Publicly available phonocardiogram (PCG) datasets remain limited in size and pathological diversity, constraining both auscultation training and the generalisation of automated heart-sound classifiers. A class-conditional diffusion model for PCG generation is developed in the log-mel domain and synthetic fidelity is assessed using complementary (i) physiology-inspired plausibility metrics, (ii) downstream label-consistency evaluation, and (iii) expert listening. Experiments use the Phy-sioNet/Computing in Cardiology Challenge 2016 dataset (3240 recordings) with recording-level splits. After preprocessing and quality control, 16,749 non-overlapping 4 s clips are mapped to a normalised 1 x 128 x 128 log-mel representation to train a conditional 2D U-Net denoiser with classifier-free guidance. Signal-level plausibility is quantified on reconstructed waveforms using three lightweight metrics: an envelope-autocorrelation rhythm score, an amplitude-based explosion score, and the dominant cycle lag. Synthetic clips preserve similar dominant cycle durations but exhibit reduced envelope periodicity and increased transient burstiness relative to real clips. For downstream evaluation, a ResNet-50 classifier achieves 92.24% accuracy on the held-out real test set and 82.8% accuracy on class-balanced synthetic batches, indicating that generated signals retain discriminative structure relevant to normal/abnormal classification. In a pilot expert listening study (60 clips, two clinicians), most synthetic clips are judged as heart-sound-like, while abnormality sensitivity is low for both real and synthetic 4 s excerpts. Overall, the results provide a practical baseline for diffusion-based PCG generation while highlighting remaining challenges in retaining abnormal acoustic cues and reducing reconstruction-induced artefacts.

Background: Respiratory sound classification plays a critical role in the clinical identification of pulmonary pathologies. However, its performance is often hindered by the limited size, severe noise, and class imbalance of real-world auscultation datasets. Although conventional audio augmentation techniques are easy to implement, they may inadvertently distort subtle pathological characteristics. Meanwhile, existing Variational Autoencoder (VAE)- or Generative Adversarial Network (GAN)-based generative approaches often suffer from limited sample fidelity and insufficient controllability over class semantics, particularly under conditions of scarce supervision. Methods: To overcome these limitations, we propose C2GA, a class-controllable generative augmentation framework. C2GA first constructs a semantically rich discrete latent space using a conditional Vector-Quantized Variational Autoencoder (VQ-VAE), in which local acoustic tokens are explicitly decoupled from global class prototypes. Subsequently, a Transformer-based autoregressive prior is trained to generate label-consistent token sequences. These generated tokens are then fused with the corresponding class prototypes and decoded into high-fidelity Mel-spectrograms for data augmentation. Conclusion: These results indicate that C2GA provides an effective and semantically reliable augmentation strategy for respiratory sound analysis. By enabling controllable and high-quality data generation, the proposed framework offers a promising solution for improving the robustness and generalization of respiratory sound classification in realistic clinical scenarios.

The localization of moving sound sources using a microphone array is typically based on modifying the signal to compensate for the Doppler effect. In the time domain this compensation is done on a sample-by-sample basis. In the frequency domain short time segments need to be used in which the Doppler effect is assumed to be approximately constant and a discrete Fourier transform is done on each segment. In contrast, the authors developed an inverse 2.5D localization method for uniformly moving single-frequency sources that works in the spectral domain and allows for the use of longer windows. This was achieved by modifying the 2.5D forward model to directly compute the effect of the motion in the static observer position. The method does neither require to modify the measured signal nor does it require quasi-stationary of the measurements within the window used. Unfortunately, this approach is not directly suitable for broad-band stochastic sources, and in the present work we will investigate how the statistical properties of a uniformly moving stochastic source change when observed at a static observer. Using a 2.5D setting, the relation between the power spectral density of the moving source and the Loève spectrum, which is a generalization of the cross-spectral density at the static receivers, was derived. Based on simulated data with speeds up to 100 m\,s$^{-1}$, the work presented here provides a proof of concept for a method based on multi-taper estimates for the Loève spectrum to localize moving broad-band stochastic sources . Currently, the method requires a stationary source signal and that the spectral density is flat within a certain range around the frequency of interest. Also, correlations between sources are currently not considered.

Acoustic side-channel attacks (ASCA) on keyboards pose a significant security risk, as keystrokes can be inferred from typing acoustics, revealing sensitive information. Prior ASCA studies are limited by small-scale datasets with restricted diversity in users, keyboards, and environments, constraining analysis across devices, microphones, and noise conditions. We introduce HEAR, a dataset designed to study ASCA along three axes: keyboard generalization, noise adaptation, and user bias. HEAR contains recordings from 53 participants using 37 laptop keyboards, collected in three realistic settings: (1) external microphone capture, (2) device microphone capture without network noise, and (3) VoIP-based streaming capture. This enables controlled evaluation across users, keyboards, and environments. On HEAR, we establish an ASCA benchmark spanning conventional features and pre-trained representations from raw audio and spectrograms in unimodal and multimodal settings. We propose DECKER, a domain-invariant keystroke inference framework with four stages: (1) Keyboard Signature Normalization to reduce device coloration, (2) domain-adversarial disentanglement to suppress keyboard identity, (3) supervised cross-keyboard contrastive alignment to enforce key consistency, and (4) Acoustic Style Randomization to synthesize unseen keyboard responses. We further explore sentence-level inference using an LLM-based post-processing layer to refine keystroke sequences via linguistic context. Results on HEAR show DECKER improves keystroke identification over strong baselines, particularly in cross-keyboard and cross-user settings, with further gains from language-model rectification. These findings highlight that ASCA remains effective across diverse users, devices, and noisy environments, underscoring its practical security risk.

