arXiv:2609.09263v1 Announce Type: cross Abstract: Speech-to-speech (S2S) models now run inside dubbing, translation, and voice agents. Unlike text models, they hear the speaker's voice, which carries the speaker's gender. A faithful system should treat a speaker as who they sound like, not as whoever usually says what they said. Testing this is harder than it looks, since most S2S models answer in a single, fixed output voice, hard-coded so it cannot drift toward a stereotype. Checking the output voice comes back clean even when the model is biased. We therefore ask two questions. When a model re-speaks the input, does the stereotype in the words shift the perceived gender of the output voice (voice rendering)? And when the model states the speaker's gender, does it follow the voice or the content (gender attribution)? We answer both with one controlled experiment crossing male and female voices with masculine-, neutral-, and feminine-stereotyped passages, on five open- and closed-source models in English, Spanish, and Mandarin. The rendered voice shows no stereotype drift. But every model decides the speaker's gender from the content, not the voice. Making the content one step more feminine (masculine -> neutral -> feminine) multiplies the odds of a "female" judgment by 1.7-24. When the content clashes with the voice, the worst model misgenders the speaker in 90% of cases. When they agree, it misgenders in only 2%. The bias thus hides in gender attribution, where fixed-voice evaluation cannot see, and where audits must look as S2S systems increasingly speak for real people.

arXiv:2609.10058v1 Announce Type: cross Abstract: Robust speech-to-text translation systems should perform reliably across diverse acoustic conditions, yet practical pipelines lack controllable tools for systematic environment exploration. Large speech models remain sensitive to unseen acoustic conditions, as training data rarely cover the full range of real environments.We present NOPEHYPE, a structured training workflow that combines a controllable environment simulator, coverage-optimal environment reduction on Power Spectral Density (PSD) templates, and a small, interpretable hyperparameter search over simulator knobs. We show that simulator-generated noise achieves performance comparable to balanced realnoise training across Whisper and SeamlessM4T models, provide principled environment prototype sets, and identify practical default simulator configurations from a structured 27-run hyperparameter sweep.

arXiv:2605.13841v3 Announce Type: replace-cross Abstract: Voice agents are increasingly deployed across enterprise applications. However, no existing benchmark jointly addresses realistic conversation simulation and comprehensive voice-specific evaluation. We present EVA-Bench, an end-to-end evaluation framework that addresses both. On the simulation side, EVA-Bench orchestrates dynamic bot-to-bot audio conversations with automatic simulation validation that detects user simulator error and appropriately regenerates conversations before scoring. On the measurement side, EVA-Bench introduces two composite metrics: EVA-A (Accuracy) and EVA-X (Experience). EVA-Bench includes 213 scenarios across three enterprise domains, a controlled perturbation suite for accent and noise robustness, and multi-trial measurements that distinguish peak from reliable capability. Across 12 systems spanning all three architectures, we find: (1) no system simultaneously exceeds 0.5 on both EVA-A pass@1 and EVA-X pass@1; (2) peak and reliable performance diverge substantially (median pass@k--pass^k gap of 0.44 on EVA-A); and (3) accent and noise perturbations expose substantial robustness gaps, with effects varying across architectures, systems, and metrics (mean $\Delta$ up to 0.314). We release EVA-Bench under an open-source license.

arXiv:2609.04867v2 Announce Type: replace-cross Abstract: Text-to-audio-video (T2AV) generation has advanced rapidly, but its evaluation still underestimates the audio modality. Existing benchmarks either treat audio as an auxiliary component of video quality or assess it in isolation from audiovisual grounding, making it difficult to diagnose where current systems truly succeed or fail in audio generation. We present PRISM-Bench, the first audio-centric diagnostic benchmark for T2AV generation. Built from a rigorously curated dataset of 900 human-verified samples, PRISM-Bench factorizes audio evaluation along two orthogonal axes: audio type (Speech, Music, and Sound) and sound-source visibility (On-screen vs. Off-screen). It evaluates generated content across four perceptual dimensions (Audio-Visual Coherence, Audio Quality, Audio Expressiveness, and Prompt Following) with 35 fine-grained criteria. To ensure reliable assessment, we adopt an enhanced MLLM-as-a-Judge protocol based on blind, side-by-side comparison against ground-truth references, demonstrating strong alignment (over 70% mean agreement) with human raters. Our evaluation of recent T2AV systems highlights a significant performance gap between frontier and open-source models. Furthermore, we demonstrate that current generation paradigms overfit to perceptual fidelity while struggling with complex grounding and control tasks, particularly in generating music and synchronized On-screen audio.

