Speech
-
Audio Token Attention Is Predictable Before the Language Model Runs
2026-09-30 03:28:23Kyoungjun Park, Yunzhe Li, Lili QiuAbstract
A large audio language model (LALM) turns a minute of speech into 750-1,500 tokens and prefills every one. Image-token pruning often cuts after the language model's first layers, where image tokens draw little attention. Audio tokens draw much more attention there, and their ranking is still far from final, so audio needs a ranking before the language model runs. Surprisingly, the attention an audio token will receive across the language model is already linearly predictable from its encoder output, before the language model runs. A linear map, fitted in closed form without labels, predicts this all-layer attention ranking at $\rho \geq .69$ on eleven of thirteen LALMs. Our method, Triage, cuts audio tokens by this prediction and, on multiple choice, cuts again at layer 2, correcting the prediction with the attention observed there. Triage sets its compression without labels, under two budgets that limit how far its output may differ from the model's own full-audio output. At the conservative budget, its word error rate and accuracy stay within .04 of full audio. At the aggressive budget, Triage beats every baseline in all twelve transcription cases. On multiple choice, at 2.2-5x compression, it outperforms DART, the strongest baseline on average, by .043 in mean accuracy. Because it cuts before the language model, it raises the audio that fits in Qwen2.5-Omni-3B's context window from 21.8 to about 62 minutes. At its most compressive point, Triage lets one GPU serve 4x as many concurrent 5-minute streams of that model. Project page: this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2609.38878
PDF
-
Talk2Agent: Benchmarking Voice Interfaces for Text Agents
2026-09-30 03:19:25Terumi Chiba, Guangzhi Sun, Zheqi Yuan, Chao ZhangAbstract
Large language model (LLM) computer-use agents are typically evaluated with clean written instructions, despite speech being an increasingly popular interface for interacting with such systems. Speech input introduces an additional failure point: transcription errors can alter task-critical entities, constraints, or targets before the agent begins reasoning, while conventional ASR metrics do not directly measure whether the information required for successful execution has been preserved. We introduce Talk2Agent, a benchmark for evaluating how effectively voice interfaces convey human-spoken instructions to LLM-based computer-use agents. Talk2Agent builds human-spoken versions of tasks from WildClawBench and OSWorld and evaluates a range of voice interfaces, including dedicated ASR models, audio-capable LLMs, contextual biasing, and LLM-based ontology repair. Because repeatedly executing long-horizon computer-use tasks is costly and stochastic, we further propose an execution-free, task-conditioned evaluation framework that projects the original task grader onto prompt-addressable intentions and measures how much task-relevant information is retained after the voice interface. On WildClawBench, Talk2Agent's execution-free native projection provides a practical, execution-grounded measure of voice-interface quality, correlating with downstream task completion and improving Pearson correlation by 0.246 over WER/CER on 32 hours of real human speech.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38867
PDF
-
Intonation Perception in Real and Synthetic Speech across Varying Familiarity Levels: A Pilot Study of Equivalence Assessment
2026-09-30 02:32:37Hanrui Zhou, Gaoyuan Zhang, Yixiang Chen, Yujie Xing, Feng Xu, Xurong Xie, Hui ChenAbstract
Language training relies on a corpus constructed by a large number linguistic materials. AI-powered voice clones provide a way to construct the corpus with relatively low cost. Singing voice conversion (SVC) model is used to generate synthetic voices. This study compares participants' performances on natural and synthetic speech in two experiments, similarity perception and intonation recognition. In the accuracy of similarity perception task, a significant interaction between speech type and intonation is found, suggesting that question may serve as a cue for speaker identification but may be influenced by synthetic features. In the accuracy of intonation recognition task, a significant interaction between speech type and familiarity is observed, indicating that speech type affects how much familiarity contributes to voice processing.
