Daily Papers — 2026-09-17

20 papers on audio, speech, music, and acoustics.

1. Reading Emotions in the Token Space: Discriminative Adaptation of SpeechLLMs for Emotion Recognition

Authors: Hasindri Watawana, Sergio Burdisso, Esaú Villatoro-Tello, Manjunath K E, Kadri Hacioglu et al.

Categories: cs.CL, cs.AI, cs.SD, eess.AS

SpeechLLMs have shown strong potential for emotion recognition, yet they read the predicted emotion off a generative decoder not suited for classification: it can emit labels outside the target set and favors frequent classes. We propose a discriminative adaptation that reads the final prompt token’s hidden state through a classification head, producing a label in one forward pass without modifying the backbone. Because this readout starts from the hidden state the model would otherwise decode, it gives a controlled comparison of generative and discriminative inference in an otherwise identical speechLLM. We keep the head a single linear layer, trading little accuracy for interpretability: each emotion becomes one direction in the LLM output token space, revealing associated tokens. On IEMOCAP, across two speechLLM architectures, it improves Macro F1 and removes hallucinations, with largest gains on realistic ASR transcripts. Our analysis reveals that these emotion directions encode indirect associations mirroring biases in web-scale text.

2. VākQA: A Benchmark and Evaluation Study for Telugu Spoken Factoid Question Answering

Authors: Bhavana Akkiraju, Ravi Sastry Kolluru, Sri Charan D, Srihari Bandarupalli, Santosh Kesiraju et al.

Categories: cs.CL, eess.AS | Paper is accepted in IEEE SLT 2026

Question answering has advanced rapidly with large language models, but predominantly for high-resource languages, in both text and spoken settings. Spoken question answering (SQA) benchmark for Telugu remains unexplored, and the reliability of automatic evaluation in this setting remains unquantified. We introduce VākQA, a Telugu SQA benchmark of 2,001 factoid question-answer pairs across six domains, with 2.53 hours of speech audio, bilingual transcriptions, and human-verified reference answers. We first validate evaluation methods against human judgements: Gemini-as-a-judge best approximates human ratings but is non-uniformly strict, while open-weight judges systematically penalize correct Telugu answers that differ in surface form from the reference. Using this validated setup, we benchmark proprietary and open-weight models across input modality, language, and domain. We observe that Telugu phrasing retains cultural specificity that is lost in translation, speech input introduces phonetic confusions that alter question meaning, and cascaded ASR-MT errors compound progressively. VākQA is publicly released.

3. CircleMatch: Prototype Matching with Circular Temporal Statistics for Tiny Keyword Spotting

Authors: Jiajun Sun, Zhe Gao

Categories: cs.SD | 5 pages, 4 figures, 2 tables

Keyword spotting (KWS), the task of identifying predefined words in speech, is a core capability of voice-enabled devices. Achieving high KWS accuracy under tight parameter budgets across different vocabulary sizes remains challenging. We present CircleMatch, a matching framework enabling KWS with very few parameters. Its encoder independently compresses frequency bands and fuses them into frame features. These features are then matched against learned class-specific prototypes to produce temporal response curves. Parameter-free circular aggregation encodes time as angles and summarizes response distributions and relative timing for classification. We develop four tiny variants, Circle-D4, Circle-D8, Circle-D16, and Circle-D32, ranging from approximately 1k to 7k parameters in the 12-class setting. Experiments with multiple random seeds on Speech Commands v1/v2 and the English and Spanish Micro subsets of the Multilingual Spoken Words Corpus demonstrate competitive accuracy with tiny models. Our qualitative analysis further suggests approximate shift equivariance of prototype responses and adaptation to temporal compression. Code and model weights are available at https://github.com/ora942878/CircleMatch.

Authors: Kunyu Feng, Yuxiang Wang, Li Wang, Wan Lin, Zhizheng Wu

Categories: cs.SD, cs.AI | 5 pages, 2 figures, 3 tables

Generalizing to unseen attacks remains challenging for audio deepfake detectors, and collecting training data covering all potential attacks is impractical. We explore recurrent refinement in an already-trained SSL-based detector without additional data or changes to its original parameters. However, directly recycling encoder outputs as inputs degrades detection in our diagnostic. We propose CoReLoop, which makes this reuse effective by adapting recurrent inputs to the frozen encoder, controlling state updates, and aligning refined outputs with the frozen classifier. By training only lightweight refinement modules and loop-specific low-rank adapters on the original data, CoReLoop enables additional refinement while preserving the detector’s original first-pass prediction. On 14 cross-domain test sets, the 24-layer model reduces pooled equal error rate (EER) from 4.85% to 3.74% with two passes, with approximately 10M trainable parameters out of 598M. To selectively apply this refinement, an optional halting head chooses the depth for each utterance, achieving 3.73% pooled EER with an average of 1.18 passes.