arXiv:2609.08977v2 Announce Type: replace-cross Abstract: In this work, we present Gander, an end-to-end model that unifies omni perception, realtime interaction, and agentic capabilities within a single framework. In contrast to turn-based conventional paradigms, Gander continuously receives streaming inputs across multiple modalities, including video, speech, and text, enabling natural full-duplex interaction in both everyday conversations and complex workflow-oriented agent scenarios. Users can interrupt the model at any time, while the model can also proactively provide intermediate feedback or ask follow up questions. To natively support these capabilities, Gander adopts two key architectural designs: 1) It employs a Cerebellum-Brain collaborative framework, in which the Cerebellum is responsible for realtime interaction and omni conversational capabilities, while the Brain handles complex reasoning and higher-level agentic tasks. The two components interact continuously through tool calling and the agent orchestration runtime. 2) The Cerebellum is built upon a streaming Thinker-Talker architecture, user inputs and model outputs are further flattened into an ordered token stream at the chunk level, providing a unified representation for low latency, continuous interaction. We conduct comprehensive evaluations of Gander across four dimensions: conversational ability, omni understanding, interactive capability, and agentic intelligence. Internal human evaluations demonstrate that Gander maintains the natural and expressive spoken dialogue capabilities of SOTA open source models while achieving competitive performance in omni interaction. Gander also demonstrates robustness in challenging real-world scenarios, including background noise interference, multi-party interactions, and backchannel communication. We release Gander together with its models, code, and data to facilitate further research and development in the community.

arXiv:2605.24652v1 Announce Type: new Abstract: Rapid advances in audio-video (AV) generation have enabled high-fidelity synthesis with synchronized sound, particularly for human-related scenarios involving speech and interactions. Yet evaluation for AV generation remains at an early stage, with only a few coarse-grained benchmarks for human-related scenarios and relying on limited preset evaluations with generic multimodal LLMs, leading to inaccurate assessments of model capabilities. To address these issues, we introduce AVBench, a fully automated benchmark tailored for human-centric AV generation. AVBench is built on two key designs for comprehensive and accurate evaluation: (i) Human-centric and fine-grained metrics. AVBench integrates ten evaluation dimensions designed for human-centered real-world scenarios, covering visual quality, audio quality, and multi-level consistency across modalities. These practical metrics capture human-related details that existing benchmarks often overlook. (ii) Specialized evaluators via preference learning. To address the lack of specialized training data, we construct large-scale supervision by transforming real-world videos into diverse training pairs with controlled perturbations. After fine-tuning on this high-quality dataset, the evaluators learn to reliably detect subtle cross-modal inconsistencies. Crucially, instead of producing discrete textual judgment, AVBench derives continuous evaluation scores from the model's prediction confidence on binary decisions. This probabilistic scoring mechanism enables a more reliable assessment than traditional VQA-style evaluation and aligns closely with human judgment. Taken together, AVBench offers automated evaluation for AV generation, demonstrates strong potential for data filtering, and serves as a differentiable reward signal for Reinforcement Learning from Human Feedback (RLHF).