We present Echo, a proof-of-concept audio system built around a single 25 M-parameter ViT encoder. The encoder is pretrained with a JEPA objective and then specialised by stages to carry speaker identity, phonetic content, and dynamic source routing in the same 512-dimensional latent space, with no per-task fine-tuning at deployment. Light heads handle diarization (ArcFace + VBx) and dynamic source separation (null-target K-set prediction). On synthetic VoxCeleb2 mixtures with unknown K, the canonical stack reaches 15.00% blind DER, 97.80% PIT separation accuracy with +9.52 dB latent SI-SDR, and a +53.50-point speaker/content factorisation gap on a held-out k-NN probe. The point of Echo is not a new SOTA on any single task but the joint coexistence of three tasks on one encoder at this footprint. We document the design stage by stage, report the dead-ends, and identify the structural wall on end-to-end ASR through the VQ bottleneck that still bounds the PoC.

Objective: laryngectomees depend on an electromechanical device to generate electrolaryngeal (EL) speech. Compared with normal speech, EL speech suffers from severe distortion, limited phonetic variation, unnatural prosody, and temporal shifts, degrading naturalness and intelligibility. Although sequence-to-sequence (seq2seq) voice conversion (VC) based EL-speech-to-normal-speech conversion (EL2SP) is promising, substantial mismatches between EL and normal speech inevitably cause cumulative mapping errors that limit performance. To address this, we describe a novel representation learning framework integrating speech and text representations to improve mapping and reconstruction quality within a seq2seq VC model. Methods: our methodology comprises two main stages: 1) representation integration and learning, and 2) reconstruction training. A network capable of incorporating auxiliary text information is first constructed with pretrained modules to learn speech--text-based integrated representations. Then, an autoencoder-style reconstruction strategy finalizes EL2SP model to inherit these representations without increasing model complexity. We introduce three fusion strategies including middle-, input-, and hybrid-level fusion strategies that progressively enhance learning. Moreover, besides standard seq2seq VC objectives, an additional reconstruction loss on the integrated representation is introduced to refine representation transfer. Results: experiments under different EL2SP datasets consistently demonstrate that our methods, combined with data augmentations, outperform baselines relying solely on speech representations. Furthermore, progressive improvements with system design depth validate the effectiveness of our methods. Significance: the proposed methods provide an extensible and practical methodology for EL speech enhancement and assistive communication technologies.

We address the challenge of generating high-fidelity, long-form soundtracks that remain coherent across scene transitions. Existing AI music systems are mainly designed for short, isolated clips and lack mechanisms to ensure narrative continuity. We present JenBridge, a modular and interpretable framework for adaptive long-form video soundtracking that ensures both high-fidelity audio generation and transition naturalness. The core architecture is a Transformer-based generative model trained with a flow-matching objective, following a two-stage paradigm: pretraining on large-scale text-audio corpora to establish robust musical priors, then adapting to the video domain with dual text-visual conditioning for precise cross-modal alignment. Crucially, to achieve long-form coherence across diverse scene changes, JenBridge incorporates a novel adaptive transition mechanism. This system features a versatile toolkit of transition styles, including a generative transition method, and uniquely employs a Large Language Model (LLM) Agent that acts as a director to select the most appropriate transition for each narrative shift intelligently. To rigorously assess this task, we propose the LVS Benchmark, a new benchmark that includes a curated dataset and novel evaluation metrics focusing on holistic and transition-aware assessment. Extensive experiments on the proposed benchmark demonstrate that JenBridge significantly outperforms existing methods in both objective and subjective metrics, particularly in terms of transition naturalness and overall narrative coherence. JenBridge represents a significant step towards fully automated, professional-quality video soundtracking.

As generative platforms such as Suno and Udio reach human-grade audio quality, the scope of AI's utility has expanded across the entire music production workflow. Beyond simple track generation, these advancements have catalyzed the adoption of AI-driven methodologies in diverse forms. These include vocal synthesis, arrangement, and professional mastering. However, current detection research remains largely confined to a binary `AI-or-human' paradigm. It fails to reflect the realities of contemporary music production workflows. In real-world production, AI tools are increasingly used to refine or master human-produced tracks, and human engineers likewise post-process AI-generated material to ensure professional quality. Moreover, users often employ adversarial tactics to bypass AI detectors, such as applying human mastering to AI-generated tracks. This creates a grey area that a simple binary classification fails to capture. In this paper, we define and investigate ``AI Music Tracking'': the challenge of identifying specific AI integration across the multifaceted spectrum of music production. To this end, we introduce HAIM, a dataset with diverse labels for stages of music production. It is designed to isolate stages of AI intervention, including hybrid production and agent-level tracking. Our evaluation of state-of-the-art detectors reveals systemic flaws. By releasing HAIM, we propose a new benchmark that shifts the field beyond binary classification toward a granular, structured evaluation of AI music.