5. Multi-Dimensional Prosody Judgment For Live Streaming Speech Synthesis

Authors: Zifan Guan, Longyu Lu, Junan Zhang, Zhizheng Wu, Meiguang Jin et al.

Categories: cs.SD, cs.AI

Evaluating live streaming speech synthesis (TTS) requires assessing fine-grained, highly expressive prosody such as emotion, intonation, and energy which traditional MOS predictors fail to capture. While proprietary Large Language Models (LLMs) like Gemini can evaluate these aspects, they are too costly for massive inference and reinforcement learning feedback. To address this, we first introduce Live-ProsodyJudge (LPJ), a cost-effective pairwise evaluator distilled from Gemini into Qwen3-Omni. However, we identify a critical flaw in standard multi-dimensional evaluation: verdict coupling. The judge tends to lazily align all individual dimension scores with its overall preference, collapsing a rich multi-dimensional rubric into a single preference bit. To resolve this, we further propose Decoupled-Live-ProsodyJudge (D-LPJ). D-LPJ eliminates the overall verdict target to prevent blind following, masks uncertain pair-dimensions during Supervised Fine-Tuning(SFT), and introduces a novel span-local GRPO strategy that applies normalized advantages strictly to their corresponding rationale spans. Evaluated on highly curated human-annotated test sets, 10 sample balanced-order LPJ achieves higher point accuracy than a single Gemini call, while D-LPJ successfully produces independent,decoupled dimension judgments. Furthermore, in a Best-of-8 TTS candidate selection tournament, the LPJ-selected utterance falls within the human top-3 in 85.29% of high-confidence cases, demonstrating its efficacy for fine-grained TTS preference optimization.

6. Robust Workflow Generation via Adversarial Learning for Audio Deepfake Detection

Authors: Xiang Li, Pin-Yu Chen, Wenqi Wei

Categories: cs.SD, cs.AI

The rapid advancement of speech synthesis and voice conversion technologies has made audio deepfakes increasingly realistic, posing serious security risks in practical applications. While existing detection methods achieve strong performance under controlled conditions, they often fail to generalize under real-world perturbations and corruptions. In this paper, we propose ROGUE, a framework that dynamically constructs robust detection workflows by orchestrating multiple detection tools. ROGUE formulates workflow generation as a sequential decision-making problem and introduces a dual-agent paradigm, where a perturbation agent generates audio perturbations and a policy agent learns to select and execute detection tools under perturbed conditions. Through adversarial learning, ROGUE enables perturbation-aware tool selection, adaptive execution strategies, and improved robustness to distribution shifts. Extensive experiments across multiple datasets and real-world corruptions demonstrate that ROGUE consistently outperforms strong baselines in both robustness and generalization. Our results highlight the effectiveness of adversarially optimized workflow generation for building reliable audio deepfake detection systems in real-world deployment settings.

7. A Deep Neural Network for Predicting Continuous Human EEG Across the Auditory Pathway in Response to Sound

Authors: Thomas J Stoll, Ross K Maddox

Categories: eess.AS | 5 pages. Submitted to ICASSP 2027

Computational models of auditory physiology commonly target specific responses or stages of the auditory pathway, limiting their ability to integrate findings across experimental paradigms and neural timescales. We present a foundation model of human auditory electrophysiology: a causal neural network trained to map binaural acoustic waveforms directly to high-sample-rate EEG. The model was trained on approximately 250 hours of EEG data from 92 subjects, with varied electrode montages and stimuli spanning tonebursts, speech, and music. We tested whether the model recovered effects of stimulus rate, frequency, and presentation method on auditory brainstem responses (ABRs); subcortical and cortical temporal response functions (TRFs) to continuous speech; and the click-evoked binaural interaction component (BIC). Predicted ABRs and TRFs reproduced established response morphology and stimulus-dependent effects, with model-grand-average correlations falling within the corresponding subject-level human distributions. The model-predicted BIC metrics closely resembled the values reported in the literature. These findings demonstrate that a single audio-to-EEG model can capture auditory physiology across paradigms and timescales, supporting future in silico experimentation and hearing technology applications.