arXiv:2605.24678v1 Announce Type: new Abstract: Speech and language technologies offer valuable opportunities for supporting mental health assessment through objective and interpretable cues. We present a systematic feature-based analysis framework leveraging perceptually grounded acoustic and linguistic characteristics, including prosody, vocal quality, semantic coherence, syntactic structure, and sarcasm. Using statistical analysis and interpretable machine learning (XGBoost with SHAP and LIME), we examine associations between speech features and validated symptom measures of depression, anxiety, and ADHD. Evaluated on both controlled benchmark datasets (StressID, DAIC-WOZ, Androids, EATD) and a real-world clinical dataset, the framework reveals stable and consistent relationships between symptom severity and vocal irregularities (e.g., shimmer, jitter), lexical-syntactic patterns, and affective tone. An ablation study conducted across all datasets further identifies the most informative feature groups. This work explores a transparent and clinically interpretable approach to speech-based mental health analysis.

arXiv:2605.23912v1 Announce Type: cross Abstract: We present Raon-Speech, a top-performing 9B-parameter speech language model (SpeechLM) for English and Korean speech understanding, answering, and generation, and Raon-SpeechChat, a high-performing full-duplex extension for natural real-time conversation. Raon-Speech successfully transforms a pre-trained LLM into a SpeechLM that both understands and generates speech while preserving strong text capabilities. It trains on 1.38M hours of highly curated English and Korean speech and text datasets with the following training stages: (1) speech modules alignment, (2) end-to-end SpeechLM pre-training with knowledge distillation, and (3) multi-task preference optimization-based post-training. Across 42 English and Korean speech and text benchmarks, Raon-Speech establishes the strongest overall profile on speech-centric tasks in our comparison against eight similarly sized recent audio foundation models, including Qwen2.5-Omni and Fun-Audio-Chat, while preserving strong text question answering performance. Building upon it, Raon-SpeechChat enables natural full-duplex conversation by continual training on 119K hours of time-aligned real and synthetic dialogue data. It proceeds through three complementary training stages: (1) causal encoder adaptation, (2) full-duplex pre-training, (3) full-duplex fine-tuning for voice and role-control. On multiple full-duplex benchmarks, Raon-SpeechChat shows its clearest strengths on the turn-taking and interruption-sensitive behaviors covered by FDB v1.0, and remains competitive across the broader full-duplex evaluation suite. We open-source all model checkpoints, the training and inference pipeline, and an interactive demo.

arXiv:2605.23954v1 Announce Type: cross Abstract: Audio Large Language Models (ALLMs) are highly vulnerable to real-world noise, which often induces severe semantic drift and hallucinations. Existing robustness methods primarily rely on waveform-level acoustic enhancement, answer-level supervision, or the internal suppression of noise representations. To address these issues, we propose echodistill, an alignment-based noisy-to-clean self-distillation framework. Echodistill leverages a frozen clean-audio teacher to provide semantic references for an inference-time noisy-audio student. Specifically, the student samples candidate responses under noisy conditions to expose its test-time behavior. These trajectories are then optimized via group-relative policy optimization (GRPO), where the token-level consistency with the teacher acts as a reward bonus. By aligning the noisy student's candidate responses with clean semantic evidence, and applying audio-aware reward shaping, our method encourages reasoning trajectories that are both correct and genuinely acoustically grounded. Echodistill significantly improves the semantic reliability and task performance of Audio LLMs under complex noise, without introducing any additional inference costs. Extensive experiments show that: (I) Compared with the strongest baseline, echodistill achieves average improvements of 4.18\%$\uparrow$ in GSR under strong noise. (II) Ablation results on Qwen-Omni further show that echodistill improves over the GRPO-only variant by 3.02\%$\uparrow$ in Acc, 3.89\%$\uparrow$ in Noisy, and 4.53\%$\uparrow$ in GSR on average. Our codes are available at https://anonymous.4open.science/r/echodistill-10DE.

arXiv:2605.24806v1 Announce Type: cross Abstract: Large audio and language models have recently demonstrated zero-shot reasoning capabilities across various domains. However, it remains unclear how the form of audio input, whether handcrafted acoustic features extracted from speech or the raw audio waveform itself, affects performance for Parkinson's disease (PD) detection across different languages. In this study, we systematically compare two input modalities for zero-shot PD detection: (i) handcrafted acoustic features extracted from speech recordings analyzed by a general-purpose LLM, and (ii) direct waveform input analyzed by audio-capable models. Experiments on PD speech datasets in four languages show that performance varies across input modalities, speech tasks, and languages. Handcrafted acoustic features provide more stable performance in a low-resource language (e.g., Bengali), whereas audio input yields dataset-dependent gains. These findings highlight the impact of input modality on zero-shot PD detection from speech.

