Audio Language Models

This work investigates the effect of batch sampling strategies during training for text-to-audio music generation under low-data and small-scale model settings. This paper describes our approach and findings for the ICME 2026 Grand Challenge on Academic Text-to-Music Generation. Training data are clustered using either text embeddings or audio embeddings, and samples with similar characteristics are grouped within the same mini-batch to mitigate gradient interference. The effects of modality and cluster granularity on clustering are analyzed. Results show that clustering based on text embeddings achieves better performance on objective evaluation metrics than clustering based on audio embeddings. In addition, different cluster granularity leads to different behaviors across evaluation criteria: a moderate number of clusters performs best on objective metrics, while a larger number of clusters tends to exhibit music with more coherent structure in listening tests.

Recognizing jazz standards from audio is a challenging form of tune-level music retrieval: different performances of the same standard may vary in tempo, key, arrangement, instrumentation, improvisational content, and even whether the head melody is present. We study this problem using a curated subset of the Jazz Trio Database designed for cross-performance standard recognition. We compare a from-scratch trained Harmonic CNN baseline against frozen pretrained music representations from recent music understanding foundation models, using both supervised probing and nearest-neighbor retrieval. Our results suggest that from-scratch spectrogram models overfit strongly to training performances, while pretrained embeddings provide better top-$k$ results but are sensitive to performer identity, which can be partially reduced with a lightweight contrastive projection. Our findings motivate jazz standard recognition as a useful stress test for music representation models and as a step toward retrieval-based standard identification. Project page: https://github.com/cagries/tipofmyear.

Large audio-language models (LALMs) frequently hallucinate by overriding acoustic evidence with language priors. While contrastive decoding (CD) offers training-free mitigation, existing methods rely on blunt perturbations like masking or noise, leaving structured audio transformations unexplored. We explore this design space by evaluating a diverse library of targeted audio perturbations and adaptively selecting the optimal negative branch for each task and example. First, we improve upon earlier prompt engineering by showing that a simple binary yes/no constraint reduces the model's tendency to falsely confirm absent audio features. Second, evaluating our library across temporal, spectral, frequency, and amplitude domains reveals that optimal transformations are highly task-dependent; for instance, reversing the audio array disrupts temporal coherence, raising accuracy on the temporal order task from 74.7% to 81.4%. Finally, we trained a light-weight perturbation selector on model hidden states to dynamically route negative branches, yielding an additional +4.3% gain on the existence task.

Audio-Language Models (ALMs) achieve strong zero-shot performance by aligning audio with textual class descriptions. Although prompt learning improves accuracy on base classes through few-shot supervised adaptation, we observe a critical trade-off: it often degrades performance on novel classes, sometimes falling below zero-shot accuracy. This exposes a base-to-novel generalization gap in prompt learning for ALMs. To address this issue, we propose \textbf{ZEBRA} (Zero-shot Entropy-Regularized Prompt Learning for Base-to-Novel Generalization), a plug-and-play framework that fuses zero-shot logits with prompt-learning logits, and employs self-entropy regularization to reduce overfitting to base classes. Experiments across multiple audio classification datasets show that ZEBRA consistently improves novel-class performance while maintaining strong base accuracy, significantly reducing the base-to-novel gap compared to standard prompt learning. The code is available at: https://github.com/asif-hanif/zebra.

Some neural audio codecs disentangle speech into latent subspaces encoding content, speaker identity, and acoustics, enabling acoustic teleportation and voice conversion. Existing evaluations rely on cross-reconstruction quality, which cannot reliably detect leakage across partitions. We extend a probing based framework to assess disentanglement by regressing room-acoustic parameters (reverberation time, clarity, and direct-to-reverberant ratio) and classifying speaker identity, using the gap between intended and unintended partitions as the disentanglement measure. Applied to an acoustic teleportation codec, we find speaker identity is largely confined to its partition, while acoustics leak into the speech embeddings due to the training objective. Acoustic embeddings blindly estimate room parameters within 0.02 s of supervised baselines, indicating physically meaningful structure emerges without explicit supervision.

Recent music audio-language models achieve high accuracy on instrument question-answering benchmarks, but it remains unclear whether this reflects robust audio grounding or benchmark-specific shortcuts. In this paper, we introduce an OpenMIC-derived diagnostic benchmark sequence for instrument grounding in music audio-language models, extending binary instrument-presence QA to genre-prior-reduced examples, confusable instrument discrimination, longer audio context, and temporal localization. Across these settings, high binary QA accuracy often fails to predict model behavior: models can exhibit option-position bias, confusable-instrument errors, and temporal response bias. These results suggest that instrument grounding should be evaluated with multi-axis diagnostic benchmarks rather than a single aggregate accuracy.

Diffusion-based text-to-audio (TTA) models achieve impressive synthesis quality but suffer from high inference latency due to iterative multi-step denoising. Existing one-step approaches alleviate this issue but still rely on paired text--audio data during distillation. To address these limitations, we propose SwiftAudio, a one-step TTA framework that performs audio-free distillation from a pretrained diffusion teacher using only text captions. Specifically, we adapt Variational Score Distillation (VSD) to the audio domain and introduce a temporal smoothness regularization objective to encourage coherent latent audio representations. This design enables the student model to inherit the teacher's generative prior without requiring paired audio supervision and allows effective training with only approximately 45K captions. Experiments on AudioCaps and Clotho demonstrate that SwiftAudio achieves state-of-the-art performance among strict one-step methods and substantially narrows the gap to multi-step diffusion systems. Project page: https://swiftaudio.org/

Strong speech-to-text (S2T) LLMs already provide robust speech perception and text reasoning, but adding speech-to-speech (S2S) output is challenging: fine-tuning the backbone can degrade the original S2T performance, while attaching a downstream talker reintroduces a serial text-to-speech bottleneck. We present PRIME-Speech, a frozen-backbone S2S conversion framework that trains only speech-generation modules. PRIME-Speech synchronizes a causal audio post-decoder with intermediate hidden states of the frozen backbone, so codec tokens are generated from the model's evolving reasoning trajectory rather than from completed text chunks. The post-decoder uses mixed hidden-state, text, and audio-history conditioning, and a training-time packing strategy with turn-level audio KV-cache and position reset stabilizes multi-turn spoken interaction without additional multi-turn S2S training data. Multi-token prediction further reduces the effective codec prediction rate and improves first-audio latency without modifying the reasoning path. Across speech translation, spoken QA, speech understanding, and multi-turn dialogue, PRIME-Speech preserves the S2T behavior of the frozen backbone while producing accurate, low-WER spoken responses.