8. Alignment-Path Distillation from Non-streaming ASR-LLMs for Streaming Speech Recognition

Authors: Yan Jia, Kai Huang, Junjie Chen, Feng-Long Xie, Xu Tang et al.

Categories: eess.AS | 5 pages, 1 figure, 2 tables

In this paper, we propose an alignment-path distillation framework for streaming automatic speech recognition (ASR) with large language models (LLMs). Interleaved streaming ASR-LLMs use forced alignments (FA) from alignment models, such as those trained with connectionist temporal classification (CTC), to construct speech-text training sequences. However, alignments obtained from a separate acoustic model may be inconsistent with those learned by LLM-based ASR. This motivates us to transfer alignment information from a non-streaming ASR-LLM to improve streaming recognition. Specifically, we extract monotonic alignment paths from a non-streaming teacher’s soft text-audio attention and use them to construct interleaved training sequences. The framework also includes logit and hidden-state distillation to learn from the teacher’s output distributions and internal representations. Experimental results show that, without logit or hidden-state distillation, training with teacher-derived alignment paths achieves a 5.2% relative error rate reduction compared with training using forced alignments. When both models use logit and hidden-state distillation, teacher-derived alignments yield a 3.9% relative error rate reduction, with similar mean emission latency but higher flicker. The complete framework achieves a 16.6% relative error rate reduction compared with training using forced alignments without logit or hidden-state distillation.

9. Beyond the Stability–Plasticity Frontier in Streaming Target Speaker Extraction

Authors: Yuesheng Ma, Linyang He, Nima Mesgarani

Categories: eess.AS

Streaming target speaker extraction must maintain a representation of whom to extract while the target may fall silent, be masked by interference, or drift acoustically away from enrollment. Existing systems typically hold this state as a stored embedding updated by hand-designed rules. Across 22 configurations, including confidence-gated and oracle-activity-gated updates, we show that this family lies on a stability-plasticity frontier: even perfect target-activity information cannot combine robustness to target absence with adaptation to enrollment-mixture mismatch. We therefore meta-train speaker-state dynamics through the closed streaming loop, exposing the updater to its own contaminated evidence. Our proposed 41k-parameter anchored fast-weights (AFW) memory moves beyond the measured heuristic frontier, gaining 3.0 dB over the best heuristic under severe mismatch while staying within 0.9 dB of static enrollment after 30 s of absence, at under 5% runtime overhead. A gated recurrent unit (GRU) control confirms that the gain is not AFW-specific, while AFW is smaller and more interpretable: under severe mismatch, its write residual grows and aligns with the target rather than the interferer. Code is publicly available at https://github.com/ym2976/anchor-fast-weight.

10. Consensus-Guided Shared-Specific Tri-View Learning for Speech Emotion Recognition

Authors: Bing Huang, Yujian Ma, Xikun Lu, Xianquan Jiang, Jinqiu Sang

Categories: eess.AS | 5 pages, 3 figures, 4 tables. Submitted to ICSAAP 2027

Speech emotion recognition (SER) benefits from heterogeneous acoustic representations, but views derived from the same utterance contain both overlapping emotional evidence and representation-dependent cues. Direct fusion may therefore propagate redundant information or obscure complementary details. To address this issue, we propose Tri-view Consensus-Guided Fusion (TriCGF) for jointly modeling spectrogram, Mel-frequency cepstral coefficients, and HuBERT representations. TriCGF organizes each view into common and view-specific components before fusion. Cross-view Consensus Learning aggregates the common components into a global reference, while View-wise Gated Integration adaptively combines this reference with each view-specific component. A soft difference regularizer further discourages excessive information overlap. Under speaker-independent evaluation, TriCGF achieves 74.19% weighted accuracy (WA) and 75.17% unweighted accuracy (UA) on IEMOCAP, and 94.36% WA and 94.28% UA on EmoDB, outperforming representative SER methods on both datasets.

11. Multimodal Conversational Context for LLM-Based ASR: Data Construction, Training, and Benchmark

Authors: Longhao Li, Jian Tang, Yuxiang Kong, Jie Chen, Binbin Zhang et al.