arXiv:2605.25962v1 Announce Type: cross Abstract: Machine unlearning removes designated concepts or knowledge from pre-trained models. Recent work has extended this paradigm to speaker identity unlearning in zero-shot text-to-speech (ZS-TTS), the task of selectively erasing a model's ability to replicate a speaker's voice. Existing methods, however, quietly assume all unlearning requests arrive at once; an unrealistic assumption, since privacy-motivated removals arrive sequentially over time. We show this assumption breaks state-of-the-art methods: unlearning each new speaker fully revives previously unlearned speakers, reintroducing the very privacy risk unlearning was meant to eliminate. We present Cumulative ORThogonal Identity Suppression (CORTIS), the first framework for continual speaker identity unlearning in ZS-TTS that requires no access to previously-unlearned speaker data. CORTIS combines Fisher-information-based parameter masking, which localizes updates to speaker-relevant weights, with orthogonal projection against subspaces spanned by prior unlearning updates. With VoiceBox, CORTIS unlearns each requested speaker while keeping previously unlearned speakers forgotten across long request sequences, substantially outperforming sequential application of prior methods. The demo is available at https://cumulativeortis.github.io/ .

arXiv:2510.02171v3 Announce Type: replace-cross Abstract: Music performance is a distinctly human activity, intrinsically linked to the performer's ability to convey, evoke, or express emotion. Machines cannot perform music in the human sense; they can produce, reproduce, execute, or synthesize music, but they lack the capacity for affective or emotional experience. As such, music performance is an ideal candidate through which to explore aspects of collaboration between humans and machines. In this paper, we introduce the witheFlow system, designed to enhance real-time music performance by automatically modulating audio effects based on features extracted from both biosignals and the audio itself. The system, currently in a proof-of-concept phase, is designed to be lightweight, able to run locally on a laptop, and is open-source given the availability of a compatible Digital Audio Workstation and sensors.

arXiv:2601.21463v3 Announce Type: replace-cross Abstract: Existing speech editing detection (SED) datasets are predominantly constructed using manual splicing or limited editing operations, resulting in restricted diversity and poor coverage of realistic editing scenarios. Meanwhile, current SED methods rely heavily on frame-level supervision to detect observable acoustic anomalies, which fundamentally limits their ability to handle deletion-type edits, where the manipulated content is entirely absent from the signal. To address these challenges, we present a unified framework that bridges speech editing detection and content localization through a generative formulation based on Audio Large Language Models (Audio LLMs). We first introduce AiEdit, https://huggingface.co/datasets/JunXueTech/AiEdit, a large-scale bilingual dataset (approximately 140 hours) that covers addition, deletion, and modification operations using state-of-the-art end-to-end speech editing systems, providing a more realistic benchmark for modern threats. Building upon this, we reformulate SED as a structured text generation task, enabling joint reasoning over edit type identification, and content localization. To enhance the grounding of generative models in acoustic evidence, we propose a prior-enhanced prompting strategy that injects word-level probabilistic cues derived from a frame-level detector. Furthermore, we introduce an acoustic consistency-aware loss that explicitly enforces the separation between normal and anomalous acoustic representations in the latent space. Experimental results demonstrate that the proposed approach consistently outperforms existing methods across both detection and localization tasks.

