Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2512.02231 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918437384617984 |
|---|---|
| author | Nguyen, Le Thien Phuc Yu, Zhuoran Hang, Samuel Low Yu An, Subin Lee, Jeongik Ban, Yohan Chung, SeungEun Nguyen, Thanh-Huy Maeng, JuWan Lee, Soochahn Lee, Yong Jae |
| author_facet | Nguyen, Le Thien Phuc Yu, Zhuoran Hang, Samuel Low Yu An, Subin Lee, Jeongik Ban, Yohan Chung, SeungEun Nguyen, Thanh-Huy Maeng, JuWan Lee, Soochahn Lee, Yong Jae |
| contents | Multimodal large language models (MLLMs) are expected to jointly interpret vision, audio, and language, yet existing video benchmarks rarely assess fine-grained reasoning about human speech. Many tasks remain visually solvable or only coarsely evaluate speech, offering limited insight into whether models can align who speaks, what is said, and when it occurs. We introduce AV-SpeakerBench, a curated benchmark of 3,212 multiple-choice questions focused on speaker-centric audiovisual reasoning in real-world videos. It features: (1) a speaker-centered formulation that treats speakers-not scenes-as the core reasoning unit; (2) fusion-grounded question design embedding audiovisual dependencies into question semantics; and (3) expert-curated annotations ensuring temporal precision and cross-modal validity. Comprehensive evaluations show that the Gemini family consistently outperforms open-source systems, with Gemini 2.5 Pro achieving the best results. Among open models, Qwen3-Omni-30B approaches Gemini 2.0 Flash but remains far behind Gemini 2.5 Pro, primarily due to weaker audiovisual fusion rather than visual perception. We believe AV-SpeakerBench establishes a rigorous foundation for advancing fine-grained audiovisual reasoning in future multimodal systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_02231 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | See, Hear, and Understand: Benchmarking Audiovisual Human Speech Understanding in Multimodal Large Language Models Nguyen, Le Thien Phuc Yu, Zhuoran Hang, Samuel Low Yu An, Subin Lee, Jeongik Ban, Yohan Chung, SeungEun Nguyen, Thanh-Huy Maeng, JuWan Lee, Soochahn Lee, Yong Jae Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning Multimodal large language models (MLLMs) are expected to jointly interpret vision, audio, and language, yet existing video benchmarks rarely assess fine-grained reasoning about human speech. Many tasks remain visually solvable or only coarsely evaluate speech, offering limited insight into whether models can align who speaks, what is said, and when it occurs. We introduce AV-SpeakerBench, a curated benchmark of 3,212 multiple-choice questions focused on speaker-centric audiovisual reasoning in real-world videos. It features: (1) a speaker-centered formulation that treats speakers-not scenes-as the core reasoning unit; (2) fusion-grounded question design embedding audiovisual dependencies into question semantics; and (3) expert-curated annotations ensuring temporal precision and cross-modal validity. Comprehensive evaluations show that the Gemini family consistently outperforms open-source systems, with Gemini 2.5 Pro achieving the best results. Among open models, Qwen3-Omni-30B approaches Gemini 2.0 Flash but remains far behind Gemini 2.5 Pro, primarily due to weaker audiovisual fusion rather than visual perception. We believe AV-SpeakerBench establishes a rigorous foundation for advancing fine-grained audiovisual reasoning in future multimodal systems. |
| title | See, Hear, and Understand: Benchmarking Audiovisual Human Speech Understanding in Multimodal Large Language Models |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2512.02231 |