Categories: eess.AS | Technical report

Conversational context provides semantic and acoustic cues across turns for automatic speech recognition (ASR), but relying on historical transcripts can propagate recognition errors and discard pronunciation and speaker information. We present a multimodal conversational-context framework for LLM-based ASR that integrates a scenario-controlled data pipeline, scalable multimodal context training, and systematic evaluation. We construct dialogues around entities and their confusable forms and interleave historical user speech with assistant text responses for supervised fine-tuning. We also introduce MM-ContextASR Bench, which evaluates contextual understanding and entity error correction across five scenarios. Experiments with Qwen3-Omni and Step-Audio-2-mini reveal limitations in handling irrelevant and erroneous history and show that our data construction and training improve context utilization, with multimodal context achieving the highest overall entity recall on both models. Further experiments on accent, dialect, and target-speaker ASR demonstrate the value of historical speech. The benchmark data and evaluation code are publicly available at https://github.com/llh666521/MM-ContextASR.

12. Model-Agnostic and Language-Agnostic Voice Pipeline Improvement for the Agriculture Domain

Authors: Aakash Singh, Lakshmi Pedapudi, Chandrashekar M S, Sanyam Singh, Naga Ganesh et al.

Categories: eess.AS, cs.AI, cs.CL, cs.SD | 20 tables, 11 figures, 23 pages

FarmerChat is Digital Green’s AI-powered agricultural advisory assistant for smallholder farmers, who access it in their own language through text, voice, or photographs. Voice is a critical channel for this population, yet field-recorded speech is challenging for general-purpose automatic speech recognition (ASR) because recordings frequently contain machinery noise, background media, competing speakers, and domain-specific agricultural vocabulary. These conditions disproportionately affect crop, pest, chemical, and quantity terms that carry the meaning of a farmer’s query. We present a modular, model-agnostic pipeline for improving ASR quality in FarmerChat without fine-tuning or replacing the underlying ASR model. The pipeline combines gated audio enhancement, speaker diarization and target-speaker selection, ASR, domain-aware correction using a weighted agricultural lexicon, and a quality gate for detecting unreliable transcripts. Only the diarization stage is fine-tuned; all other stages use off-the-shelf models behind common interfaces. We evaluate the pipeline on human-annotated FarmerChat recordings in Hindi, Telugu, and Odia using word error rate (WER) and a domain-weighted error rate that gives greater importance to agricultural terminology. The largest improvements occur on multi-speaker recordings, where target-speaker selection prevents competing speech from entering the transcript. Across the full corpus, the pipeline reduces WER by 16-23% relative on three cloud ASR models and by 5% on an on-device model. On multi-speaker recordings, the reductions are 32-42% for the cloud models and 16% for the on-device model. All reported reductions are statistically significant. These results show that targeted preprocessing, speaker selection, and domain-aware post-processing can substantially improve agricultural speech transcription while preserving the underlying ASR model.

13. Foreground Voice Activity Detection: Learning Speaker Selectivity from Supervision

Authors: Guangzhao Yang, Muhammad Huzaifah, Yu Pan, Jinya Sakurai, Ningjie Bai

Categories: eess.AS, cs.SD

Voice activity detection (VAD) fronts most voice-agent pipelines, yet production detectors treat all human speech, background talkers included, as valid activity; in crowded settings this floods recognition, stalls turn-taking, and triggers false barge-in. We formalize Foreground VAD (FVAD): a frame-synchronous, enrollment-free task in which only the dominant speaker, defined by sustained presence rather than instantaneous loudness, is positive, and which reduces to conventional VAD when a single speaker is present. We show that foreground selectivity is largely governed by training supervision: the crucial ingredient is an augmentation recipe pairing foreground-only labels with competing-speaker mixing, generated fully automatically without human annotation. To quantify selectivity we introduce the Background False-Alarm Rate (BG-FAR), gated by foreground F1, and build a controlled benchmark, Mix-Interference, complemented by an adapted VOiCES for real-world far-field evaluation. Across equal-size backbones, Mamba and LSTM perform on par while a longer-context attention model is no better, suggesting that training supervision plays a substantially larger role than temporal modeling capacity in achieving foreground selectivity. The resulting lightweight streaming model, Mamba-FVAD, outperforms commercial VADs and enrollment-based speaker-aware systems in foreground selectivity while staying competitive on conventional VAD, at 1-2 ms per-frame CPU latency.