arXiv:2602.18527v2 Announce Type: replace-cross Abstract: Current audio-visual large language models (AV-LLMs) are predominantly restricted to 2D perception, relying on RGB video and monaural audio. This design choice introduces a fundamental dimensionality mismatch that precludes reliable source localization and spatial reasoning in complex 3D environments. We address this limitation by presenting JAEGER, a framework that extends AV-LLMs to 3D space, to enable joint spatial grounding and reasoning through the integration of RGB-D observations and multi-channel first-order ambisonics. A core contribution of our work is the neural intensity vector (Neural IV), a learned spatial audio representation that encodes robust directional cues to enhance direction-of-arrival estimation, even in adverse acoustic scenarios with overlapping sources. To facilitate large-scale training and systematic evaluation, we propose SpatialSceneQA, a benchmark of 61k instruction-tuning samples curated from simulated physical environments. Extensive experiments demonstrate that our approach consistently surpasses 2D-centric baselines across diverse spatial perception and reasoning tasks, underscoring the necessity of explicit 3D modelling for advancing AI in physical environments. Our source code, pre-trained model checkpoints, and datasets are available at https://github.com/liuzhan22/JAEGER.

arXiv:2605.04700v2 Announce Type: replace-cross Abstract: Jailbreak attacks on audio language models (ALMs) optimize audio perturbations to elicit unsafe generations, and they typically update the entire waveform densely throughout optimization. In this work, we investigate the necessity of such dense optimization by analyzing the structure of token-aligned gradients in ALMs. We find that gradient energy is highly non-uniform across audio tokens, indicating that only a small subset of token-aligned audio regions dominates the optimization signal. Motivated by this observation, we propose Token-Aware Gradient Optimization (TAGO), which enables sparse jailbreak optimization by retaining only waveform gradients aligned with audio tokens that have high gradient energy, while masking the remaining gradients at each iteration. Across three ALMs, TAGO outperforms baselines, and substantial sparsification preserves strong attack success rates (e.g. on Qwen3-Omni, $\mathrm{ASR}_{l}$ remains at 86% with a token retention ratio of 0.25, compared to 87% with full token retention). These results demonstrate that dense waveform updates are largely redundant, and we advocate that future audio jailbreak and safety alignment research should further leverage this heterogeneous token-level gradient structure.

arXiv:2605.18916v2 Announce Type: replace-cross Abstract: We investigate Counterfactual Video Foley Generation, which aims to adopt a sound-source identity that contradicts the visual evidence while remaining temporally synchronized to a silent video. Existing Video&Text-to-Audio (VT2A) models struggle with this, often remaining anchored to the visually implied sound source when video and text contents disagree. We present ConterFlow, an inference-time dual-phase sampling scheme for pretrained flow-matching VT2A models. Phase 1 builds a video-derived temporal structure while suppressing the visually implied source; Phase 2 drops video conditioning to focus entirely on shaping audio timbre toward the target prompt. ConterFlow substantially improves counterfactual Video Foley generation compared to naive negative prompting and state-of-the-art baselines. To evaluate replacement quality, we propose a metric leveraging a text-audio co-embedding space to measure both target-prompt evidence and residual visually implied source leakage. Video demonstrations and code are available at https://gyubin-lee.github.io/counterflow-demo/

arXiv:2604.04025v1 Announce Type: new Abstract: Background music shapes attention, affect, and approach behavior in commercial environments, yet the neural plausibility of AI-generated music for such settings remains poorly characterized. We present an in-silico pilot study that combines Wubble, a generative music system, with TRIBE v2, a publicly released whole-brain encoding model, to estimate cortical response profiles for prompt-conditioned retail music. Five fully instrumental tracks were generated to span low-to-high arousal, sparse-to-dense arrangement, and neutral-to-positive valence prompts, then analyzed with audio-only TRIBE v2 inference on loudness-normalized waveforms. Analysis focused on fsaverage5 cortical predictions summarized over auditory, superior temporal, temporo-parietal, and inferior frontal HCP parcels. The fast bright major-pop condition produced the largest whole-cortex mean activation (0.0402), the strongest prefrontal ROI composite response (0.0704), and the highest parcel means in IFJa (0.1102), IFJp (0.0995), A5 (0.0188), and area 45 (0.0015). Pairwise spatial correlations ranged from 0.787 to 0.974, indicating that prompt variation modulated predicted cortical states rather than yielding a single undifferentiated response profile. Predicted cortical surface maps further revealed visually distinct spatial organization between low-arousal and high-arousal conditions. These results support a cautious claim of cortical neurological plausibility: prompt-conditioned AI music can systematically shift predicted auditory-temporal-prefrontal patterns relevant to salience and valuation. Although the study does not establish subcortical reward engagement or consumer behavior, it provides a reproducible framework for neural pre-screening and pre-optimization of commercial music generation against biologically informed cortical proxies.