We propose UniVocal, a unified framework that implicitly infers vocal modes from text context to pioneer Speech-Singing Code-Switching (SCS) Synthesis - a task where transitions are autonomously driven by textual semantics, akin to seamless human language blending. Unlike single-mode generation or systems relying on switching-control tags, our proposed UniVocal implicitly infers vocal modes solely from text context. To achieve this, we employ a data-efficient two-stage curriculum learning strategy that progressively trains a competitive TTS system to acquire the desired SCS capability. Addressing data scarcity, we introduce a scalable pipeline to synthesize diverse code-switching data that is both semantically and acoustically natural, alongside a new multi-scenario benchmark, SCSBench. To address limitations of semantic tokenizers in capturing acoustic details, we also introduce refined cent token and Chain-of-Thought (CoT) generation for planning prosody before content generation, effectively enhancing empathetic speech generation and singing melody. Experimental results demonstrate that UniVocal achieves state-of-the-art performance on SCSBench while maintaining competitive performance on regular speech and singing tasks. Audio samples are available at https://project-univocal-demo.github.io/demo/. The code and dataset are released at https://github.com/FunAudioLLM/FunResearch/tree/main/UniVocal.

This paper presents an overview of DCASE 2026 Challenge Task 2, titled "Noise-aware unsupervised anomalous sound detection (UASD) for machine condition monitoring." The task aims to advance noise-robust anomalous sound detection for machine condition monitoring under the unsupervised setting, where only normal machine sounds are available for training. Reliable detection under noisy conditions is crucial for practical deployment, but previous DCASE Task 2 settings provided limited information about environmental noise, potentially limiting UASD performance in highly noisy situations. To address this limitation, DCASE 2026 allows participants to exploit two-channel audio samples simultaneously captured at locations near and far from the target machine. Since the distant microphone is expected to contain relatively stronger environmental noise and weaker direct machine sounds, it may help distinguish environmental noise components from the target machine sounds. After the challenge submission deadline, challenge results and an analysis of the submitted systems will be added.

Multi-pitch estimation (MPE) typically predicts which pitches are active in a mixture, but not which instrument or source produced them. This paper investigates a lightweight slot-attention framework for multi-instrument MPE (MI-MPE), where a mixture CQT is mapped to an unordered set of source-like pitch maps. The model uses permutation-invariant Hungarian matching to avoid fixed output semantics and treats the number of slots as an upper bound on the number of active sources. We further study two modular extensions: a self-supervised timbre encoder that provides training-time targets for slot-level timbre embeddings, and a polyphony branch that regularizes the pitch density of mixture- and slot-level predictions. Experiments show that Hungarian matching substantially improves instrument family decomposition on URMP. Stem-level prediction remains more challenging: timbre and polyphony supervision improve selected configurations, but do not consistently resolve source assignment. The results suggest that slot-based architectures are a promising direction for source-aware MPE, while highlighting the need to couple auxiliary musical cues to slot identity more carefully.

We present a multimodal dataset of 1020 hours of simultaneously recorded scalp electroencephalography (EEG), facial electromyography (EMG), and speech audio from three healthy native Japanese speakers during open-vocabulary overt speech. Recordings were acquired with three EEG systems-an ultra-high-density system (g.Pangolin) and two cap-type systems (g.SCARABEO and eegosports), spanning 62-128 channels-across many sessions over several months. Each session provides time-synchronized EEG, facial EMG, and audio, together with speech-event annotations and transcriptions. Although collected with speech decoding as a primary motivation, the dataset also supports work on multimodal signal processing, artifact modeling, longitudinal and cross-device adaptation, and EEG representation learning. Technical validation included power spectral density and event-related potential analyses across participants, devices, and tasks, which showed the expected 1/f spectral profile, task-related alpha-band attenuation, and time-locked evoked responses. The dataset is released in Brain Imaging Data Structure (BIDS) format via OpenNeuro under a CC0 waiver to support both speech-related and broader EEG research.

Deep learning has greatly advanced automatic speech recognition (ASR), enabling widespread deployment on edge devices such as smartphones and smart home systems. However, the computational and energy demands of deep neural networks pose significant challenges for such resource-constrained deployments, introducing latency and limiting real-time interaction. Neuromorphic computing offers a promising solution by introducing activation sparsity through spiking neural networks (SNNs) and event-driven neural networks, converting dense operations into sparse computations. However, a study that evaluates the hardware benefits of different neuromorphic strategies remains lacking for ASR. This paper explores spiking and event-driven neuromorphic neural networks to improve activation sparsity in the state-of-the-art SpeechMamba model for ASR. We introduce an event-driven SpeechMamba with FATReLU activation, achieving over 60% activation sparsity with less than 1% accuracy degradation on LibriSpeech. We also propose a spiking SpeechMamba that attains over 70% sparsity while using 30% fewer parameters than comparable SNNs. Finally, we develop a cycle-accurate event-driven simulator enabling flexible algorithm-hardware co-exploration, which helps us identify computational bottlenecks and yields over 10% additional efficiency improvements.