Abstract (translated)
URL
https://arxiv.org/abs/2610.08839
PDF
-
A barrier or a booster? Familiarity effects on Mandarin emotion prosody recognition using AI-powered voice cloning
2026-09-30 02:26:48Feng Xu, Gaoyuan Zhang, Shanshan Xue, Yixiang Chen, Hanrui Zhou, Xurong Xie, Hui ChenAbstract
Emotion prosody perception requires simultaneous processing of acoustic cues and speaker identity. While listeners effortlessly decode natural speech, AI synthetic voices introduce cognitive complexities due to subtle acoustic atypicalities. It remains unclear how these synthetic features interact with a listener's prior social knowledge and memory of a familiar speaker. This study investigated how speech sources (human vs. AI) and speaker familiarity affect emotion recognition accuracy and cognitive load. A within-subject task with Mandarin-speaking adults evaluated behavioral (accuracy, reaction time) and physiological data (heart rate variability). Results showed that human voices yielded significantly higher accuracy and faster processing times than AI voices, while HRV did not significantly differentiate between conditions. These findings show that decoding synthetic speech is gated by top-down social cognition, highlighting limitations in current AI synthesis technologies.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38794
PDF
-
Acoustic Biomarkers for Alzheimer's Disease Detection from Spontaneous Speech
2026-09-30 02:15:12Anh L Tran, Arjee Javellana Restar, Bereket KindoAbstract
Alzheimer's disease (AD) remains a major public health challenge, and early identification is crucial for timely intervention and care planning. Recent work has shown that speech-based digital biomarkers can capture subtle changes in vocal production, prosody, and timing that often precede or accompany cognitive decline. In this study, we investigate whether acoustic features extracted from spontaneous speech can reliably distinguish individuals with AD from healthy controls. Using a structured speech dataset derived from the Cookie Theft picture-description task, we extract clinically informative low-level descriptors spanning prosodic, spectral, and temporal domains, and evaluate a range of machine learning classifiers. Our results indicate that acoustic biomarkers carry strong discriminative value, with the best-performing models achieving clinically meaningful separation between diagnostic groups. These findings support the use of non-invasive, low-cost audio-based screening tools and motivate further work on multimodal and longitudinal assessment frameworks for dementia risk detection.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12500
PDF
-
RAST: Resolution-Aware Privileged Structure Transfer for Low-Resolution Audio Activity Recognition
2026-09-30 02:12:47Ji Hwan Park, Gautham Krishna Gudur, Yufei Shen, Dawei Liang, Edison ThomazAbstract
Audio is increasingly used for human activity recognition (HAR) because it captures object interactions, environmental events, and contextual cues in everyday environments. High-resolution (HR) audio provides rich acoustic information for model development but incurs substantial energy and storage costs and may expose sensitive speech content. Low-resolution (LR) audio offers a more privacy-preserving and resource-efficient alternative for deployment, but reduced sampling rates can remove acoustic cues essential for activity recognition, leading to significant performance degradation. We formulate this training-deployment mismatch as sensor-resolution privileged learning, in which HR audio is available during training, while inference relies exclusively on LR audio. We propose RAST, a resolution-aware transfer framework that compresses HR teacher representations by preserving token-level information and neighborhood structure before performing localized HR-LR alignment. Experiments on the SAMoSA and AudioIMU datasets show that RAST consistently outperforms LR-only training and direct teacher-transfer baselines, improving LR-only recognition by up to approximately 7.8% while requiring only LR audio at inference.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38780
PDF
-
Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning
2026-09-29 23:25:21Jian Chen, You Zhang, Mark VintonAbstract
TTS systems with autoregressive semantic modeling have demonstrated strong zero-shot voice cloning performance and rich expressive variation, but their sequential decoding incurs substantial latency. Non-autoregressive alternatives offer much faster generation, yet often rely on more restrictive reference conditioning, such as requiring transcripts of the reference speech during inference. We present Tacit-TTS, an efficient transcript-free zero-shot voice cloning system distilled from IndexTTS2. Our model replaces autoregressive text-to-semantic decoding with masked non-autoregressive generation, introduces training-free acoustic length estimation, and accelerates the flow-matching renderer through ReFlow distillation. Across two English and two Mandarin datasets, Tacit-TTS achieves competitive zero-shot quality while generating speech over 10x faster than IndexTTS2 for utterances longer than 5 seconds. Its transcript-free conditioning further supports cross-lingual and non-lexical references. We validate this capability using references from eight other languages, infant babble, and synthetic gibberish, where transcript-dependent systems often degrade or fail due to unreliable ASR transcripts.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38658
PDF
-
Marking Contour Tones in Yor`{u}b'{a}
Abstract
Yorùbá is a tonal language in which contour tones pose persistent orthographic challenges. These are especially notable for personal names and lexical items whose conventional spellings avoid vowel lengthening that would otherwise provide a host syllable for the second tone. A particular concern is a class of names in which the conventional spelling does not just omit tonal information but inverts the meaning of said name, sometimes asserting the opposite of what the name intends. This paper describes the problem, illustrates the inadequacy of current solutions, and proposes the adoption of the caron and circumflex marks. These are symbols with precedent in Yorùbá phonological scholarship since Olmsted (1951), used as orthographic conventions on single vowels to encode rising and falling contour tones, making them accessible for the first time through standard keyboard input and computational text processing. The proposal is supported by an implementation in the WriteYoruba keyboard and the TTSYoruba speech synthesizer, whose architecture and listener evaluation are reported separately (Tubosun et al., 2026).