arXiv:2604.03329v1 Announce Type: cross Abstract: Violence detection benefits from audio, but real-world soundscapes can be noisy or weakly related to the visible scene. We present CoLoRSMamba, a directional Video to Audio multimodal architecture that couples VideoMamba and AudioMamba through CLS-guided conditional LoRA. At each layer, the VideoMamba CLS token produces a channel-wise modulation vector and a stabilization gate that adapt the AudioMamba projections responsible for the selective state-space parameters (Delta, B, C), including the step-size pathway, yielding scene-aware audio dynamics without token-level cross-attention. Training combines binary classification with a symmetric AV-InfoNCE objective that aligns clip-level audio and video embeddings. To support fair multimodal evaluation, we curate audio-filtered clip level subsets of the NTU-CCTV and DVD datasets from temporal annotations, retaining only clips with available audio. On these subsets, CoLoRSMamba outperforms representative audio-only, video-only, and multimodal baselines, achieving 88.63% accuracy / 86.24% F1-V on NTU-CCTV and 75.77% accuracy / 72.94% F1-V on DVD. It further offers a favorable accuracy-efficiency tradeoff, surpassing several larger models with fewer parameters and FLOPs.

arXiv:2604.03333v1 Announce Type: cross Abstract: Symbolic music generation has made significant progress, yet achieving fine-grained and flexible control over composer style remains challenging. Existing training-based methods for composer style conditioning depend on large labeled datasets. Besides, these methods typically support only single-composer generation at a time, limiting their applicability to more creative or blended scenarios. In this work, we propose Composer Vector, an inference-time steering method that operates directly in the model's latent space to control composer style without retraining. Through experiments on multiple symbolic music generation models, we show that Composer Vector effectively guides generations toward target composer styles, enabling smooth and interpretable control through a continuous steering coefficient. It also enables seamless fusion of multiple styles within a unified latent space framework. Overall, our work demonstrates that simple latent space steering provides a practical and general mechanism for controllable symbolic music generation, enabling more flexible and interactive creative workflows. Code and Demo are available here: https://github.com/JiangXunyi/Composer-Vector and https://jiangxunyi.github.io/composervector.github.io/

arXiv:2604.04229v1 Announce Type: cross Abstract: Learning aligned multimodal embeddings from weakly paired, label-free corpora is challenging: pipelines often provide only pre-extracted features, clips contain multiple events, and spurious co-occurrences. We propose HSC-MAE (Hierarchical Semantic Correlation-Aware Masked Autoencoder), a dual-path teacher-student framework that enforces semantic consistency across three complementary levels of representation - from coarse to fine: (i) global-level canonical-geometry correlation via DCCA, which aligns audio and visual embeddings within a shared modality-invariant subspace; (ii) local-level neighborhood-semantics correlation via teacher-mined soft top-k affinities, which preserves multi-positive relational structure among semantically similar instances; and (iii) sample-level conditional-sufficiency correlation via masked autoencoding, which ensures individual embeddings retain discriminative semantic content under partial observation. Concretely, a student MAE path is trained with masked feature reconstruction and affinity-weighted soft top-k InfoNCE; an EMA teacher operating on unmasked inputs via the CCA path supplies stable canonical geometry and soft positives. Learnable multi-task weights reconcile competing objectives, and an optional distillation loss transfers teacher geometry into the student. Experiments on AVE and VEGAS demonstrate substantial mAP improvements over strong unsupervised baselines, validating that HSC-MAE yields robust and well-structured audio-visual representations.