Diffusion Transformers (DiTs) have significantly advanced audio-driven portrait animation, but their high computational cost leads to substantial inference latency. Although training-free diffusion caching accelerates inference significant, existing methods are primarily developed for text-conditioned generation and overlook the spatial and modality imbalances inherent in audio-driven portrait animation. In this paper, we propose SyncCache, a training-free caching acceleration method tailored for DiT-based portrait animation that explicitly exploits asymmetric dynamics. Specifically, high-frequency dynamics driven by audio conditions and concentrated in human regions are more challenging and critical to cache and reuse than the low-frequency visual background in portrait animation. First, we introduce Spatially-Asymmetric Probing to prioritize error sensitivity in dynamic human region. Second, through Modality-Decoupled Caching, we bypass heavy DiT block by reusing stable inter-block residuals, while continuously recomputing lightweight audio blocks to preserve precise lip synchronization. Furthermore, we introduce a cache ratio to control cache capacity and formulate memory-adaptive cache selection as an offline dynamic programming problem without online overhead. Extensive experiments demonstrate that SyncCache achieves superior speed-quality trade-offs, delivering up to 4.12x acceleration on HunyuanVideo-Avatar and 3.75x on Wan-S2V with near-lossless visual fidelity and precise audio alignment.

Text-based singing voice editing (SVE) aims to revise sung lyrics while preserving the original melody, total duration, and non-edited regions. In this paper, we propose MeloDISinger, a flow-matching-based SVE model for melody-aware and duration-preserving editing. Its core module, MeloDRP, predicts fixed-budget duration ratios, enabling explicit span-wise duration control. For melody-aware duration allocation, MeloDRP fuses phonetic cues with pseudo-MIDI melodic context through cross-attention, while temporal-overlap supervision encourages soft phoneme--note correspondences. We further use a flow-matching mel decoder for audio infilling to synthesize edited regions while preserving surrounding context. In addition, we introduce a duration-aware edited-lyric generation pipeline using WhisperX and an LLM to construct feasible evaluation scenarios. Experiments demonstrate state-of-the-art performance in both objective and subjective evaluations.

Accurately predicting the temporal evolution of clinical biomarkers is crucial for the early diagnosis and management of neurodegenerative diseases such as Alzheimer's disease. However, this relies on longitudinal data to capture biomarker changes over time, which is often sparse and irregular due to the high cost, labor-intensive nature, and patient burden. To address these challenges, we propose ENC-ODE, an Event-level Neurodegenerative modeling in Continuous time with neural Ordinary Differential Equations. ENC-ODE predicts future biomarker evolution by modeling clinical events through diagnosis-conditioned continuous dynamics. A target-conditioned attention mechanism weights and aggregates event-level predictions for the target time and modality without history compression. Extensive experiments on Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset demonstrate that ENC-ODE outperforms representative sequence models while offering a scalable and neuroscientifically grounded solution for clinical support. The code is available at https://github.com/JardinDelSol/enc-ode.

Industry-scale video and live-streaming moderation imposes requirements that are difficult to satisfy with generic pretrained public models or external APIs, including adaptation to platform-specific data distributions, policy-specific objectives, and product-level safety constraints. As a result, platforms must undertake internal model development, naturally turning to shared public research for guidance. However, existing multimodal foundation-model studies primarily report architectures, training recipes, data scaling strategies, and benchmark results, but provide less systematic guidance on how failures should be localized and translated into targeted model-development interventions. Interventions are essential because deployment failures are rarely self-explanatory. Similar failures can originate from different causes. Without targeted interventions, improvement reduces to heuristic trial-and-error, where benchmark improvements are weakly attributable, and failures are difficult to trace to their underlying causes. To address this gap, we present a diagnostic methodology for industry-scale Audio-Visual-Language Models AVLM development. The methodology maps model failures into a taxonomy of observable failure signatures and links each class of failure to an intervention space. We instantiate this methodology across the development and alignment lifecycle of an AVLM foundation model for a large-scale video and live-streaming platform. The resulting system supports over 100 regions and is designed for noisy, ambiguous, and highly diverse content drawn from global platform traffic.

Self-supervised learning enables audio representations that transfer across domains and tasks. We present BEST-RQ-2, an evolution of BEST-RQ that retains frozen randomprojection-based discrete targets while introducing a two-step contextualize-then-predict pretraining scheme. A ViT context encoder processes only the unmasked spectrogram regions, and a lightweight predictor infers targets for the masked regions; the predictor is discarded after pretraining. Replacing the original Conformer encoder with a ViT shifts performance across domains, slightly reducing speech performance while improving music and environmental sounds, with comparable average scores. The main improvement comes from decomposing masked prediction into separate contextualization and prediction stages. On the X-ARES and XARES-LLM benchmarks, BEST-RQ-2 consistently outperforms one-stage baselines in overall transfer while keeping inference compute unchanged. Code and model checkpoints are publicly available.

Audio-Visual Speech Recognition takes two input modalities, acoustic and visual streams, where visual information from lip movements aids recognition when audio is noisy. Recently, LLM-based AVSR models have emerged as a promising paradigm by connecting pre-trained audio-visual encoders to an LLM, achieving strong results in clean conditions. However, these models are predominantly optimized for clean acoustic conditions, with limited attention to making the LLM backbone robust to noise. No explicit mechanism is employed to produce stable representations under corrupted audio, leading to performance degradation in noisy environments. To address this, we propose VIB-AVSR, which integrates Variational Information Bottleneck layers at targeted positions within the LLM backbone to regularize representations. VIB-AVSR reduces degradation under noisy conditions across multiple SNR levels and noise types, without requiring architectural modifications or additional training data.

We present Proteus, a framework developed at Resemble AI for automated robustness testing of our audio deepfake detection system. Given a detector, Proteus systematically searches over sequences of everyday audio transformations (codec transcoding, additive noise, reverberation, dynamic-range compression, and VoIP simulation) to find combinations that fool the detector while preserving speech quality. We propose two complementary search strategies: (1) a breadth-first search that exhaustively maps augmentation effectiveness across the parameter space, and (2) a Q-learning agent designed to efficiently discover deeper attack chains by exploiting structural patterns in the BFS data. We report findings from continuous deployment of Proteus against our production detector, showing that specific augmentation chains can reliably flip detection verdicts while preserving speech intelligibility and speaker identity. We discuss how these findings are used to harden the detector through targeted retraining.

Recent advances in language--audio retrieval have been largely driven by contrastive dual-encoder architectures that align audio and text in a shared embedding space. While effective, existing retrieval embeddings are primarily optimized for audio--caption matching, limiting their ability to support diverse retrieval objectives and controllable retrieval behaviors. We present ALM2Vec, a universal audio embedding framework derived from pretrained large audio--language models (LALMs). By transferring the audio understanding, instruction-following, and reasoning capabilities acquired through large-scale multimodal training, ALM2Vec learns a unified embedding space for retrieval across audio domains and task types. Beyond conventional text--audio retrieval, ALM2Vec incorporates natural-language instructions into the embedding process, enabling instruction-aware retrieval for scenarios such as audio question answering and aspect-conditioned retrieval. Experimental results show that ALM2Vec achieves competitive performance on standard audio and speech retrieval benchmarks while exhibiting promising compositional and controllable retrieval capabilities, highlighting its potential as a unified audio embedding model for retrieval across domains, tasks, and user intents.

Long-form audio exhibits an inherent hierarchy: fine-grained events form sub-activities, which in turn constitute higher-level activities. Prior work often models these levels separately, leading to cross-level inconsistencies and requiring supervision at multiple levels. We formulate the problem as hierarchical parsing from event-level evidence: given detected event segments with class posteriors, we infer an order-consistent Act-Sub-Event parse tree. We propose Hierarchical Activity Grammar, encoding hierarchical composition and temporal-order constraints, and perform grammar-guided decoding that combines event evidence with a grammar prior. This yields a temporally grounded parse tree from which sub-activity segmentation and activity classification are derived, without requiring sub-activity or activity labels for training. Experiments on the long-form MultiAct audio dataset demonstrate improved temporal-order consistency (Edit score) and produces interpretable hierarchies.