Automatically distinguishing child-directed speech from adult-directed speech in long-form recordings is key to scalable analyses of children's language environments. Existing approaches process utterances in isolation and have been evaluated primarily on English. We address these gaps along three dimensions. First, we fine-tune and evaluate six-self supervised models on a multilingual dataset of 182 children, showing that in-domain pre-training on child-centered recordings substantially outperforms models trained on adult speech. Second, we demonstrate that incorporating surrounding context substantially improves classification, with an absolute gain of 13.8% in average F1-score. Third, we evaluate our model in a realistic end-to-end pipeline, from adult speech detection to addressee classification, showing that performance drops under automatic segmentation but still consistently outperforms a rule-based baseline.

Modern audio processing networks are commonly deployed on accelerators whose peak throughput is obtained through dense linear algebra, whereas conventional acoustic frontends -- a Short-Time Fourier Transform (STFT) followed by sparse Mel aggregation -- remain structurally heterogeneous. This mismatch can introduce memory-bandwidth, dispatch, and intermediate-allocation overheads on contemporary accelerator backends. This work introduces MelT, a single-stage frontend framework in which Mel-spaced Non-Uniform Discrete Fourier Transform (NDFT) bases are precomputed and applied to time-domain acoustic frames through dense General Matrix Multiplication (GEMM) operations. The contribution is not the NDFT operator itself; rather, it is the formulation of Mel-spaced NDFT projection as a GEMM-native audio frontend and its evaluation as a hardware-efficient alternative to conventional STFT+Mel pipelines. Evaluated across platforms ranging from Apple A18 Pro edge hardware to NVIDIA H100 datacenter acceleration, MelT attains up to a $3.75\times$ speedup in inference latency and a $3.52\times$ reduction in energy consumption while maintaining downstream classification accuracy.

Empathetic spoken dialogue systems must infer a user's emotional state to respond appropriately, yet everyday speech often carries weak, neutral, or ambiguous affective cues. To address this, we introduce Sympatheia, a speech-to-speech dialogue framework conditioned on affect inferred from the user's speech and, when available, explicit affect specifications provided as a continuous valence--arousal (VA) control signal by a multimodal sensing module or user interface. To train our model, we construct Sympatheia-18k, an emotion-conditioned synthetic spoken dialogue corpus with 12 emotion anchors. This dataset includes an emotional split for learning affective speech behavior, and a neutral split that pairs emotionally neutral queries with multiple emotion-conditioned responses to isolate explicit emotion control in emotionally ambiguous cases. Empirical results show that Sympatheia outperforms speech conversational baselines in generating responses whose semantic content and spoken delivery are both emotionally appropriate. We further show that the same VA interface can integrate emotion estimates from diverse sensing modules, including facial expression, biosignals, and textual affect descriptions, improving response alignment when speech alone provides limited emotional evidence. These results suggest that continuous affect conditioning is an effective practical step for building emotionally adaptive voice assistants.

A simple way to improve the performance of almost any machine learning model is not to train a single but several models with diverse algorithms which will make slightly distinct kinds of predictions and errors on the same data, and thus improve the average predictions and robustness. However, making predictions using a whole ensemble of models is cumbersome and computationally too expensive to allow deployment to a large number of users, especially if the models are large neural nets. In response to this, we introduce a layer and point wise projection mapping, which maps student and teacher representations into an aligned high-dimensional embedding space during training process. The proposed approach combined with LoRA injection reduces the student model trainable parameters to less than 1% of the teacher model, while significantly improving word error rate (WER) compared to other distillation methods, as demonstrated in ablation studies. Unlike a mixture of experts, our method can be trained rapidly and in parallel.

We address the problem of out-of-distribution (OOD) detection for target observations embedded in a subspace of the high dimensional data space. Using continuous normalizing flows (CNFs), we propose a Lagrangian sub-flow (LSF) framework designed to isolate and estimate the density for the relevant components in the representation and using the remaining components as context. Through experimentation with models for speech synthesis, we show that CNFs, similarly to other deep generative models (DGMs), are susceptible to the "likelihood paradox", where high likelihood is erroneously assigned to OOD samples. This is attributed to the inductive bias of DGMs that prioritize low-level structural details over high-level semantic coherence. To mitigate this phenomenon, we propose a number of geometric diagnostic signals based on the velocity field over the sub-flow trajectory. Based on these signals, we design metrics for the challenging task of zero-shot phoneme-level mispronunciation detection. Finally, we demonstrate the superiority of these metrics compared to likelihood-based methods on a real-world mispronunciation detection benchmark.