14. State-Space-Based FIR Filtering on a Quantum Computer

Authors: Roope Salmi, Davide Rocchesso, Vesa Välimäki

Categories: eess.AS, cs.SD, eess.SP, quant-ph | Submitted to ICASSP 2027

Many signal processing tasks require intensive computations. Quantum computing promises to accelerate certain tasks, but algorithms must be designed around the limitations of quantum mechanics. This paper provides a quantum implementation of finite impulse response (FIR) filters, which are a widely used tool in classical signal processing. The filter can be parallelized and composed as part of larger quantum algorithms, with potential for speedups using quantum amplitude estimation and future fault-tolerant quantum hardware. To accomplish these properties, we introduce a state-space framework wherein sample-based signal processing is performed with unitary quantum circuits. We present the quantum delay gate as the quantum analog of the delay line in discrete-time signal processing. Unitary circuits intrinsically describe lossless systems, but we also implement lowpass and other bounded filters by emulating a projection operator. The FIR filter implementation is tested and demonstrated with a quantum circuit simulator.

15. MTVA-Bench: Evaluating the Language Model Inside Cascaded Voice Agents

Authors: Pritish Mishra, Ishaan Kumar, Akshat Mandoli, Sudarshan Kamath

Categories: cs.AI

Generally, most voice agents are cascaded systems, i.e., an ASR model transcribes the caller’s audio, a language model reads the transcript and decides what to say and which backend tools to call, and a TTS model speaks the reply. Nearly all of the decision making happens in the language model, but existing evaluations measure it either too broadly or too narrowly. End-to-end voice benchmarks score the full pipeline, so recognition errors and model errors mix into a single number. LLM benchmarks isolate the model but they do not evaluate what makes real phone calls hard, such as transcription issues, caller’s voice being split across messages and the requirement that replies follow the language and script specified. We introduce the Multi-Turn Voice Agent Benchmark (MTVA-Bench), which evaluates the language model on the same conditions it faces inside a cascaded system. The caller is played by an LLM following a set of rubrics and tool calls are answered by a mock backend which responds to the arguments the model actually sent. The benchmark contains 49 agents working across 490 reviewed scenarios and supports 7 languages. Scoring is a combination of deterministic checks on tool calls with two LLM judges, one that scores scenario specific rules and one that grades conversation quality without access to the task. Both judges must cite specific messages from the transcript. Task and conversation scores are weighted equally, since a call can complete its task and still go badly for the caller. In a seven-model study, six of the models select the correct tool within 6.4 points of one another, but their overall scores span 24.4 points. Most of the gap comes from argument values, action ordering, rule compliance, and what the model says around its tool calls.

16. From Intent to Action: Benchmarking LLM Safety in Vehicle Voice Command Authorization

Authors: Diba Afroze, Xingli Zhang, Yazhou Tu, Xiali Hei

Categories: cs.AI, cs.CL, cs.RO

Large language models (LLMs) are increasingly integrated into vehicle voice assistants. But linking natural-language requests to vehicle functions creates a safety-critical authorization problem. Before executing a command, the system must choose whether to execute, refuse, clarify, require confirmation, defer to manual control, trigger an emergency response, or make no tool call. To our knowledge, prior evaluations do not isolate this pre-action decision across speaker role, authentication status, vehicle state, and tool availability. We introduce a 202-scenario benchmark with Reference Decisions under a seven-class taxonomy. We evaluate two local open-weight models and three API-based LLMs using Decision Alignment and safety-specific error metrics. Alignment ranges from 40.1% for Llama 3.2 3B to 89.1% for Gemini 3.1 Pro Preview. The API-based models score between 83.2% and 89.1%, with no statistically significant differences among them. Even these models produce two to three False Executes among 161 non-execution scenarios, and persistent errors remain in confirmation and manual-control decisions. A controlled Llama 3.2 3B ablation increases alignment to 40.1% under the structured authorization policy, versus 28.2-29.2% under schema-only and generic-safety baselines, but it does not eliminate False Executes. Structured LLM decisions are therefore insufficient as a standalone safety mechanism, and deployment requires an independent enforcement layer that verifies tool permissions and vehicle-state constraints before invoking any vehicle function.

17. Design of the IBM Granite 5.0 TurboCTC ASR Model

Authors: Brian Kingsbury, George Saon, Masayuki Suzuki, Hong-Kwang J. Kuo, Takashi Fukuda et al.

Categories: cs.CL | 5 pages, 2 figures, submitted to ICASSP 2027

We describe the architecture, training methodology and inference speedups of Granite 5.0 Turbo CTC, a 470 million parameter encoder-only model with an excellent speed-accuracy tradeoff. The architecture uses pyramidal temporal subsampling within Conformer blocks using strided depthwise convolutions, block-diagonal (chunk-wise) self-attention, and conditioning on intermediate predictions from the middle layer. Training highlights are the use of only publicly available data, the novel use of a Muon optimizer, and balanced data sampling. Inference speedups include replacing 1 x 1 convolutions with linear layers and optimizing the attention computation in the Conformer blocks. Collectively, these result in a model that is on the speed-accuracy Pareto frontier of the Open ASR leaderboard for English short-form ASR while being twice as fast as the fastest competitor. The model can be used under a permissive license and downloaded from https://huggingface.co/ibm-granite/granite-speech-5.0-470m-turboctc.