Early detection of dementia through speech analysis offers a non-invasive screening alternative, but capturing both acoustic and linguistic biomarkers remains challenging. We propose a multimodal framework leveraging Whisper for dual-purpose extraction: acoustic representations from encoder outputs and transcripts via automatic speech recognition (ASR). For the acoustic pathway, temporal networks with attention pooling aggregate variable-length sequences into fixed-dimensional embeddings. For the linguistic pathway, we prompt a large language model (LLM) to extract interpretable features spanning lexical diversity, syntactic complexity, semantic coherence, and discourse patterns. A gated fusion network integrates both modalities. On ADReSS and ADReSSo, our method achieves F1-scores of 89.47% and 90.14%, demonstrating effective integration of acoustic and LLM-augmented linguistic features. Ablation shows that multimodal fusion consistently outperforms either modality alone.

Discrete audio representations have become increasingly popular for building multimodal text-audio systems and integrating audio capabilities into Large Language Models (LLMs). However, numerous studies report performance degradation on various downstream tasks due to information loss during discretization. To address this, we propose a novel approach combining temporally compressed discrete tokens with dimensionality-reduced continuous residuals. Our framework consists of a hybridized discrete-continuous focal modulation codec and a hybrid Transformer. This architecture performs autoregressive inference in the discrete domain, coupled with non-autoregressive prediction and continuous residual upsampling. Experimental results show that our approach significantly improves the retention of speaker characteristics compared to discrete-only methods, while simultaneously reducing the number of required autoregressive steps.

Learning discrete speech representations that preserve similarity across variable-length utterances is central to query-by-example spoken term detection (QbE-STD). While wav2tok introduced CTC-based sequence alignment to enforce token consistency, its tightly coupled clustering and alignment training recipe limits scalability. We propose wav2tok 2.0, a scalable alignment-aware speech tokenizer built on the BEST-STD backbone. wav2tok 2.0 employs staged training, first learning discriminative, speaker-invariant representations via contrastive learning and vector quantization, and then enforcing pairwise token consistency using a CTC alignment loss and a novel DTW-aligned framewise prediction objective with adaptive weighting. Experiments show that wav2tok 2.0 consistently outperforms BEST-STD and general-purpose tokenizers on QbE-STD while remaining efficient and scalable.

While pre-trained models excel in specialized tasks, learning universal representations across diverse acoustic domains remains challenging. To address this, we propose WQ-Fusion, a robust dual-encoder framework for cross-domain audio representation learning. Overcoming the limitations of static concatenation, WQ-Fusion integrates whisper and qwen via an Adaptive Feature Modulation module and a novel element-wise gated attention mechanism. This design enables dynamic feature selection, allowing the model to selectively emphasize relevant acoustic and semantic dimensions. Extensive experiments on the Interspeech 2026 Audio Encoder Capability Challenge (Track A) benchmark demonstrate that by effectively routing heterogeneous information, WQ-Fusion achieves a superior overall score of 0.836, significantly outperforming the strongest single-encoder baseline.

Soroll-IA is a weakly labeled environmental audio dataset recorded in a real-world industrial port environment in Valencia (Spain) using two fixed sensing nodes. The dataset comprises approximately 22 hours of audio segmented into 7,396 clips and covers 26 sound event classes representative of industrial port acoustic activity commonly observed in such environments, such as crane sirens, train movements, traffic, and other logistical and industrial sounds. Recordings were captured under highly challenging acoustic conditions, including strong background noise, long-distance sources, and frequent event overlap. All audio clips were annotated by domain experts following a weak labeling strategy, where tags indicate the presence of sound events within a clip without temporal localization. To account for inter-annotator variability, two ground-truth versions are released: one without cross-validation, where a class is considered present if annotated by at least one expert, and a second, more conservative version based on cross-validation, where agreement by at least two-thirds of the annotators is required. The dataset is intended to support research in audio tagging, weakly supervised sound event detection, and machine learning under realistic industrial acoustic conditions. Benchmark results are provided using two complementary architectures: CNN14 representing high-capacity convolutional models for audio tagging, and MobileNetV2, selected for its suitability in real-time classification on low-resource edge devices. To the best of current knowledge, Soroll-IA constitutes an available dataset dedicated exclusively to industrial port acoustic environments, aiming to foster advances in robust environmental sound analysis for safety-critical and operational monitoring applications. The dataset is available online and collected under Attribution-NonCommercial 4.0 International license.

Self-supervised learning (SSL) has emerged as an essential paradigm for music information retrieval (MIR). While current SSL models achieve state-of-the-art performance across various MIR tasks, they typically treat audio as 1D sequences, either operating on time-domain waveforms or on flattened time-frequency-domain spectrograms. This discards the rich spatial and structural information in time-frequency representations and overlooks a fundamental intuition in music production. In particular, music is naturally represented as time-frequency grids in MIDI-based workflows, a structure that tightly corresponds to 2D spectrograms and inherently makes many MIR tasks trivial. Motivated by this intuition, we propose PupuM2D, a musical Masked Modeling Duo (M2D) model that is trained directly on 2D music spectrograms. Instead of applying masked language modeling (MLM) to 1D sequences, PupuM2D learns robust representations by predicting the latent embeddings of masked 2D music spectrogram patches from unmasked contexts. To optimally adapt such a framework to music signals, we also apply domain-specific modifications to model architecture, training scheme, and inference paradigm, with comprehensive ablation studies showing their effectiveness. Evaluations on the MARBLEv2 benchmark show that PupuM2D outperforms the sequence-based SSL models across multiple MIR tasks in linear probing. Additionally, case studies of the attention maps also confirm that PupuM2D captures musically meaningful patterns within the 2D time-frequency domain. Codes and checkpoints are available at: https://www.yichenggu.com/PupuM2D/.

Recent Large Audio Language Models (LALMs) have achieved remarkable progress in audio perceptual tasks across individual acoustic layers, including speech, sound, and music. However, existing benchmarks predominantly evaluate these layers in isolation, overlooking the complex contextual relationships that arise when multiple acoustic sources co-occur in real-world auditory scenes. Real-world auditory interpretation requires Context-Aware Auditory Scene Understanding (CASU): the ability to comprehend the holistic scene by integrating sound layers. To evaluate this capability, we introduce the CASU benchmark, which assesses whether Audio LLMs can interpret auditory scenes composed of speech, acoustic events (e.g., announcements), and background environments (e.g., traffic), and reason about the logical relationships between these layers. We propose a scalable pipeline for constructing time-accurate, semi-synthetic audio streams by composing real-world scene sounds with synthetic speech. Building on this data, we design four tasks that probe scene understanding: contextual question answering, entity extraction from the scene, speaker role inference, and counterfactual reasoning where scene is manipulated. Experiments across multiple LALMs demonstrate that effective auditory scene understanding requires integration over all auditory layers, rather than reliance on speech or sound alone, underscoring the necessity of CASU for advancing complex audio understanding in LALMs.