Face-to-face speech comprehension is inherently multimodal, integrating acoustic signals with visible articulation, facial expression, head motion, and other socially relevant cues. While audiovisual speech systems typically focus on the mouth region as the primary visual source of linguistic information, affective facial expressions are often treated separately as emotion-recognition targets. This paper investigates whether upper-face affective information contributes to audiovisual sentence recognition beyond audio and mouth-region cues, particularly under acoustic degradation. Using the CREMA-D audiovisual emotional speech corpus, we train feature-based sentence classifiers under four cue conditions: audio only (A), audio plus mouth/lower-face features (A+M), audio plus upper-face features (A+U), and audio plus both mouth and upper-face features (A+M+U). Models are evaluated on clean audio and pink-noise conditions at +10 dB, +5 dB, and 0 dB SNR using actor-independent splits. Results show that mouth/lower-face features provide substantial robustness benefits under degraded audio. At 0 dB SNR, A+M improves accuracy over A by 0.0794, with an actor-bootstrap 95% confidence interval of [0.0296, 0.1298]. Upper-face affective cues exhibit a more nuanced effect. Although the direct accuracy gain of A+M+U over A+M is small, full-face models consistently improve calibration across SNR levels and outperform shuffled upper-face controls under noisy conditions. These findings suggest that affective facial information may support multimodal robustness and confidence estimation under acoustic uncertainty without directly encoding lexical content. More broadly, the study highlights the potential role of socially expressive facial cues in human-centered audiovisual interaction systems.

Sound design workflows frequently oscillate between time-consuming library searches and the complexity of procedural synthesis, with practitioners typically relying on disconnected tools to address each challenge separately. This paper introduces Quality Audio Prototyping (QuAP), a working prototype that unifies content-based audio retrieval and procedural sound generation within a single interface, reducing the procedural distance between a narrative concept and its sonic realisation. QuAP integrates a similarity-based retrieval engine with real-time procedural audio models, complemented by a rule-based assistant that provides perceptually informed parameter guidance, offering definitions and recommendations derived from empirical optimisation rather than requiring prior synthesis knowledge. Preliminary evaluation confirms the viability of this approach: subjective assessment demonstrated statistically significant quality improvements in five of six embedded synthesis models, and an encoder ablation study established the preferred retrieval architecture on a sound effect dataset. A user evaluation with 16 practitioners confirmed the tool's workflow utility, with all participants agreeing that the parameter assistant preserved creative agency while lowering the barrier to procedural interaction.

Distributed Acoustic Sensing (DAS) enables large-scale monitoring through optical fibers, but its high dimensionality and complex spatio-temporal patterns make event classification demanding. Existing deep learning approaches-CNNs, recurrent models, and Transformer variants-either fail to capture long-range dependencies or require processing raw DAS matrices at prohibitive cost. We propose DAStatFormer, a hybrid multibranch Transformer that combines compact multidomain statistical features with Gated Transformer Networks. Instead of raw signals, we extract 24 ANOVA-selected attributes per channel from the temporal, waveform, and spectral domains, reducing data size by orders of magnitude while preserving discriminative information. Each domain is processed via dedicated step-wise and channel-wise attention branches, fused by an adaptive gating mechanism. Experiments on the open $Φ$-OTDR benchmark and a real-scenario DAS dataset show that DAS-tatFormer achieves up to 99.4% accuracy and near-perfect real-world performance, while using significantly fewer parameters and lower inference cost than models such as DASFormer and DeepViT. These results demonstrate its suitability for scalable, real-time DAS-based monitoring. We release our code at https://github.com/MichelD-git/DAStatFormer

Diffusion and flow-matching based text-to-speech (TTS) models excel in naturalness but often lack explicit emotion control, as emotional signals remain entangled with speaker identity. We discover that emotion embedding emerges as a linearly decodable direction of frozen hidden states, nearly orthogonal to the direction embedding speaker identity. This inspires a plug-and-play framework DUET for emotion control over pretrained diffusion and flow-matching based TTS models. During generation, DUET unifies dual-space control to achieve fine-grained emotion intervention in a single per-step update: hidden space steering shifts generation along the target emotion direction, while mel-space guidance refines spectral details through gradients backpropagated from a differentiable vocoder. We validate DUET on five architecturally diverse pretrained TTS backbones across three datasets, where it outperforms 10 supervised state-of-the-art emotional TTS baselines across paradigms and achieves the highest human-rated emotion appropriateness. To further showcase its qualitative behavior, we deploy DUET on an Ameca humanoid robot, where it produces richly expressive emotional speech on the humanoid, demonstrating the strong potential for plug-and-play affective interaction for embodied agents.

We present Chatterbox-Flash, a zero-shot text-to-speech model obtained by fine-tuning a pretrained autoregressive TTS decoder into a block-diffusion decoder, enabling parallel token generation within each block while retaining block-by-block streaming. We find that naively transferring mainstream block-diffusion decoding to discrete speech tokens degrades quality, as a long-tail token distribution biases parallel position selection toward a few high-frequency tokens. To mitigate this without architectural modification, we introduce two inference-time techniques: prior-calibrated scoring, which subtracts the block-level marginal token distribution, and an early-decoding schedule, which adaptively terminates iteration based on calibrated confidence. On standard zero-shot TTS benchmarks, Chatterbox-Flash attains high-fidelity synthesis comparable to strong autoregressive and non-autoregressive baselines, while supporting streaming inference with time-to-first-packet on par with streaming AR systems and substantially lower real-time factor. Code and audio samples are available at https://github.com/resemble-ai/chatterbox-flash.

