OmniCapBench is a new deep‑structured evaluation framework for audio‑visual captioning that replaces free‑form text predictions with sets of atomic, verifiable units across entity references, visual shots, and audio events. It uses deterministic constraint checks and localized LLM‑based semantic comparisons to provide reliable, fine‑grained scoring. With 786 densely annotated videos, the benchmark can distinguish perception errors such as temporal grounding failures, identity drift, cross‑modal misalignment, and hallucinations, revealing that current multimodal large language models excel at local perception but struggle with long‑horizon audio‑visual reasoning.
arXiv:2607. 10299v1 Announce Type: new Abstract: Recent advances in large-scale multimodal models have drivenremarkable progress in vision-language tasks; however, comprehensiveomni-modal understanding remains under-explored, largely due to thescarcity of datasets with rich, explicitly aligned auditory cues.
SonicCaps is a large-scale audio captioning dataset featuring approximately 15 million captions paired with 700,000 audio clips, created using the Qwen3-Omni multimodal language model. The dataset emphasizes diversity by generating around 24 captions per clip through structured prompt engineering and few-shot generation, covering main descriptions, rephrased variants, and semantic tags. Human evaluations rate SonicCaps higher than existing datasets, and training CLAP models on it improves audio retrieval and zero-shot classification across public and commercial benchmarks.
arXiv:2606. 07643v1 Announce Type: cross Abstract: Recent advances in Omni-Multimodal Large Language Models (Omni-MLLMs) have enabled strong integration of vision, audio, and language.
VidOmni-Bench is a new benchmark for fine‑grained video understanding that asks models to verify whether each event in dense video captions is supported by the video. It contains 500 videos covering five complexity types and durations from 4 seconds to 90 minutes, and uses human‑verified sentence‑level labels to create hard negatives. Experiments show that Video‑LLMs often hallucinate events, struggle to detect incorrect descriptions, and exhibit varying weaknesses depending on video complexity and duration.
arXiv:2605. 00873v2 Announce Type: replace-cross Abstract: The rapid advancement of photorealistic Text-to-Video (T2V) generation brings in an urgent need for up-to-date evaluation methods.
Audio Description (AD) provides spoken narration of visual events during dialogue gaps, making movies accessible to visually impaired audiences. The problem requires determining both what (which visua...
OmniHallu is a unified framework for detecting hallucinations in multimodal large language models across both comprehension and generation tasks involving image, video, and audio modalities. It introduces OmniHallu-Bench, a 10,000-sample benchmark with claim-level human annotations for six cross-modal tasks (I2T, V2T, A2T, T2I, T2V, T2A). The system uses a multi‑agent architecture that decomposes outputs into atomic claims, verifies them with modality‑specific experts, and aggregates evidence through structured reasoning, while a preference‑optimized verifier reduces expert calls by 66% with minimal performance loss.
The paper introduces Training-Free Omni (TFO), a plug‑and‑play framework that transforms a frozen vision‑language model (VLM) into a speech‑centric omni model without modifying its architecture or requiring multimodal re‑alignment. TFO leverages Whisper to generate confidence‑filtered, timestamped transcripts and routes them through the VLM’s existing language interface, leaving the visual pathway untouched. Evaluations on 56 benchmarks across 21 languages show that TFO matches or surpasses native omni models on audio‑visual tasks, improves audio‑only performance, and preserves strong visual and reasoning capabilities.
SONIC‑O1 is a new benchmark designed to evaluate multimodal large language models on audio‑video understanding. It contains 60 hours of 231 clips across 13 real‑world conversational domains, with 4,958 human‑verified annotations and demographic metadata. The benchmark tests open‑ended summarization, multiple‑choice question answering, and temporally grounded reasoning, revealing performance gaps between model families and across demographic groups.
arXiv:2607. 27109v2 Announce Type: cross Abstract: With the development of audio large language models (AudioLLMs), audio captioning needs to move from brief descriptions toward open-ended and fine-grained free-form descriptions.
The paper introduces an evaluation framework for structured audio captions that separates acoustic and semantic aspects, such as timestamped sound event descriptions. It covers five axes—tag sets, descriptions, reasoning, numeric measurements, and spectral profiles—using large language model judges for semantics and deterministic metrics for temporal and acoustic features. Controlled perturbations validate that the metrics are robust to paraphrases but sensitive to real semantic and acoustic errors.
The paper introduces Cue2Narrate, a two‑stage pipeline that jointly predicts what visual events to narrate and when to insert the narration in long, untrimmed movie clips. It uses a dual‑head audio‑visual localizer to identify visual cue and narration windows, followed by a LoRA‑adapted vision‑language model that generates concise audio descriptions, trained with a Description Ranking Loss. The authors also present the LongLSMDC benchmark, comprising up to 8‑minute clips, and show that Cue2Narrate outperforms video‑only and audio‑only baselines by 5–12 points in average mAP and improves AD generation over fine‑tuned base VLMs.