Large Audio-Language Models (LALMs) have been widely used as judge models for the automatic evaluation of generated speech. However, prior approaches predominantly focus on holistic naturalness, leaving fine-grained paralinguistic distinctions underexplored. We introduce ParaPairAudioBench, a pairwise benchmark of 5,175 audio pairs across five paralinguistic dimensions: Style, Rate, Emphasis, Age, and Gender. Our experiments show that current LALM judges still lag behind human judgments by 32%p on average and exhibit severe calibration failures, particularly in Tie cases where the correct decision is to abstain. To further analyze lexical versus acoustic reliance, the benchmark includes both same-transcript and cross-transcript conditions. ParaPairAudioBench enables multi-dimensional, calibration-aware assessment of the reliability of LALM-as-a-Judge for paralinguistic speech evaluation.

Interactive music and live performance relies on real-time human expression, but modern generative music AI remains largely absent from this domain due to its prohibitive inference latency and offline rendering paradigm. To provide pioneer musicians with a novel medium for interactive composition, we should fundamentally change these static models into dynamic, playable instruments. In this paper, we propose a framework that bridges this gap. To achieve the low latency required for live interaction without sacrificing structural coherence, we formulate distillation within a streaming autoregressive latent space. Our approach gets rid of the need for expensive paired audio-latent datasets by utilizing prompt-only inputs to synthesize teacher-guided, chunk-wise trajectories on the fly. Because live instruments require high acoustic fidelity, we introduce music-aware consistency objectives, which combine latent, spectral, and temporal-difference losses, to preserve crucial qualities like timbre, transients, and rhythmic stability during accelerated single-step streaming generation. Implemented via parameter-efficient adaptation, our distillation reduces generation steps to achieve a low real-time factor. Crucially, by operating as a continuous autoregressive stream, the system can seamlessly assimilate dynamic human inputs on the fly, allowing users to instantly steer the musical trajectory without interrupting the audio flow. Ultimately, this work recontextualizes generative text-to-music models not as passive prompt-and-wait systems, but as responsive instruments, opening new frontiers for live human-AI musical co-creation.

This work investigates the interpretability of a Wav2Vec 2.0based speech intelligibility assessment model for oral and oropharyngeal cancer patients through canonical correlation analysis. By measuring the correlation between the model embeddings and eGeMAPS low-level descriptors (LLDs) as an interpretable reference, we analyze how acoustic information is encoded across the model layers. The analysis is conducted at two levels: individual LLDs layer-wise, and group-level: prosodic, spectral, and voice quality. Results show that the learned representations are most strongly correlated with spectral and prosodic features, with the first MFCC coefficient yielding the highest correlations across all layers. At the group level, spectral and prosodic groups achieve correlations of 0.77 and 0.71 respectively, while voice quality reaches 0.65. Beyond model interpretability, this work also offers practical guidance on acoustic feature selection for pathological speech assessment.

Music source restoration (MSR) requires jointly addressing source unmixing and the inversion of non-linear production effects. Current methods struggle to achieve accurate target signal reconstruction while maintaining semantic consistency. To address this limitation, we propose DTT-BSR+, a two-stage cascade MSR system that decouples distribution fitting from signal reconstruction into separate stages. A generative DTT-BSR separator in the first stage produces stems matching the prior of clean sources, and a modified Demucs network in the second stage enhances the first stage output using time-domain and multi-resolution spectral losses. DTT-BSR+ improves multi-mel signal-to-noise ratio (MMSNR) over the single-stage DTT-BSR across all stems, and surpasses the state-of-the-art X-LANCE MSR system on five stems. We also reveal through Fréchet Audio Distance (FAD) decomposition an implicit trade-off between signal reconstruction accuracy and semantic distribution fitting across stems.

Large audio-language models (LALMs) can reason about audio, yet it remains unclear whether they can perform comparative judgments between two speech signals along emotional, environmental, linguistic, prosodic, and interpersonal dimensions. We study this question in the context of speech emotion recognition (SER), where the model determines which utterance exhibits higher arousal, valence, or dominance. We introduce a reasoning-guided ordinal SER framework that conditions an LALM on paired speech inputs. The model is trained using reasoning traces generated from both semantic audio descriptions and acoustic evidence derived from GeMAPS features, enabling interpretable comparative decisions. Beyond direct supervision, we also employ direct preference optimization to encourage stronger separation for emotional differences. Experiments show that the proposed framework improves preference prediction while requiring only 5% of the training data used by conventional ordinal SER systems.

We study the limits of CTC-internal scoring for N-best hypothesis selection and locate the information bottleneck separating acoustic confidence from linguistic plausibility. Eleven CTC-internal and acoustic-feature scoring strategies produce no statistically significant WER improvement over greedy decoding on LibriSpeech dev-other at G=16 (all p > 0.05). The exhaustion is systematic: CTC's Spearman $ρ$ between hypothesis score and per-utterance WER degrades from -0.574 at G=4 to -0.270 at G=128, a 53% loss driven by blank-path proliferation. This establishes that the discriminative capacity of CTC-internal representations is saturated: no recombination of acoustic signals can close the oracle gap. Confirming that the bottleneck is linguistic, not acoustic, external linguistic information introduced via MBR decoding breaks through it. MBR-CER decoding with a RoBERTa pseudo-log-likelihood (PLL) posterior ($τ$=10, G=128) achieves 5.42% WER on held-out LibriSpeech test-other (greedy 5.96%, $Δ$=-0.535 pp, p<0.0001, 9.0% relative). RoBERTa PLL $ρ$ degrades only 21% over the same range, retaining discriminating power where CTC loses it. Applied without retuning across two Zipformer architectures, three domains (LibriSpeech, TED-LIUM 3, VoxPopuli), and four MUSAN noise levels, the recipe gives significant gains in 11 of 13 conditions. On the training side, standard MWER training via the CTC forward-backward algorithm implements Rao-Blackwellized REINFORCE at the output projection (variance about 3x below Viterbi). Yet sequence-level fine-tuning fails at near-converged checkpoints: all four MWER configurations on CR-CTC collapse (+6.18 to +8.90 pp WER), as a training oracle gap of 0.007 pp provides no usable reward signal.

Unifying speech, sound, and music generation in one model is hindered by tradeoffs between fidelity, end-to-end training, in-context conditioning, and variable-length synthesis that no current paradigm fully resolves. To address this challenge, we present AudioCALM, a universal audio generation framework that extends autoregressive (AR) next-token prediction from discrete tokens to continuous audio latents: a thin flow-matching head replaces the softmax to predict rectified-flow velocities at each position, and a block-causal AR-Flow attention pattern produces arbitrary-length output. Joint training of multiple audio generation tasks faces an asymmetric text--audio mismatch: speech transcripts align to specific time spans and demand tight, time-aligned attention, whereas sound and music captions describe only overall semantics and rely on diffuse, holistic attention; mixing the two disproportionately degrades sound and music generation. We address this asymmetry at two levels: a data reformulation strategy that unifies all three tasks under a single description-style conditioning interface, and a novel architecture Asymmetric Mixture-of-Modality-Experts (A-MoME), which adds a dedicated residual expert for speech while sound and music share the backbone, incurring no inference overhead on non-speech inputs. Experimental results demonstrate that AudioCALM matches modality-specific state-of-the-art and outperforms prior unified baselines on speech, sound, and music generation benchmarks.