Abstract (translated)
URL
https://arxiv.org/abs/2609.38627
PDF
-
Towards Model as a Library: Offline, Community-Sourced AI for Low-Resource African Languages
Abstract
Large language models are frequently proposed as a route to AI-powered services for African communities, but they are least reliable exactly where the need is greatest: all African languages remain low-resource by any standard measure, and models trained on scraped, standardised text systematically misrepresent the dialectal and regional variation of how people actually speak. We introduce \textbf{Model as a Library (MaaL)}, a software architecture that packages small, community-enrolled speech models as versioned on-device dependencies, enabling offline structured data collection that cannot generatively hallucinate, for populations that current language models serve worst. Rather than relying on web-scraped corpora, MaaL's vocabulary is enrolled directly from a small number of example recordings by the speakers themselves, at the point of deployment. We describe the architecture and its central mechanism - keyword spotting that turns a closed-vocabulary text form into a voice form, filled and submitted entirely on-device - and propose transpiling the closed-vocabulary elements already present in widely-deployed digital form tools into MaaL schemas, a low-friction path to voice-first, offline data collection for the low-literacy populations these tools already reach. This is a position and system-design paper: we describe the concept, the mechanism, and an analytical feasibility case, and identify what a working implementation still requires.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38574
PDF
-
VAmoS Part Deux: Harder, More Realistic Voice-Agent Simulation
2026-09-29 20:33:57Joshua Meyer, Sahar Shayegan, Ritiz Tambi, Ali Khan, Sun Kim, Victor Shih, Mehdi Jamei, Andi PartoviAbstract
Voice agents in production must handle several requests, background speech, and customers who lose patience. We introduce VAmoS Energy, a benchmark that combines these challenges in 100 calls about utility billing and payment assistance. Each caller makes two to four requests. The agent has sixteen tools backed by a stateful Stripe billing twin and the Apache Fineract loan engine, with account access blocked until caller verification succeeds. The tasks use public household electricity data and a policy based on Pennsylvania's residential billing rules. An LLM-as-a-verifier checks the agent's actions and spoken figures against explicit requirements. On a calibration run, it agrees with a code verifier on 99.1% of checks. Across fourteen voice stacks and three repeats per task, completion ranges from 17.3% to 44.7%. Grok Voice leads, and Gemini 3.8 Live and GPT-Live follow at about the same cost per call. Background television reduces pooled completion from 38.7% to 8.6%. The simulated caller often accepts an incorrect result because it hears the agent's words but cannot inspect its actions. These findings show why voice agents need evaluation across the whole call, including what they say, what they change, and how they handle competing speech.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38512
PDF
-
Role-guided Speaker Deletion Verification in Clinical Psychiatry Speech Recordings with Audio Language Models
2026-09-29 20:20:24Joseph T Colonel, Daniel Katzman, Kelsey Kirker, Adam N Davidson, Shalaila S Haas, Cheryl Corcoran, Ren\'{e} S Kahn, Guillermo Checci, Baihan LinAbstract
Clinical research in psychiatry increasingly relies on large scale collection of spoken language data to identify acoustic and linguistic biomarkers. Yet evolving consent and protocol requirements can oblige investigators to remove a designated speaker from multi-speaker recordings and to verify said removal at a scale infeasible for manual review of entire corpora. We study this verification problem for role-driven dyadic clinical dialogue in psychiatry and investigate it with two parallel, symmetric pipelines: confirming that clinician speech has been removed from psychiatric interview recordings, and confirming that patient speech has been removed from the same recordings. Each pipeline redacts the raw audio for its target role and then scans the surviving output with audio-language and large-language models to identify missed deletions. We evaluate this approach on a corpus of 48 dyadic recordings drawn from psychiatry settings, testing four open-weight models in an inference-only setting: Gemma-4-12B, Gemma-4-31B, Nemotron-3-Nano, and Nemotron-3-Nano-Omni. A disjunctive OR ensemble over fourteen model-view configurations had a combined F1 of 0.478 (precision 0.330, recall 0.870), an improvement over individual model estimates driven by recall gains that point to substantial complementarity across models and context views.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38491
PDF
-
GestAdapt: Workspace-Conditioned Co-Speech Gesture Generation for Humanoid Robots
