EL2SP.
Get papers like these the day they come out, not a week later.
Follow this category- 6CAST-ASD: Causal Attention Over Speakers and Time for Online Active Speaker Detection
We present CAST-ASD (Causal Attention over Speakers and Time), a causal joint model for active speaker detection (ASD).
- 7MedSonor: Learning to Use Clinical Audio in Medical LLMs
We present MedSonor, a medical dialogue model integrating medical knowledge, spoken dialogue, and clinical audio understanding.
- 8Beyond Fixed Targets: On-Policy One-Shot Speaker Personalization for Autoregressive Text-To-Speech Models
Diversifying the training data with zero-shot pseudo-pairs eliminates this failure but yields little improvement in speaker similarity and increases word error rate because the targets remain fixed and teacher-forced. We therefore propose two on-policy approaches for one-shot speaker personalization: Group Relative Policy Optimization (GRPO) with a speaker-similarity reward and On-Policy Self-Distillation (OPSD) with soft supervision from a reference-conditioned teacher.
- 9SSEmb: Training A Generalizable Speech Retrieval Model Using Only Automatic Speech Recognition Data
Training these retrievers typically depends on large collections of spoken query-document pairs that are costly to acquire and difficult to extend across domains and languages. We introduce \ours, which replaces such task-specific supervision with speech-transcript pairs from standard ASR corpora.
- 10AV-FlowEdit: Training-Free Prompt-Driven Audio-Visual Editing with Event-Aware Spectral Trajectory Preservation
We present AV-FlowEdit, a training-free and inversion-free framework for prompt-driven audio-video editing.
- 11What Does LALM Post-training Learn? Rethinking Audio Understanding and Reasoning Capabilities
Across different models and post-training methods, we find that much of the reported improvement persists even without informative audio, but largely disappears when the question format changes.
- 12AnyEdit: A Unified Framework for Speech and Singing Voice Editing with Real-World Environmental Consistency
We present AnyEdit, a unified, separation-free framework for speech and singing voice editing that preserves real-world acoustic environments, including noise, reverberation, and background music.
- 13What Matters in Preference Post-Training of Continuous Autoregressive Speech Language Models
Continuous autoregressive speech language models combine an autoregressive (AR) backbone with a diffusion or flow-matching head.
- 14Long-Prompt Zero-Shot TTS: Benchmarking Prosody Reproduction Beyond Voice Cloning
We introduce LP-TTS-Bench, a bilingual benchmark of 37 speakers with 63.95 prompt hours and 2,728 held-out test utterances.
- 15E3-System: Expressive EmoData, Explicitly Controllable, Easy-to-Use TTS
Text descriptions offer a natural control interface but provide weak acoustic grounding, whereas speech style prompts provide concrete prosody but entangle style with the prompt speaker's timbre. To bridge this gap, we propose E3-System, a unified intent-to-speech framework for expressive zero-shot TTS that couples three requirements: Expressive style modeling, Explicit timbre-style control, and Easy-to-use natural-language interaction.
- 16MSVAE: Rethinking Musical Semantics in Continuous Latent Spaces
We address this difficulty from the representation side. We propose MSVAE (Music Semantic Variational Autoencoder), which structures its continuous latent around two complementary semantic streams: work-level compositional content and lyric-level linguistic-vocal content.
- 17SyncReward: Human-Aligned Synchronization Reward for Audio-Visual Generation
A taxonomy-guided construction pipeline covers four sound-event families: Speech, Onset, Instrument, and Ambient.
- 18SIMPLE: Semantic Structure-Aligned Prompt Learning for Open-Vocabulary Audio-Visual Event Localization
In this paper, we propose SIMPLE, a semantic structure-aligned prompt learning framework that aligns cross-category semantic topology between learnable and hand-crafted prompts across asynchronous audio and visual streams.
- 19PAG-Agent: Zero-Shot Personalized Music Album Generation Via Aesthetic Profile Distillation
We present PAG-Agent, a zero-shot multimodal framework that uses an explicit Personal Aesthetic Profile to bridge user preference inference and music generation control.
- 20Fish Audio Drama 3 Technical Report
We introduce Fish Audio Drama 3, an open-source speech generation model that turns text into directed performances, built on a 35B-parameter mixture-of-experts backbone that activates 3B parameters per token.