Continuous Variational Autoencoders (VAEs) serve as the fundamental continuous tokenizer for modern neural audio generation systems, enabling high-fidelity reconstruction while providing a compact, smooth latent space for downstream generative priors. However, continuous VAEs face a fundamental conflict among compression rate, reconstruction fidelity, and latent space topology, which we formalize as the Rate-Distortion-Regularity Trilemma. This trilemma stems from a topological mismatch: the isotropic Gaussian prior in standard VAEs imposes a flat latent geometry that fails to accommodate audio's hierarchical nature, where low-frequency components are structured and compressible while high-frequency components are stochastic and incompressible, leading to disordered information packing in which crucial semantic features are interleaved with high-entropy noise. To address this challenge, we propose Structured Topology-Aware Regularization (STAR), a general training strategy that reshapes latent space geometry by imposing a growth-based constraint field, routing structural and textural information into channel subspaces with matching capacities. STAR is applicable to any VAE architecture and effectively resolves the trilemma, as demonstrated in CNN-based VAEs. We further present STAR-VAE, which combines STAR with a hybrid CNN-Mamba architecture for local feature extraction and linear-complexity global context modeling, and STAR-Gen, an LLM-based Flow Matching framework that leverages STAR-VAE's structured latent space for high-fidelity generation without vector quantization artifacts. Experiments across diverse audio domains show that STAR-VAE achieves state-of-the-art reconstruction fidelity and enhanced semantic information preservation, while the structured latent space improves both traditional diffusion models and STAR-Gen for text-to-audio generation.

Large automatic speech recognition (ASR) models such as Whisper must be deployed across hardware with widely varying memory and inference-speed constraints. We present a compression framework that jointly parametrizes Whisper deployment along \emph{six} dimensions: model size, temporal resolution, encoder token stride, low-rank adaptation capacity, weight precision and sparsity pattern. All axes are jointly optimized using NSGA-III with respect to three deployment objectives: word error rate (WER), inference FLOPs, and memory footprint. Across 50 of the 1,680 candidate configurations evaluated, we characterize the conditional effect of each axis and identify compression combinations that dominate naive single-axis scaling, while finding that 1:4 structured sparsity fails to recover acceptable accuracy under the tested recovery budgets. We report measured WER and resident memory, use analytical EffFLOPs as the search-time compute surrogate, and separately validate representative inference configurations using measured real-time factor (RTF).

We frame the system as diagnostic data curation for a large audio-language model: before fine-tuning, we probe Qwen3-Omni-30B-A3B-Instruct under normal, empty-audio, and shuffled-audio conditions to identify how the model's answers change when audio evidence is removed or mismatched. These model confusion patterns are used to bucket training samples into text-prior, shuffle-leak, strong audio-dependent, and hard or misleading cases. Our strongest train-only system fine-tunes only on strong-audio items, where the normal audio-question pair is correct but both counterfactual variants fail, plus a small number of empty-audio negatives and a text-only response normalizer for parse-failed generations. On the official development set, the best train-only system reaches 67.27% accuracy after response normalization, compared with 65.90% for our local Qwen3-Omni baseline. Final submissions additionally include models trained using train+development splits and a three-model ensemble.

Self-supervised learning (SSL) models have become a central component of modern speech processing systems, as they enable the learning of rich acoustic representations without reliance on labeled data. Despite their success on adult speech, it remains unclear how effectively these models capture speaker-related attributes such as age and gender in children's speech, which differs substantially from adult speech due to ongoing physiological and cognitive development. Higher pitch, increased articulatory variability, and age-dependent acoustic changes make children's speech a particularly challenging domain. In this work, we present a comprehensive analysis of how age and gender information is encoded across layers of four widely used SSL models: Wav2Vec2, HuBERT, Data2Vec, and WavLM. Layer-wise features are extracted and evaluated using a lightweight CNN on two benchmark children's speech corpora, PFSTAR and CMU Kids. To analyze feature compactness and redundancy, PCA is applied to identify redundancy and highlight the dimensions that contribute most to classification performance. Experimental results show that age- and gender-related information is unevenly distributed across SSL layers, with early to mid-level layers encoding the strongest paralinguistic cues. HuBERT achieves the best overall performance for age classification, while Wav2Vec2 and HuBERT lead gender classification on PFSTAR and CMU Kids, respectively. Beyond single-split evaluation, we further demonstrate that these findings remain stable under speaker-wise cross-validation, layer aggregation, and cross-database evaluation, indicating robustness to data imbalance and domain mismatch. Finally, we show that reliable age and gender classification is achievable even from short speech segments of 1--3 seconds.

We present InstructFX2FX, a system for sequential audio effect refinement through multi-turn natural-language instructions. Existing text-to-effect systems are largely single-shot, mapping one textual descriptor to one preset. Real audio engineering is instead sequential: engineers refine an existing effect chain through successive instructions. This poses a stateful problem that single-shot systems do not address: given the current effect parameters state and a new instruction, update the sound while preserving what earlier instructions already achieved. InstructFX2FX addresses this with a hybrid architecture that divides labor between a language model and CLAP-guided optimization. The LLM serves as a high-level planner that selects effects and proposes the initial parameter state, motivated by recent evidence that LLMs can outperform CLAP-based optimization for single-turn text-to-effect mapping; CLAP-guided optimization then refines the existing parameter state, providing a more stable and robust refinement mechanism than LLM reprompting. In the demo, attendees drive a dry recording through successive natural-language instructions: after each turn, they choose how strongly the effect is applied, then issue the next instruction based on what still differs from the sound they intend. In a preliminary evaluation on SocialFX-derived descriptor pairs, CLAP-guided refinement achieves lower DSP-feature MMD than an LLM+LLM initialize-then-reprompt baseline on 9 of 10 pairs. Trajectory analysis further shows that, for differentiable effects, optimization tends to gradually move the audio toward the new target while retaining the effects of the previous instruction, highlighting the potential for gradual refinement.

Recent speech research involves increasingly large datasets, complex models, and diverse experimental workflows. However, existing frameworks require substantial engineering effort to support such experiments. We present ESPnet3, a speech and audio research framework built on a modular system architecture with configuration-driven dataset composition and unified Python-based workflows. ESPnet3 introduces a DataOrganizer abstraction for flexible dataset integration and dataset sharding for memory-efficient large-scale training, while allowing recipe-specific logic through lightweight stage overrides. In OWSM pre-training experiments, ESPnet3 reduces per-epoch training time by \emph{21.1 minutes} compared to ESPnet2 and achieves \emph{>80% GPU utilization} in multi-node training. Fine-tuning experiments show that new models and datasets can be integrated with around \emph{46 lines of additional code}. ESPnet3 will be publicly released with model checkpoints and training logs.