2026-09-29 18:55:37Bosong Ding, Xianglin Zhang, Miao Xin, Murat Kirtay, Giacomo SpiglerAbstract
Co-speech gestures for robots must adapt not only to speech and embodiment, but also to the workspace available for performing the motion. Since the same speech can be accompanied by different gestures, a robot can respond to workspace constraints, e.g., gestures for speech next to a wall. In these scenarios, the robot should gesture in a suitable motion rather than simply correcting an unconstrained one. To achieve this goal, we present GestAdapt, a workspace-conditioned framework that conditions co-speech gesture generation on a prescribed wrist workspace. The GestAdapt framework learns from six complementary co-speech corpora through a shared motion representation and supports retargeting to different robot embodiments. Quantitative evaluation shows that generated motions remain close to the real-motion distribution while respecting the workspace. In a user study, gestures generated under modified workspace constraints receive a mean quality score of 3.24/5, above our no-workspace variant (2.43/5) and below the reference motions (3.68/5). In a real robot evaluation, all compared motions are retargeted to the Reachy2 humanoid robot under identical workspace constraints. Motions generated with our framework rank first in 69.7\% of comparisons, higher than our no-workspace variant baseline and retargeted ground-truth motions constrained afterward. Overall, the results support adapting gestures to the available workspace during generation, rather than modifying unconstrained trajectories afterward to satisfy workspace constraints, potentially compromising gesture naturalness.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38400
PDF
-
TALK-Dem: Benchmarking Embodied Task Planning under Dementia-Associated Communication Patterns
2026-09-29 18:33:21Guangxin Zhao, Yiran Hu, Yuan Cao, Chenxi Jiang, Jianfei Yang, Yegang Du, Yasuyuki Taki, Yoshifumi Kitamura, Lin Gu, Zhi ZhengAbstract
Existing LLM-driven robot task planners rely on a taken-for-granted assumption of an ideal user whose instructions are clear, complete, and task-focused. However, when interacting with real-world users, especially those experiencing cognitive impairments, such as people living with dementia (PLWD), the planners often make mistakes and even pose physical safety risks. We proposed TALK-Dem (Talking Attributes and Linguistic Knowledge in Dementia), the first benchmark for evaluating LLM-driven robot task planning under dementia-associated verbal communication. TALK-Dem contains 4,800 instructions and covers five typical communication patterns, including Referential Imprecision, Object Substitution, Empty Speech, Topic Drift, and Intrusion, at three intensity levels. Experiments across six open-weight LLMs reveal a substantial robustness gap. Across communication patterns, open-weight models exhibited performance drops of up to 22.3 percentage points compared to ideal instructions. This revealed a critical gap and even danger for real-world applications, especially in assistive robotics, where locally deployable models are necessary due to privacy concerns and connectivity constraints. To mitigate this issue, we proposed the Context-Aware Retrieval from Experience (CARE) method, which retrieves relevant previously resolved tasks to provide task-specific interpretation and planning context. CARE generally outperformed standard prompting baselines across the six open-weight models, improving average task success by 18.1 percentage points over the vanilla prompt. These results highlighted the importance of both evaluating communication robustness and developing effective adaptation strategies for locally deployable assistive robots. The TALK-Dem dataset is publicly available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38371
PDF
-
EmoRES-TTS: Residual-Enhanced Vector Steering for Emotional Speech Generation
2026-09-29 17:59:05Kuan-Po Huang, Haohe Liu, Puyuan Peng, Haibin Wu, Zhaoheng Ni, Hung-yi Lee, Jinwon Lee, Neha ChachraAbstract
Emotion-conditioned text-to-speech (TTS) models may fail to express the requested emotion reliably, and improving controllability by additional training is costly in both computation and emotion-labeled speech training data. We therefore study vector steering, a training-free approach that modifies the internal representations of a frozen model. CoCoEmo, a conventional vector steering method for emotion TTS, treats each emotion vector as an indivisible direction controlled by a single global strength, limiting adherence to the requested emotion. In this work, we first discover that an emotion vector can be decomposed into a shared component that moves speech away from neutral expression and a residual component that directs generation toward the requested emotion. Building on this finding, we propose Emotion Residual-Enhanced Steering for TTS (EmoRES), a novel method that controls the two components without retraining the