AVOC: Enhancing Hour-Level Audio-Video Understanding in Omni-Modal LLMs via Retrieval-Inspired Token Compression

• URL: http://arxiv.org/abs/2606.24286v1
• Date: Tue, 23 Jun 2026 08:06:58 GMT
• Title: AVOC: Enhancing Hour-Level Audio-Video Understanding in Omni-Modal LLMs via Retrieval-Inspired Token Compression
• Abstract summary: AVOC is a framework for long-form audio-video understanding in Omni-modal Large Language Models.<n>AVOC achieves state-of-the-art performance on long-form audio-video benchmarks.
• Score: 22.818321106501788
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multimodal Large Language Models have achieved remarkable progress in short-form audio-video understanding, yet long-form audio-video comprehension remains challenged by limited context windows and severe information redundancy. To address these bottlenecks, we propose AVOC, a framework for long-form audio-video understanding in Omni-modal Large Language Models. AVOC introduces a learnable token compression module between the modality encoders and the LLM backbone. We reframe multimodal token compression as a top-$K$ retrieval problem: given a fixed context budget, the module must retrieve a compact subset of tokens that best supports answering the user query. We draw inspiration from three classical Information Retrieval criteria for selecting informative units from a large candidate pool: relevance, importance, and diversity. AVOC instantiates each criterion as a tailored mechanism for audio-video understanding, and integrates them into a unified retrieval-style compression pipeline. Experiments show that AVOC achieves state-of-the-art performance on long-form audio-video benchmarks, surpassing the second-best model by 4.9 and 5.5 points in average accuracy on OmniVideoBench and LVOmniBench, respectively. Moreover, AVOC maintains robust performance on Audio-Video Needle-in-a-Haystack task at durations up to one hour.

Related papers

• Empowering Long-form Omni-modal Understanding with Robust Audio Perception [48.23485436432239]
  AVDC (Audio-Visual Decoupled Captions) is a large-scale dataset designed to disentangle visual and auditory semantics.<n>AVDC captures both modality-specific nuances and complex cross-modalinteractions.<n>AVDC-QA-CoT is a Chain-of-Thought augmented question-answering dataset to foster audio-visualreasoning.
  arXiv  Detail & Related papers  (2026-07-11T13:04:24Z)
• AVTok: 1D Unified Tokenization for Holistic Audio-Video Generation [47.70199516002865]
  We present textbfAVTok, a novel unified tokenizer designated for holistic audio-video generation.<n>AVTok features a dual-stream transformer-based architecture with shared encoder-decoder and modal-specific learnable queries.<n>Experiments demonstrate that AVTok excels both in audio-video reconstruction and when integrated into downstream pipelines for audio-to-video, video-to-audio, and class-conditional joint audio-video generation.
  arXiv  Detail & Related papers  (2026-06-29T18:35:17Z)
• LVOmniBench: Pioneering Long Audio-Video Understanding Evaluation for Omnimodal LLMs [68.35684758116453]
  This dataset comprises high-quality videos sourced from open platforms that feature rich audio-visual dynamics.<n>We rigorously evaluate the capabilities of OmniLLMs across domains, including long-term memory, temporal localization, fine-grained understanding, and multimodal perception.<n>Open-source models generally achieve accuracies below 35%, whereas the Gemini 3 Pro reaches a peak accuracy of approximately 65%.
  arXiv  Detail & Related papers  (2026-03-19T17:58:13Z)
• AudioMarathon: A Comprehensive Benchmark for Long-Context Audio Understanding and Efficiency in Audio LLMs [53.248502396225724]
  AudioMarathon is a benchmark designed to evaluate both understanding and inference efficiency on long-form audio.<n>We evaluate state-of-the-art LALMs and observe clear performance drops as audio length grows.<n>The results show large gaps across current LALMs and highlight the need for better temporal reasoning.
  arXiv  Detail & Related papers  (2025-10-08T17:50:16Z)
• MAGNET: A Multi-agent Framework for Finding Audio-Visual Needles by Reasoning over Multi-Video Haystacks [67.31276358668424]
  We introduce a novel task named AV-HaystacksQA, where the goal is to identify salient segments across different videos in response to a query and link them together to generate the most informative answer.<n> AVHaystacks is an audio-visual benchmark comprising 3100 annotated QA pairs designed to assess the capabilities of LMMs in multi-video retrieval and temporal grounding task.<n>We propose a model-agnostic, multi-agent framework to address this challenge, achieving up to 89% and 65% relative improvements over baseline methods on BLEU@4 and GPT evaluation scores in QA task on our proposed AVHaystack
  arXiv  Detail & Related papers  (2025-06-08T06:34:29Z)
• Adaptive Audio-Visual Speech Recognition via Matryoshka-Based Multimodal LLMs [33.12165044958361]
  Recent advances in Large Language Models (LLMs) show strong performance in speech recognition, including Audio-Visual Speech Recognition (AVSR)<n>To address this, we propose Llama-MTSK, the first Matryoshka-based Multimodal LLM for AVSR.<n>Inspired by Matryoshka Representation Learning, our model encodes representations at multiple granularities with a single architecture.<n>For efficient fine-tuning, we introduce three LoRA-based strategies using global and scale-specific modules.
  arXiv  Detail & Related papers  (2025-03-09T00:02:10Z)
• STORM: Token-Efficient Long Video Understanding for Multimodal LLMs [116.4479155699528]
  STORM is a novel architecture incorporating a dedicated temporal encoder between the image encoder and the Video-LLMs.<n>We show that STORM achieves state-of-the-art results across various long video understanding benchmarks.
  arXiv  Detail & Related papers  (2025-03-06T06:17:38Z)
• AVS-Mamba: Exploring Temporal and Multi-modal Mamba for Audio-Visual Segmentation [62.682428307810525]
  We introduce AVS-Mamba, a selective state space model to address the audio-visual segmentation task.<n>Our framework incorporates two key components for video understanding and cross-modal learning.<n>Our approach achieves new state-of-the-art results on the AVSBench-object and AVS-semantic datasets.
  arXiv  Detail & Related papers  (2025-01-14T03:20:20Z)
• SAVEn-Vid: Synergistic Audio-Visual Integration for Enhanced Understanding in Long Video Context [19.224601064352846]
  We introduce SAVEn-Vid, the first-ever long audio-visual video dataset comprising over 58k audio-visual instructions.<n>We present AVBench, a benchmark containing 2,500 QAs designed to evaluate models on enhanced audio-visual comprehension tasks within long video.<n>Experiments demonstrate that SAVEnVideo outperforms the best Video-LLM by 3.61% on the zero-shot long video task (Video-MME) and surpasses the leading audio-visual LLM by 1.29% on the zero-shot audio-visual task (Music-AVQA)
  arXiv  Detail & Related papers  (2024-11-25T09:22:13Z)
• Mirasol3B: A Multimodal Autoregressive model for time-aligned and contextual modalities [67.89368528234394]
  One of the main challenges of multimodal learning is the need to combine heterogeneous modalities. Video and audio are obtained at much higher rates than text and are roughly aligned in time. Our approach achieves the state-of-the-art on well established multimodal benchmarks, outperforming much larger models.
  arXiv  Detail & Related papers  (2023-11-09T19:15:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.