In this paper, we introduce Synthetic Heart Sound Detection (SHAC), a task aimed at identifying phonocardiograms (PCGs) synthesized using neural audio codecs (NACs). To facilitate research in this direction, we release CARDIOFAKE, the first benchmark dataset for SHAC containing both real and codec-synthesized PCGs. We benchmark spectral representations (MFCC, LFCC) and self-supervised learning (SSL) representations (e.g., WavLM) for the task. Furthermore, we propose GROOT, a fusion framework that integrates spectral and SSL features for leveraging their complementary behavior. Experiments show that GROOT, combining MFCC and WavLM, achieves state-of-the-art performance, outperforming individual representations and competitive baselines.

Generating a coherent multi-shot video requires structured cross-shot memory. Subject appearance, scene context, and speaker identity must persist across cuts. Existing approaches either train end-to-end over fixed-length sequences and cannot scale, generate shot-by-shot with memory banks that grow linearly, or orchestrate pretrained generators under an LLM planner without a multi-shot-aware backbone. We present UnityShots, a memory-driven multi-shot audio-video generation system built on LTX-2.3, trained on annotated cinematic and music-video shots. The video stream maintains two fixed-size slots, a long-term memory (LTM) slot anchored to the opening shot and a short-term memory (STM) slot holding the immediately preceding tail, both updated at every cut by a boundary-conditioned gate that fuses visual cut probability and beat-tracker signals. The audio stream injects a reference speaker token at every shot to preserve vocal timbre without a sliding audio bank. A discrete cut-type prior, learned through AdaLN, becomes an inference-time control knob over transition strength. We release a benchmark of $200$ multi-cultural multi-shot sequences spanning six ethnic regions and ten or more languages, with per-shot reference identities, reference audio, and per-boundary transition labels. Evaluated across I2V, T2V, and R2V conditioning modes, UnityShots leads open-source baselines on every cross-shot coherence metric and matches the strongest closed-source system on the multi-shot axes.

Automatic Speech Recognition (ASR) systems, despite achieving remarkable accuracy in general-purpose domains with native speech (L1), struggle in domains like Air Traffic Control (ATC) due to strong channel noise, a presence of non-native (L2) English accents, and data scarcity. We propose a synthetic data generation pipeline with acoustical properties simulations specifically designed to address this lack of real data to improve recognition accuracy in the ATC domain. Our approach leverages a combination of neural generation techniques, including Text-to-Speech, Voice Conversion, L2-to-L1 accent conversion, and a novel controllable L1-to-L2 accent conversion framework built to simulate accented speech. Our experiments with the Whisper model on the ATCO2 corpus demonstrate that fine-tuning with either synthetic data alone, or a mix of real and synthetic data, significantly improves the word error rate over out-of-the-box and real data only baselines respectively.

This paper evaluates Reservoir Computing (RC) as an autonomous, 'feature-free' framework for audio processing, designed to eliminate traditional, handcrafted feature extraction stages. We investigate whether the high-dimensional temporal dynamics inherent in a reservoir can function as a robust end-to-end processor for the direct classification of raw acoustic signals. By bypassing computationally intensive representations like MFCCs, this approach seeks to mitigate significant intellectual and pre-processing bottlenecks in traditional signal pipelines. Our study evaluates and compares shallow, sequential, and parallel deep reservoir architectures to determine their capacity for hierarchical feature representation. Experimental results demonstrate that the proposed parallel approach consistently outperforms shallow and sequential baselines while maintaining low model complexity. These findings highlight the potential of RC as an efficient and scalable alternative for time-domain audio processing, offering a promising pathway toward deployable, low-power acoustic systems with minimal preprocessing requirements.

Dysarthria severity assessment is essential for therapy planning and longitudinal monitoring, yet manual perceptual rating is time-consuming and variable across clinicians. Although deep learning models achieve strong performance, their black-box nature limits clinical adoption. Existing speech explainability methods typically provide acoustic feature importance scores that are difficult for end-users to interpret. We propose an influence-based, instance-level explainability framework that explains each decision through supportive and competing training samples. Using gradient-based influence approximations, we compute per-utterance influence scores to identify supportive and competing training samples for each prediction. Controlled deletion experiments from 5 to 20 percent validate the explanations, showing that removing highly influential samples systematically shifts predictions. This approach provides auditable explanations by linking decisions to perceptible reference cases.

Differentiable audio processors are habitually designed and optimised in machine-learning frameworks, but deploying them as real-time audio effects still often requires non-automatic implementation in a dedicated digital signal processing language. The translation is error-prone, demands an onerous verification process, and detaches research prototypes from usable production tools. That being so, we present ADAC, a compiler that lowers a trained model to a framework-agnostic intermediate representation and emits efficient FAUST code whose impulse response matches the source model to within floating-point arithmetic noise, direct paths included. The optimisation loop is made audible by replacing the model in a running plugin after each gradient step. The exported processor carries a small set of macro-controls that leave its stability intact. A stability certificate computed from the shipped parameters is checked before the plugin is built. At the demonstration, a feedback delay network is trained and exported to a working plugin.

Current text-guided audio editing methods rely on paired training data, predefined operation templates, and separate processing pipelines across speech, music, and sound. We present Bagpiper-Edit to enable open-ended audio editing via free-form natural language instructions. We reformulate audio editing as a rich-caption rewriting task by treating a rich caption as the semantic representation of an audio clip. The user request is translated into an edited caption, which then guides Bagpiper-Edit to generate the target edited audio with the original audio as contextual acoustic anchor. This unlocks the potential of free-form editing, and circumvents the need for paired audio-editing training data, enabling powerful zero-shot editing capabilities. Evaluations across speech, audio, and free-form editing show Bagpiper-Edit maintains good consistency to the original audio and achieves similar performance to other expert models in most cases. Demo: https://bagpiper-edit.github.io, Codes: https://github.com/espnet/espnet/pull/6417 & https://github.com/HsunGong/espnet

Large Audio Language Models (LALMs) have demonstrated strong performance across a wide range of audio tasks. As they are increasingly deployed in real-world applications, ensuring their safety alignment has become more important. Although refusal mechanisms serve as a key safeguard by preventing LALMs from responding to harmful requests, they can also lead to {\em over-refusal}, where models incorrectly reject benign queries. This issue is especially challenging in the audio domain because speech that appears harmful in isolation may become benign when interpreted together with the surrounding acoustic context, such as background sounds. To study this problem, we introduce \textbf{AOR-Bench} (\textbf{A}udio \textbf{O}ver-\textbf{R}efusal \textbf{Bench}mark), the first benchmark for over-refusal specifically designed for LALMs. AOR-Bench contains 3,000 pseudo-harmful audio samples across six scenario categories. Evaluating 12 representative LALMs from six major model families, we find that over-refusal is widespread (Figure~\ref{fig:overall_performance}) and uncover several important patterns in their safety judgments. As a preliminary effort to mitigate this issue, we further explore two lightweight strategies (e.g., Chain-of-Thought and activation steering) to reduce over-refusal.