backbone. On IEMOCAP, EmoRES outperforms CoCoEmo across all four objective emotion metrics on the IndexTTS-2 and CosyVoice2 backbones. Rank correlation improves by 26.13 and 12.97 percentage points, corresponding to relative gains of 118.8% and 33.1%, while emotion hit rate improves by 12.95 and 6.92 points, corresponding to relative gains of 20.1% and 9.8%. Human evaluation further shows a relative improvement up to 35.0% in the rate at which listeners correctly identified the dominant requested emotion and up to a 17.3% improvement in fidelity, while listeners prefer EmoRES for naturalness in up to 63.8% of pairwise comparisons. Component ablations further demonstrate that effective control benefits from preserving the shared component while strengthening the residual of the emotion steering vectors.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38157
PDF
-
Pruning for Efficiency, Paying in Fairness: Demographic Disparities in Pruned Speech-LLMs
2026-09-29 17:46:50Ganesh Pavan Kartikeya Bharadwaj Kolluri, Michael Kampouridis, Ravi ShekharAbstract
Speech-LLMs are expensive to run, making compression important for real-world deployment. However, compressed models are usually selected using aggregate word error rate (WER), which can hide how pruning affects different demographic groups. In this work, we systematically study the effect of audio encoder pruning on SLAM-ASR for different demographic groups. Using the Fair-Speech and Common Voice datasets, we found that the pruning does not affect all demographic groups equally; the gap between best- and worst-performing groups increases in fold. These disparities appear across all three encoder scales, but only the largest model initially hides them behind aggregate WER. LoRA adaptation improves WER for every group, but benefits groups already performing well more strongly and widens for certain groups. On Common Voice English, Danish, and Dutch, accent gaps persist but do not clearly widen, showing that the fairness effects of pruning vary across datasets and must be measured directly. Our findings suggest that for pruned models, deployment decisions should include per-group WER, with the worst-performing group's error rate as an explicit criterion.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38106
PDF
-
GLaS-JEPA: Gaussian-Regularized Speech SSL without Engineered Prediction Targets
2026-09-29 15:22:51Gaspard Bott\'e, S\'everin Baroudi, Samir Sadok, Francesco Paissan, Thomas Hueber, Xavier Alameda-Pineda, Ricard Marxer, Mirco RavanelliAbstract
Speech self-supervised learning aims to learn general-purpose representations for downstream speech tasks. However, current approaches rely on complex, carefully designed prediction targets. We challenge this necessity with GLaS-JEPA, a framework that directly predicts the current encoder's continuous representations at masked positions, without contrastive learning, discrete targets, or separate EMA target encoders. We prevent representation collapse using SIGReg representation-space regularization, eliminating the need for engineered target-generation mechanisms. Pretrained on 960 hours of LibriSpeech, our 57M-parameter model achieves a 6.89% WER on frozen-encoder SUPERB ASR and a 25.87% CER on slot filling, outperforming the best non-distilled sub-90M baselines by 43.1% and 22.0%, respectively. These results demonstrate that highly competitive speech representations can emerge from a radically simplified training recipe.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37798
PDF
-
Learning as Deepfakes Evolve: RF-Prompt for Continual Audio Deepfake Detection
2026-09-29 13:43:38Yuankun Xie, Xiaoxuan Guo, Xiaopeng Wang, Siqing Qin, Shaole Li, Kong Aik LeeAbstract
Continual audio deepfake detection requires learning newly emerging deepfake methods while retaining discrimination of previously encountered speech. Existing dataset-incremental evaluation changes both real-speech domains and deepfake mechanisms, making their effects difficult to distinguish. We construct five task organizations over identical training, development, and evaluation pools to study these factors under a controlled sample budget. Our proposed Real-Anchored Mechanism-Incremental (RAMI) protocol reflects the practical setting in which available real speech provides a recurring mixed-domain reference while new deepfake mechanisms arrive incrementally. We further propose RF-Prompt, an asymmetric continual prompt-learning method that preserves reusable real-speech knowledge through a shared real prompt and expands mechanism-specific knowledge through inherited fake experts with orthogonal residuals. Input-adaptive soft fusion combines the accumulated experts into a fixed number of injected tokens without requiring task identity at inference. On RAMI, RF-Prompt achieves 10.110% average EER and 10.370% pooled EER, outperforming all evaluated continual-learning baselines. Across the five controlled protocols, RAMI yields the lowest common-average and pooled EER. Component ablations, limited-data experiments, and cross-backbone evaluations further validate the proposed design.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37586