18. Dictionary-Constrained Grapheme-to-Phoneme for Unsegmented Languages from LLM-Annotated Data

Authors: Rui Hu, Zhenpeng Zhan, Xiaolong Lin

Categories: cs.CL | Submitted to ICASSP 2027

Grapheme-to-phoneme (G2P) conversion turns raw text into its phonemic form and is an essential part of both text-to-speech (TTS) and automatic speech recognition (ASR) systems. It is required to be fast, stable and context-aware. For unsegmented languages such as Japanese, G2P additionally couples word segmentation with highly context-dependent polyphone disambiguation, and the scarcity of accurately annotated data remains a bottleneck. In this paper, we present a context-aware neural G2P method that scores paths of a discriminative conditional random field (CRF) over a word lattice constructed from dictionaries. To tackle data scarcity, we utilize large language models (LLMs) to generate more than 2 million sentences. Experimental results demonstrate that our method strongly outperforms conventional morphological analyzer-based methods and neural sequence models. On the Joyo-Kanji-Yomi benchmark, our method reaches 99.62% target word reading accuracy, 0.32% target word phoneme error rate (PER) and 0.14% sentence PER.

19. Full-Duplex Speech Models Take the Floor When Asked, Not When Needed

Authors: Linkai Peng, Baorian Nuchged, Kaiqi Fu, Yuyang Yao

Categories: cs.CL, cs.HC | 5 pages

Full-duplex speech models listen and speak at once, promising always-on assistants. Yet they must also decide when they should speak. Human listeners speak when addressed or when the speaker stops, but also self-select to correct a false claim, supply a missing word, or warn of danger. We ask whether full-duplex models do the same. To separate the reason to speak from the opportunity, we construct context-matched English monologues in which only the trigger utterance varies within a topic, define 10 conditions from turn-allocation rules, and compress inter-word pauses to limit opportunities created by silence. Across five model families, being addressed and silence are far more reliable triggers than false facts or hazards. Frame-level text-token probabilities in Moshi and PersonaPlex are lower for false facts than for Neutral when averaged over the first 2\,s after trigger end. Pauses or permission to interrupt do not close this gap either. Given the floor, Moshi and PersonaPlex answer most direct questions, yet the proportion of non-empty false-fact replies that challenge the claim is only .14–.15, and the proportion of hazard replies that warn of danger is .04–.07. This paper thus identifies a gap in both speech initiation and response content. Closing it requires genuine content understanding and intervention decisions grounded in it.

20. AVTrace: Diagnosing Audio-Visual Temporal Reasoning in Omni Models

Authors: Longyin Zhang, Parth Sakhare Mahendra, Chengwei Wei, Ning Zhang, Lim Ming Chong et al.

Categories: cs.CV, cs.AI

Omni models can describe video content, but can they locate events in time, preserve event order, and judge audio-visual synchronization? We introduce AVTrace (Audio-Visual Temporal Reasoning Assessment and Capability Evaluation), a silver-standard diagnostic suite spanning onset and span grounding, synchronization, next-step prediction, cross-modal localization, chain parsing, and event-conditioned comprehension. It contains 34,114 training examples and category-balanced development and test splits of 3,500 and 7,000 examples. We evaluate five open omni models under their respective input configurations using reference-blind response normalization followed by deterministic scoring. All five off-the-shelf systems score below the test split’s majority-label baseline of 0.556 on synchronization verification, and obtain low scores on chain parsing and event-conditioned grounding and comprehension. Development-set perturbations reveal task-dependent sensitivity in Qwen3-Omni-30B to modality removal and changes in visual input processing, without isolating their underlying causes. Parameter-efficient temporal post-training improves Gemma4-E4B-it on several benchmark metrics. On three external image benchmarks, task metrics change modestly, including some degradations, while teacher-forcing perplexity decreases. Together, these findings show that semantic reference-text overlap should not be treated as a proxy for temporal localization, and that AVTrace can identify task-specific weaknesses while providing a testbed for temporal post-training.

音频 · 语音 · 音乐 · 声学