Deep learning-based audio classification systems, including automatic speaker verification, are vulnerable to adversarial attacks. Realistic real-time threat assessment remains difficult because optimization-based methods, such as projected gradient descent (PGD) and Carlini-Wagner, require costly iterative updates in the high-dimensional waveform domain. Generative attacks allow single-shot synthesis but often introduce perceptible artifacts or depend on computationally intensive architectures, while diffusion and autoregressive approaches incur high inference latency. To address this gap, we propose a generative attack framework operating in the continuous latent space of a neural audio codec. A conditional generator synthesizes class-specific perturbations in a single forward pass and decodes them into adversarial waveforms. Our method achieves targeted attack success rates up to 99% with sub-7 ms inference, outperforming generative baselines while reducing latency by 24x.

Acoustic environments often contain multiple overlapping sound events, and the same acoustic scene can be described using diverse textual expressions, making audio-text alignment inherently ambiguous. This paper proposes a probabilistic audio-language pretraining framework to model many-to-many correspondence ambiguity in audio-text alignment. Unlike conventional contrastive methods that learn deterministic point embeddings, our approach represents each modality as a distribution and learns uncertainty-aware cross-modal alignment. Rather than relying on masking-based uncertainty simulation, we mix audio-text pairs to create overlapping sounds that better reflect real acoustic mixtures and capture semantic inclusion relations among sound events. We further introduce a multi-level inclusion loss to enforce representations consistent with these relations. Experiments on audio-text retrieval benchmarks show that the proposed method outperforms deterministic baselines.

Audio editing aims to modify specific content in an existing audio clip according to a text instruction or description while preserving the remaining acoustic content. Despite the remarkable progress of diffusion models, existing training-based editing methods mainly rely on the local inductive biases and cross-attention interaction in convolutional U-Net backbones, which often hinder long-range semantic alignment and precise understanding and localization of instructions. In contrast, diffusion transformers provide stronger global modeling and multimodal fusion, but existing editing architectures usually adopt a simple stack of diffusion transformer blocks. Applying joint attention over concatenated audio and text tokens in all blocks results in quadratic complexity with respect to token length. To balance editing performance and efficiency, we propose a novel instruction-guided audio editing framework based on rectified flow matching (RFM), named RFM-Editing 2, built on a hybrid two-stage diffusion transformer. The proposed model performs joint attention over audio and text tokens to establish coarse semantic alignment at the low-resolution stage, then switches to alternating joint-attention and cross-attention blocks to refine editing details at the high-resolution stage. This coarse-to-fine strategy enables efficient and accurate instruction-guided audio editing. Experiments show that the proposed framework achieves notable performance gains on challenging editing tasks involving overlapping audio events and complex instructions, while substantially improving editing efficiency.

The challenge associated with recognizing dysarthric speech primarily arises from pronounced acoustic variability attributed to impaired articulatory precision. Past research has demonstrated improved recognition through the use of hybrid DNN/HMM sequence discriminative training. This paper presents a comprehensive investigation of various combinations of acoustic features tailored to different Acoustic Models, offering suitable feature selections for each. The incorporation of Pitch features notably improved recognition performance, especially for sentence recognition tasks involving dysarthric speech. Through a systematic examination of the TORGO database, we have demonstrated the potential to enhance the performance of the state-of-the-art Factorized Time Delay Neural Network (F-TDNN) model for recognizing dysarthric speech. Our methods, implemented with the F-TDNN model, resulted in a 4.65% relative improvement in isolated word recognition and a 4.63% relative improvement in sentence recognition for dysarthric speech, compared to previous research. This improvement effectively compensates for speech variability, attributable to our deliberate selection of the number of overlapping frames between consecutive training example chunks.

AudioLLMs enable speech recognition conditioned on textual prompts such as domain descriptions or entity lists. However, it remains unclear whether these models genuinely utilise such context or rely on parametric knowledge learned during pretraining. Existing benchmarks cannot answer this question because they evaluate transcription under fixed prompting conditions and rarely include explicit contextual inputs. We introduce IndicContextEval, a 56-hour multilingual benchmark of natural speech from 555 speakers across 8 Indian languages and 23 professional domains. We design a 7-level prompting framework that progressively introduces contextual signals, including metadata, natural-language descriptions, entity lists in English and native script, and adversarial prompts with incorrect entities. Evaluating five models reveals substantial differences in context utilisation behaviour, highlighting the need for explicit evaluation of contextual grounding in AudioLLMs.

While Audio Large Language Models (Audio LLMs) excel at multimodal understanding, they suffer from text dominance, a bias where models blindly favor text over acoustic evidence, causing hallucinations. However, the internal mechanisms underlying how these models behave when audio and textual inputs contradict each other remain unexplored. In this work, we present the first mechanistic analysis of this phenomenon by tracing the propagation of internal representations across layers. Our investigation reveals three key findings: (i) text dominance is systematically and empirically across models; (ii) while text and audio rely on functionally distinct pathways, they ultimately converge into a shared semantic space in late layers; and (iii) the text pathway does not erase audio information, but rather actively suppresses intact audio representations. Building on these insights, we leverage back-patching, a training-free intervention that routes late-layer audio activations back into earlier layers. This amplifies the audio representations, enabling them to overcome textual suppression. Our evaluation shows that back-patching consistently reduces text dominance, paving the way for mechanistic multimodal alignment under conflict.

The advancement of speech generation technologies has made artificial speech increasingly realistic. Although modern classification models can achieve high accuracy when it comes to deepfake detection, they do not produce evidences such as indicating where spoof cues appear in the spectrogram and what they imply acoustically, limiting their usefulness in forensic settings. Manual analysis of full spectrograms is resource-intensive, so evidence should narrow attention to the most diagnostic regions. Moreover, existing explainability methods have limited capabilities in connecting contextual attributes to localized evidence, making explanations harder to verify. To overcome this limitation, we propose GRIDEX, a pipeline that, when given a deepfake spectrogram, generates forensic explanations of its anomalies. The pipeline (i) selects top-K anomalous regions in the spectrogram and (ii) produces an explanation for each anomaly. The explanations follow a schema of categorical acoustic fields, including temporal, spectral, phonetic information and interpretation text. To our knowledge, this is the first framework to generate structured forensic explanations using regional grounding for deepfake spectrograms. GRIDEX is trained with a two-stage learning paradigm that combines supervised fine-tuning (SFT) with Group Relative Policy Optimization (GRPO). Experiments on our dataset show improved artifact localization and explanation quality over strong vision-language model (VLM) baselines. The dataset and code will be released upon publication.

Efficient processing of continuous audio streams remains a key challenge for real-time and resource-constrained systems. This paper introduces a neuromorphic trigger for audio event detection, based on a spiking neural network (SNN) that selectively gates input to downstream models. The proposed neuromorphic trigger acts as a flexible low-cost front-end, identifying salient audio segments and enabling these to be processed by a more computationally intensive model for tasks such as classification. The trigger is implemented as a lightweight fully connected SNN using a close-open filter for postprocessing, and is evaluated on two representative tasks: Anomalous Sound Detection (ASD) and Sound Event Detection (SED). For ASD, the trigger achieves a one-second segment-based F1 score of 0.97 on a class-agnostic form of the URBAN-SED dataset, demonstrating high reliability in identifying relevant audio regions. For SED, the trigger is combined with the Dang classifier on the DCASE 2017 Challenge Task 2 dataset, showing a potential $42.6\times$ reduction in FLOPs while reducing the lower bound of the event-based error rate from 0.41 to 0.25. These results highlight the potential of neuromorphic triggers as real-time, energy-efficient front-end filters, enabling substantial reductions in computational cost.