Unified speech foundation models require a holistic tokenization space that is both learnable by language models and decodable into high-quality waveforms. Existing speech tokenizers, however, often fail to satisfy these requirements simultaneously, leading to increased architectural complexity and more involved training designs. We propose HoliTok, a continuous Holistic speech Tokenization model designed for unified generation-understanding modeling. HoliTok encodes 48~kHz speech into a compact 25~Hz sequence of 128-dimensional latents. It is trained with a progressive strategy that jointly preserves signal-level fidelity, incorporates semantic information, and maintains strong latent learnability. Based on this tokenization, we build a unified AR+DiT model for speech synthesis and recognition, where the same latent sequence supports both generation-specific and unified generation-understanding tasks. Experiments show that HoliTok achieves competitive reconstruction fidelity, improves generative learnability for high-quality and controllable synthesis, and, among the evaluated representations, is the only one that operates robustly in our unified generation-understanding architecture without additional optimization tricks. These results suggest that HoliTok serves as an effective speech tokenizer and a foundational representation interface for unified spoken language modeling. The code is available at: https://github.com/bovod-sjtu/HoliTok.

Tokenizing music to fit the general framework of language models is a compelling challenge, especially considering the diverse symbolic structures in which music can be represented (e.g., sequences, grids, and graphs). To date, most approaches tokenize symbolic music as sequences of musical events, such as onsets, pitches, time shifts, or compound note events. This strategy is intuitive and has proven effective in Transformer-based models, but it treats the regularity of musical time implicitly: individual tokens may span different durations, resulting in non-uniform time progression. In this paper, we instead consider whether an alternative tokenization is possible, where a uniform-length musical step (e.g., a beat) serves as the basic unit. Specifically, we encode all events within a single time step at the same pitch as one token, and group tokens explicitly by time step, which resembles a sparse encoding of a piano-roll representation. We evaluate the proposed tokenization on music continuation and accompaniment generation tasks, comparing it with mainstream event-based methods. Results show improved musical quality and structural coherence, while additional analyses confirm higher efficiency and more effective capture of long-range patterns with the proposed tokenization.

Joint audio-video generation models are rapidly approaching professional production quality, raising a central question: do they understand audio-visual physics, or merely generate plausible sounds and frames that violate real-world consistency? We introduce AV-Phys Bench, a benchmark for evaluating physical commonsense in joint audio-video generation. AV-Phys Bench tests models across three scene categories: Steady State, Event Transition, and Environment Transition. It covers physics-grounded subcategories drawn from real-world scenes, plus Anti-AV-Physics prompts that deliberately request physically inconsistent audio-video behavior. Each generation is evaluated along five dimensions: visual semantic adherence, audio semantic adherence, visual physical commonsense, audio physical commonsense, and cross-modal physical commonsense. Across three proprietary and four open-source models, we find that Seedance 2.0 performs best overall, but all models remain far from robust physical understanding. Performance drops sharply on event-driven and environment-driven transitions, and even strong proprietary systems collapse on Anti-AV-Physics prompts. We further introduce AV-Phys Agent, a ReAct-style evaluator that combines a multimodal language model with deterministic acoustic measurement tools, producing rankings that closely align with human ratings. Our results identify cross-modal physical consistency and transition-driven scene dynamics as key open challenges for joint audio-video generation.

Audio-text retrieval systems based on Contrastive Language-Audio Pretraining (CLAP) achieve strong performance on traditional benchmarks; however, these benchmarks rely on caption-style queries that differ substantially from real-world search behavior, limiting their assessment of practical retrieval robustness. We present Omni-Embed-Audio (OEA), a retrieval-oriented encoder leveraging multimodal LLMs with native audio understanding. To systematically evaluate robustness beyond caption-style queries, we introduce User-Intent Queries (UIQs) - five formulations reflecting natural search behaviors: questions, commands, keyword tags, paraphrases, and exclusion-based negative queries. For negative queries, we develop a hard negative mining pipeline and propose discrimination metrics (HNSR, TFR) assessing models' ability to suppress acoustically similar distractors. Experiments on AudioCaps, Clotho, and MECAT show that OEA achieves comparable text-to-audio retrieval performance to state-of-the-art M2D-CLAP, while demonstrating clear advantages in two critical areas: (1) dominant text-to-text retrieval (+22% relative improvement), and (2) substantially superior hard negative discrimination (+4.3%p HNSR@10, +34.7% relative TFR@10), revealing that LLM backbones provide superior semantic understanding of complex queries.

Voice cloning is often evaluated in terms of overall quality, but less is known about accent preservation and its perceptual consequences. We compare standard and heavily accented Mandarin speech and their voice clones using a combined computational and perceptual design. Embedding-based analyses showed larger original-clone distances for accented speakers in several speaker-discriminative embedding spaces, but this difference disappeared after normalizing against each speaker's within-original baseline variability. In the perception study, clones are rated as more similar to their originals for standard than for accented speakers, and intelligibility increases from original to clone, with a larger gain for accented speech. These results show that accent variation can shape perceived identity match and intelligibility in voice cloning even when it is not reflected in baseline-normalized speaker-embedding distance, and they motivate treating accent preservation as an explicit component of speaker identity preservation, rather than assuming that it is fully captured by off-the-shelf speaker-discriminative embeddings.