PDF
-
Look What You Made Us Cluster: Hate Narrative Extraction from Reddit Discourse
2026-09-29 12:43:31Annabelle K. L. Chua, Forster J. Khoo, Joel C. R. Tan, Huey Ting Ang, Kheng Hwee Tan, Joel Y. A. Sim, Shirley W. H. Ow, Ria Mundhra, Elsie C. K. Toh, Youfeng Xu, Lynnette H. X. NgAbstract
Narrative extraction allows us to identify online hate narratives, supporting the construction of rigorous detection systems. Existing computational approaches, however, are limited in precision as they rely on semantic representations, which tend to capture only surface-level meaning. To detect more precise and interpretable narratives, we present an extraction pipeline that represents narratives as entity-evaluation pairs. Narratives are extracted using a Large Language Model (LLM) reasoning process that extends Aspect-Based Sentiment Analysis, identifying the aspect, classifying its judgement type as the basis for evaluation, and deriving the evaluation accordingly. Extracted narratives are then clustered using Leiden, following which clusters are resolved to an intended level of granularity through an LLM-guided refinement process. We illustrate this narrative pipeline with English Reddit comments from 2024 that criticize Taylor Swift, analyzing a representative cluster that exhibits hate speech patterns to demonstrate its interpretive value.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37408
PDF
-
What Comes Next? Omni-StoryBench for Evaluating Story-Grounded Omnimodal Generation
2026-09-29 11:53:15Sieun Hyeon, Yejoon Lee, Mintaek Lim, Woojin Kim, Jaeik Kim, Jaeyoung DoAbstract
Omnimodal evaluation should go beyond independent text, image, and speech production: individually plausible outputs may not express a coherent shared event. We introduce Omni-StoryBench, a story-grounded omnimodal benchmark evaluating whether models can coherently continue stories across image, narration, and speech. Each instance provides a current storybook page and structured next-page conditions, requiring models to generate the next illustration, narration, and spoken character utterance. Omni-StoryBench contains 900 rigorously validated story transitions from openly licensed children's books, with ground-truth next-page references and speech metadata. We evaluate systems with modality-specific metrics and consistency-centered LLM-as-a-judge rubrics for context preservation, condition following, reference consistency, and cross-modal coherence. Across 32 baseline configurations spanning orchestration, semi-orchestration, and native any-to-any paradigms, we find orchestration with strong VLM planning most reliable, while current native omnimodal models often struggle with output completeness and controllability. Our analysis shows text-side performance is associated with image and speech quality, but image generation and visual continuity form the clearest observed bottleneck among the evaluated configurations. These results position Omni-StoryBench as a system-level benchmark measuring coherent omnimodal generation beyond isolated modality quality.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37317
PDF
-
JoyAI-Voice 2.0: A Full-Continuous Autoregressive Speech Generation Model with Semantic-Acoustic Joint Representation
2026-09-29 10:12:21Yafeng Chen (Intern), Boya Dong (Intern), Yankun Huang (Intern), Hao Li (Intern), Jingdong Li (Intern), Xiangyu Liang (Intern), Hao Ni (Intern), Wenchao Wang (Intern), Yuxuan Wang (Intern), Zhangyu Xiao (Intern), Wei Deng (Intern), Nan Duan (Intern), Yu Gu (Intern), Wenhao Guan (Intern), Weisheng Han (Intern), Yabin Li (Intern), Yuan Liu (Intern), Jiaxin Ye (Intern), Fan Yu, Lin ZhuAbstract
We present JoyAI-Voice~2.0, an end-to-end anthropomorphic speech generation model built upon a fully continuous, dual-encoder architecture. Raw speech is encoded into continuous latents and partitioned into patches. Each patch is decomposed by a semantic-acoustic dual encoder into a semantically purified representation and an acoustic representation, which are fused and jointly fed to a causal autoregressive Transformer for planning. The Transformer predicts the conditioning for the next patch, and a local diffusion Transformer renders its full latents for 48\,kHz synthesis. The model is trained with a joint flow-matching and stop-prediction objective, followed by supervised fine-tuning and reinforcement learning with DiffusionNFT to improve model performance. It achieves the lowest average word error rate of 2.51\% on Seed-TTS, a 14.9\% relative reduction over the strongest baseline, and state-of-the-art attribute fidelity on InstructTTSEval in both Chinese and English, leading on 5 of 10 perceptual dimensions with the highest overall score of 0.893 on MDVD-Eval.
Abstract (translated)
URL
https://arxiv.org/abs/2610.08834
PDF