Large Audio Language Models (LALMs) achieve strong performance on a variety of audio understanding tasks but continue to struggle with temporal reasoning, a fundamental capability central to human auditory perception. Understanding the causes of these failures remains challenging as existing benchmarks report performance gaps without probing underlying mechanisms. To address this, we introduce a benchmark with 1,657 questions across three foundational tasks designed specifically for mechanistic analysis. Examining model outputs across varying input settings (behavioral analysis) reveals that models often under-utilize audio when textual cues are available. We also provide the first causal mechanistic analysis of temporal reasoning failures in LALMs. Comparing attention upweighting against scaling, we find that redistributing attention across audio tokens is more effective than increasing audio attention. Targeting task-relevant tokens yields further gains. These findings suggest that modality imbalance alone cannot explain failures. Attention scaling at bottleneck layers improves accuracy from 55.9% to 59.1% without fine-tuning, demonstrating a promising direction for future work.

Text-to-audio (TTA) generation, synthesizing audio from natural language, has been widely studied for its ability to capture precise user intent. To effectively advance TTA models, it is essential to reliably evaluate generated audio without relying on costly human subjective ratings, motivating the development of automatic evaluation metrics that correlate well with human judgments. While recent CLAP-based metrics provide practical reference-free solutions, their coarse-grained text-audio similarity matching often correlates poorly with human ratings. To address this, we propose ELSA, a reference-free evaluation metric for fine-grained text-audio alignment. ELSA decomposes generated audio guided by distinct acoustic events derived from the text query and assesses event-level alignment. Experiments across four TTA benchmarks show that ELSA reveals a higher correlation with human subjective ratings than prior metrics, highlighting its effectiveness for reliable TTA evaluation.

Long-form serialized audio drama, with arcs that run for 200 to 800 episodes, is a major creative medium and a setting where frontier large language models (LLMs) fail. We benchmark 21 models, spanning classical, fine-tuned, open-frontier, closed-frontier, and reasoning tiers, on a uniform set of structural narrative metrics. All closed-frontier systems saturate at a plot-beat F1 in the band [0.78, 0.81] and collapse by about -0.20 F1 at horizon h=200. We introduce NarrativeWorldBench, an open benchmark of nine narrative-structure metrics evaluated across horizons h in {10, 20, 50, 100, 200}, with cross-lingual evaluation across four Indic languages (Hindi, Tamil, Telugu, Marathi). We introduce N-VSSM, a Narrative Variational State-Space Model that maintains a structured 256-dimensional latent world state over more than 200 episodes via a Mamba-2 backbone with an event-conditioned posterior and an 8B decoder. N-VSSM holds plot-beat F1 >= 0.84 across all horizons at 4x lower compute than the closed-frontier band. A learned Cultural Transfer Function lifts cross-language fidelity by +0.20 to +0.23 Likert points. In a within-subjects writer study (n = 12 professional authors, 240 trials), N-VSSM is preferred over Claude Opus 4.5 on long-arc consistency 71% of the time and rated +1.3 Likert points higher on controllability.

This work investigates modelling strategies in continuous and discrete latent spaces in the vector quantisation (VQ)-based neural audio codec (NAC) speech enhancement (SE), along with the role of VQ regularisation. We propose cNAC-SE and dNAC-SE frameworks that predict continuous representations and discrete tokens in latent space, respectively. Theoretical analysis and visualisations in latent space are performed to exhibit their inherent modelling mechanisms. Experimental results show that the fully fine-tuned cNAC-SE model consistently outperforms all dNAC-SE variants across diverse test conditions and achieves leading performance among established generative approaches in DNS-MOS metrics. Comparison with the discriminative counterpart shows that VQ enhances robustness through an intrinsic effect of clean-prior-constrained regularisation, independent of discrete token processing. This highlights the transferable value of VQ regularisation to other continuous modelling methods.

With the growing focus on audio in multimedia applications, numerous advanced works on audio generation have emerged. Existing studies typically treat text-to-audio (TTA) and other related audio generation tasks, such as instruction-based audio editing, as independent challenges, adopting task-specific architectures or modules. This absence of a unified modeling paradigm substantially increases the overhead and complexity of building a system for both audio generation and editing, while also leading to limited scalability. To address this issue, we introduce AudioWeave, a unified model for TTA and audio editing without additional task-specific components. Specifically, we propose a joint condition modeling approach with a factorized position embedding, enabling the diffusion transformer backbone to operate under heterogeneous inputs of TTA and audio editing. We further propose a progressive multistage training strategy to mitigate task competition and catastrophic forgetting caused by interference among multiple tasks. This in turn helps maintain the performance of each individual task and may even lead to improvements in certain aspects. Experimental results on TTA task and six audio editing tasks show that our unified model achieves competitive performance with task-specific models, laying a groundwork for further exploration of unified audio generation models.

Sperm-whale codas are conventionally characterized by click count and inter-click intervals (ICIs), leaving recurring differences in constituent click waveforms unresolved. This study tests whether acoustic organization is nested across two scales: within codas, where recurring click-waveform differences may complement ICI timing, and across codas, where recurring whole-coda forms may themselves carry sequence dependence. Candidate recurring click and whole-coda groupings were identified from 1,483 codas without prespecifying waveform categories, then evaluated with native-rate spectral/envelope measurements, exact nuisance matching, held-out timing contrasts, and sequence controls. At the first tier, recurring click-waveform groups differed in spectral slope, bandwidth, flatness, high/low-band energy, and envelope structure within matched date, social unit, individual, and sample-rate strata. Their composition added information about whole-coda grouping beyond timing, while timing remained informative when click composition was fixed. The richer description also carried held-out social-unit-associated information beyond timing. At the second tier, direct native-rate waveform summaries recovered the recurring whole-coda forms well above context-preserving nulls, whereas conventional timing did not; the forms also cross-cut published timing-defined coda types. The preceding two-coda context then added held-out predictive information beyond the immediately preceding coda, while a third preceding coda provided no reliable further gain. Together, these results support two-tier acoustic organization: recurring waveform differences and ICI timing jointly organize individual codas, and acoustically grounded whole-coda forms in turn show bounded second-order predictive dependence across sequences.

Audio-Language Models (ALMs) have shown remarkable success in zero-shot audio classification by aligning audio waveforms with text. Recent efforts to improve downstream performance focus on learning optimal text prompts. However, previous approaches focus on the text encoder, leaving the potential of learnable prompts within the audio encoder unexplored. In this paper, we propose a novel framework that introduces trainable prompts into the audio encoder to capture task-specific acoustic features. We demonstrate that integrating audio-side prompt learning with existing text-side approaches enhances few-shot adaptation. Through extensive experiments across 11 datasets show that integrating our method as a plug-and-play module alongside existing text prompt tuning generally leads to performance improvements. These findings suggest that explicitly modulating the audio representation space effectively complements text-only prompting approaches. The code is available at https://github.com/hyebin-c/aspl.