Spatial audio enhances immersion by reproducing 3D sound fields, with Ambisonics offering a scalable format for this purpose. While first-order Ambisonics (FOA) notably facilitates hardware-efficient acquisition and storage of sound fields as compared to high-order Ambisonics (HOA), its low spatial resolution limits realism, highlighting the need for Ambisonics upscaling (AU) as an approach for increasing the order of Ambisonics signals. In this work we propose DiffAU, a cascaded AU method that leverages recent developments in diffusion models combined with novel adaptation to spatial audio to generate 3rd order Ambisonics from FOA. By learning data distributions, DiffAU provides a principled approach that rapidly and reliably reproduces HOA in various settings. Experiments in anechoic conditions with multiple speakers, show strong objective and perceptual performance.

Recent advances in end-to-end trained omni-models have substantially improved audio capabilities by strengthening text-audio modality alignment. However, whether such alignment inadvertently facilitates the transfer of safety vulnerabilities across modalities remains underexplored. This question is critical as text-based jailbreak attacks are considerably more mature than audio-based ones; if they transfer systematically, current audio safety evaluations may underestimate risks originating from the text modality. In this paper, we introduce the Alignment Curse, a formally characterized and empirically validated principle showing that stronger modality alignment enables more effective transfer of attacks from text to audio, revealing a fundamental tension between capability and safety. Motivated by this principle, we conduct a comprehensive black-box evaluation of three attack categories on recent omni-models (e.g., Qwen2.5-Omni, Qwen3-Omni): text attacks, text-transferred audio attacks, and audio attacks. We find that text-transferred audio attacks perform comparably to, and often better than, audio-based attacks, exhibiting a clear advantage under audio-only access. This suggests that text-based vulnerabilities play a pivotal role in shaping audio safety risks. Finally, we empirically analyze the relationship between modality alignment and transfer effectiveness across attack methods and models, observing consistent support for the Alignment Curse: tighter modality alignment leads to more effective cross-modality attack transfer.

Scaling Multimodal Large Language Models (MLLMs) to long-form speech is bottlenecked by the explosive growth of input tokens. Unlike images or videos, audio lacks overlapping information, making extreme 1-token compression highly susceptible to the loss of fine-grained acoustic cues. To overcome this, we propose FastSLM, a token-efficient architecture featuring the Hierarchical Temporal Abstractor (HTA). HTA progressively distills non-overlapping acoustic features across multiple temporal scales, achieving an extreme compression rate of 1.67 tokens per second a 97% reduction without losing critical context. Experimental results show that FastSLM achieves competitive performance with state-of-the-art models on long-form benchmarks despite operating with significantly fewer FLOPs and parameters. The source code and model checkpoints are available at https://anonymous.4open.science/r/FastSLM-8BD3.

Knowledge distillation (KD) is one of the most effective paradigms for compressing large-scale foundation models into deployable architectures. In the context of Automatic Speech Recognition (ASR), previous studies have predominantly focused on forcing the student model to strictly mimic the predictive distribution of a massive teacher model. However, this static dependency often presents an inherent trade-off: while the student rapidly acquires basic linguistic representations, it simultaneously inherits the teacher's domain-specific blind spots and over-confident hallucinations, leading to a severe decline in out-of-distribution generalization capacity. To effectively mitigate this issue, we propose Adaptive Self-Knowledge Distillation (ASKD), a dynamic curriculum framework. ASKD systematically decays the dependency on the teacher's distribution as training progresses-thereby unlocking the student's independent reasoning capacity-and subsequently employs a self-knowledge distillation phase to act as a structural regularizer. By applying ASKD, we distill the massive Whisper architecture into a compact variant, ASKD-Whisper. In our comprehensive evaluations across diverse acoustic domains, ASKD-Whisper not only achieves a 5x speedup in inference latency but also outperforms its teacher model by yielding a 1.07% lower word error rate (WER). These results demonstrate that ASKD effectively prevents teacher-induced overfitting and establishes a new state-of-the-art for generalizable model compression.

Audio Language Models (ALMs) offer a promising shift towards explainable audio deepfake detections (ADD), moving beyond \textit{black-box} classifiers by providing transparency to their predictions via reasoning traces. However, such reasoning may not support the model predictions, reflecting poor coherence, or, worse, may rationalize incorrect predictions with plausible but misleading explanation. Moreover, the behavior of ALM reasoning under adversarial attacks remains under-explored, raising questions about the practical reliability of such explanation capabilities. To address this gap, this study introduces \textbf{SARA} (\textbf{S}hift \textbf{A}nalysis of \textbf{R}easoning in \textbf{A}udio), a diagnostic framework that evaluates ALM reasoning across three dimensions: acoustic perception, reasoning-verdict coherence and dissonance. We test five open-source ALMs against both acoustic and linguistic adversarial attacks. We show that acoustic attacks significantly degrade reasoning-verdict coherence (average decrease of 14.20\%), frequently inducing internal logical conflicts. Conversely, linguistic attacks achieve higher attack success rates while maintaining reasoning coherence. We further demonstrate that the textual coherence of generated reasoning traces also serves as a latent indicator of adversarial inputs, enabling effective detection of perturbed audio (0.78 in F1) \textit{without accessing the raw acoustic signal}. These findings suggest that reasoning traces provide diagnostic utility that persists even when final classification outputs are compromised.

