LongVQUBench
Benchmarking Long-Term Video Quality Understanding of Vision-Language Models

Nanyang Technological University, Singapore

LongVQUBench features perceptual quality reasoning questions that require integrating visual evidence across extended video durations moving beyond what a single frame can reveal.

GQU · Global Quality Understanding

Which distortion type has the strongest influence on the viewer's global perception of video quality?

•   A.   Spatial sharpen  
• B.   Temporal jitter
• C.   Frame rate inconsistency
• D.   Color banding

The evaluation of long-term video quality understanding remains an open challenge for large vision–language models (LVLMs). Existing video quality benchmarks predominantly focus on short clips and isolated distortions, overlooking the temporal continuity, cumulative degradation, and reasoning complexity inherent in long-duration content. To address these limitations, we present LongVQUBench, a comprehensive benchmark for long-term video quality understanding. LongVQUBench contains over 1,200 diverse videos spanning movies, documentaries, surveillance footage, egocentric recordings, and animated content, accompanied by 1,500 multiple-choice and open-ended questions for validation and testing.

To assess perceptual reasoning across different temporal scopes, we introduce three progressively complex evaluation levels: (i) local event quality understanding (LQU) for analyzing localized distortions; (ii) cross-event quality reasoning (CQR) for integrating multiple degraded events; and (iii) global quality understanding (GQU) for holistic perceptual evaluation over extended durations. Furthermore, a needle distortion question-answering (NDQA) paradigm is embedded across all three levels, where subtle spatial or temporal artifacts are sparsely inserted to probe fine-grained detection and reasoning capabilities. Extensive experiments on 14 state-of-the-art LVLMs reveal significant performance degradation with increasing video length and reasoning depth, highlighting their limited capacity for long-range temporal integration and perceptual attribution.

Hierarchical Evaluation Framework

Three progressively complex levels probe a model's capacity from local distortion detection to global perceptual quality.

Representative questions across the three evaluation levels of LongVQUBench.

Dataset Statistics

Videos sourced from LongVideoBench, MLVU, and LongVideo-Reason-Eval, documentaries, surveillance, vlogs, cooking, animated infographics, and news.

BibTeX

@inproceedings{nema2026longvqubench,
  title     = {LongVQUBench: Benchmarking Long-Term Video Quality Understanding of Vision-Language Models},
  author    = {Nema, Arpita and Zhu, Hanwei and Zhang, Xi and Lin, Weisi},
  booktitle = {European Conference on Computer Vision (ECCV)},
  year      = {2026},
  url       = {https://longvqubench.github.io}
}

📋 Copy BibTeX