Topics / Sound / Oct 26 - Nov 1, 2026

Weekly issue

Top 20 Sound Papers: Oct 26 - Nov 1, 2026

Updated · Category: Sound (arXiv cs.SD) · Ranked by potential impact

Prefer new papers every day? It's free.Sound Daily Papers →
  1. 1
    EmbeddingGemma 2: Unified and Efficient Multimodal Representations

    We present EmbeddingGemma 2, an open multimodal embedding model which natively represents text, images, audio, video, and combinations thereof in a unified space.

  2. 2
    In-Context Adapter Generation for Compositional Generalization in Large Audio Language Models

    To exploit fine-grained task information, we introduce in-context adapter generation (ICAG), a token-level hypernetwork architecture; we call the resulting method T2L-ICAG.

  3. 3
    Spatial Audio Generation and Editing with Executable Plans and Self-Distilling Transfusion

    We present AMBIT, a unified framework for native First-Order Ambisonics (FOA) generation and editing with three connected contributions. First, a shared Transfusion backbone couples autoregressive planning with rectified-flow rendering through executable ScenePlans that specify source content, activity, and 3D trajectories.

  4. 4
    Discovering Interpretable Acoustic Features in Autoregressive Voice-Cloning TTS Models

    We introduce a feature-based interpretability framework for autoregressive voice-cloning TTS models, analyzing features from sparse autoencoders (SAEs), PCA, and random directions via sparse probing, steering, and quantitative auto-interpretability.

  5. 5
    S3-S2ST: Semantically Supervised Streaming Speech-to-Speech Translation

    We present S3-S2ST, a model that learns to translate and speak from semantically aligned streaming supervision.

Get papers like these the day they come out, not a week later.

Follow this category
  1. 6
    AV-QTSS: Class-Query Target Sound Spatialization in Multi-Source 360$^\circ$ Videos

    We propose a query-conditioned dense binding localizer that predicts a spherical probability map over the whole scene, coupled to the renderer so both are optimized jointly.

  2. 7
    What Makes A Good Audio Prompt for Zero-Shot TTS? Learning Prompt Selection from Synthesis Outcomes

    We propose a small network that learns from the outputs a TTS model produced to rank a recording's candidate prompts from their audio alone.

  3. 8
    A Tutorial on Generative Models for Music Audio

    It introduces neural audio codecs and autoencoders, autoregressive and masked token models, diffusion and flow matching, sampling acceleration, and preference optimization in a shared notation.

  4. 9
    Understanding Everyday Communication Patterns, Stress, and Productivity in Hybrid Work Through Egocentric Wearable Audio Sensing

    To improve diarization performance under low-dimensional feature constraints, we apply a teacher-student knowledge distillation framework that transfers knowledge from large-scale speech foundation models into a compact ResNet-based model, achieving consistent improvements in diarization performance over the baseline across three public egocentric speech datasets.

  5. 10
    Beyond Utterance-Level Speaker Similarity: Perception-Aligned Segmental Evaluation of Accent in Speech Generation

    We introduce a segmental evaluation protocol combining speaker and accent similarity ratings with subword annotations of perceived accent differences provided by listeners from target accent regions.

  6. 11
    SonoWorld4D: From One Video to A 4D Audio-Visual Scene Via Diffusion Synchronization

    However, sound, despite being intrinsically coupled with dynamic objects and events, remains largely unmodeled in prior work. We introduce Video-to-4DAVScene, a new task that turns a video into an explorable, time-varying 4D audio-visual scene, and present the SonoWorld4D framework for this task.

  7. 12
    Perceptual Differences and Preferences in Speech Enhancement

    We investigate discrimination and preference as complementary dimensions of SE evaluation through crowdsourced triangle and pairwise preference tests across four SE systems, with comparisons to clean references.

  8. 13
  9. 14
    DRAMA-Eval: A Benchmark for Context-Aware, Multi-Turn, Multi-Speaker Speech Generation

    DRAMA-Eval utilizes structured script contexts and offers a holistic evaluation framework covering objective, local, and global aspects of generation quality, enabling the assessment of acoustic fidelity, sentence-level execution, and long-range coherence.

  10. 15
    DIAL-S2ST: Learning Latency Control for Simultaneous Speech-to-Speech Translation

    We propose DIAL-S2ST, a post-training method for learning inference-time latency control in simultaneous speech-to-speech translation models.

  11. 16
    TAC2: Timestamped Audio Captioning As Composable Functions

    We introduce TAC2, a foundation model that casts timestamped audio captioning as composable functions to construct granular, speaker-attributed acoustic timelines directly from in-the-wild audio.

  12. 17
    EmpDPO: Enabling Emphasis Control in TTS Via Direct Preference Optimization

    In this work, we study how to add emphasis control post hoc to a pretrained TTS model that has never been trained with emphasis annotations.

  13. 18
    SelectSep: Target-Selective Video Queries for Audio Source Separation

    We propose SelectSep, a target-selective VQSS framework that jointly uses text and a single click coordinate to specify the target source and instance without requiring segmentation at inference time.

  14. 19
    Gangqinpu: A Multi-Genre, Multi-Task Piano Dataset with Aligned Audio, MIDI, and Musical Scores

    We introduce Gangqinpu, a multi-genre piano dataset containing 18,786 score-performance pairs and 1,007.1 hours of source audio, with automatic note-level alignments and derived beat/downbeat annotations.

  15. 20
    From Speech Tokens to Stolen Voices: Attacker-Chosen-Text Speaker Impersonation Via Representation Inversion

    If an attacker obtains them, it may use this information to imitate the user's voice, threatening their voice privacy. To investigate this risk, we propose \stia, a Speech Token Inversion Attack that generates speech with attacker-specified text in a victim's voice using a single short leaked representation, without access to the original audio or its transcript.