General-purpose audio representations aim to map acoustically variable instances of the same event to nearby points, resolving content identity in a zero-shot setting. Unlike supervised classification benchmarks that measure adaptability via parameter updates, we introduce VocSim, a training-free benchmark probing the intrinsic geometric alignment of frozen embeddings, with no parameters updated and no labels used (a label-free PCA whitening is fit per subset to correct anisotropy). VocSim aggregates 125k single-source clips from 19 corpora spanning human speech, animal vocalizations, and environmental sounds, isolating content representation from source separation (polyphonic mixtures are out of scope). We evaluate embeddings with Precision@k for local purity and the Global Separation Rate (GSR) for point-wise class separation, calibrated by lift over an empirical permutation baseline. A simple pipeline of frozen Whisper features, time-frequency pooling, and label-free PCA yields strong zero-shot performance with stable GSR rankings across domains (Kendall's tau = 0.60). However, on blind low-resource speech (Shipibo-Conibo, Chintang), local retrieval collapses while remaining above chance, exposing a cross-lingual speech generalization gap. As external validation, our top embeddings predict avian perceptual similarity, improve bioacoustic classification, and achieve state-of-the-art on the HEAR benchmark. We release data, code, and a public leaderboard.

This study presents a systematic evaluation of time-frequency feature design for binaural sound source localization (SSL), focusing on how feature selection influences model performance across diverse conditions. We investigate the performance of a convolutional neural network (CNN) model using various combinations of amplitude-based features (magnitude spectrogram, interaural level difference - ILD) and phase-based features (phase spectrogram, interaural phase difference - IPD). Evaluations on in-domain and out-of-domain data with mismatched head-related transfer functions (HRTFs) reveal that carefully chosen feature combinations often outperform increases in model complexity. While two-feature sets such as ILD + IPD are sufficient for in-domain SSL, generalization to diverse content requires richer inputs combining channel spectrograms with both ILD and IPD. Using the optimal feature sets, our low-complexity CNN model achieves competitive performance. Our findings underscore the importance of feature design in binaural SSL and provide practical guidance for both domain-specific and general-purpose localization.

Individual Head-Related Transfer Functions (HRTFs) are starting to be introduced in many commercial immersive audio applications and are crucial for realistic spatial audio rendering. However, one of the main hesitations regarding their introduction is that creating individual HRTFs is impractical at scale due to the complexities of the HRTF measurement process. To mitigate this drawback, HRTF spatial upsampling has been proposed with the aim of reducing the measurements required. While prior work has seen success with different machine learning (ML) approaches, these models often struggle with long-range preservation of local spatial variation patterns across neighbouring source directions and generalization at high upsampling factors. In this paper, we propose a novel transformer-based architecture for HRTF upsampling, leveraging the attention mechanism to better capture spatial correlations across the HRTF sphere. Working in the spherical harmonic (SH) domain, our model learns to reconstruct high-resolution HRTFs from sparse input measurements with significantly improved accuracy. To enhance spatial coherence, we introduce a neighbour dissimilarity loss that promotes magnitude smoothness, yielding more realistic upsampling. We evaluate our method using both perceptual localization models and objective spectral distortion metrics. Experiments show that our model outperforms existing methods across several evaluation metrics in generating realistic, high-fidelity HRTFs.

Lyrics translation requires both accurate semantic transfer and preservation of musical rhythm, syllabic structure, and poetic style. In animated musicals, the challenge intensifies due to alignment with visual and auditory cues. We introduce Multilingual Audio-Video Lyrics Benchmark for Animated Song Translation (MAVL), the first multilingual, multimodal benchmark for singable lyrics translation. By integrating text, audio, and video, MAVL enables richer and more expressive translations than text-only approaches. Building on this, we propose Syllable-Constrained Audio-Video LLM with Chain-of-Thought SylAVL-CoT, which leverages audio-video cues and enforces syllabic constraints to produce natural-sounding lyrics. Experimental results demonstrate that SylAVL-CoT significantly outperforms text-based models in singability and contextual accuracy, emphasizing the value of multimodal, multilingual approaches for lyrics translation.

Zero-shot learning enables models to generalise to unseen classes by leveraging semantic information, bridging the gap between training and testing sets with non-overlapping classes. While much research has focused on zero-shot learning in computer vision, the application of these methods to environmental audio remains underexplored, with poor performance in existing studies. Generative methods, which have demonstrated success in computer vision, are notably absent from zero-shot environmental sound classification studies. To address this gap, this work investigates generative methods for zero-shot learning in environmental audio. Two successful generative models from computer vision are adapted: a cross-aligned and distribution-aligned variational autoencoder (CADA-VAE) and a leveraging invariant side generative adversarial network (LisGAN). Additionally, we introduced a novel diffusion model conditioned on class auxiliary data. Synthetic embeddings generated by the diffusion model are combined with seen class embeddings to train a classifier. Experiments are conducted on five environmental audio datasets, ESC-50, ARCA23K-FSD, FSC22, UrbanSound8k and TAU Urban Acoustics 2019, and one music classification dataset, GTZAN. Results show that the diffusion model outperforms all baseline methods on average across six audio datasets. This work establishes the diffusion model as a promising approach for zero-shot learning and introduces the first benchmark of generative methods for zero-shot environmental sound classification, providing a foundation for future research.