محفوظ في:
التفاصيل البيبلوغرافية
المؤلفون الرئيسيون: Zhou, Yang-Hao, Li, Haitian, Lin, Rexar, Huang, Heyan, Zhou, Jinxing, Yuan, Changsen, Lan, Tian, Zhou, Ziqin, Li, Yudong, Xu, Jiajun, Liao, Jingyun, Cheng, Yi-Ming, Chen, Xuefeng, Mao, Xian-Ling, Feng, Yousheng
التنسيق: Preprint
منشور في: 2026
الموضوعات:
الوصول للمادة أونلاين:https://arxiv.org/abs/2602.00607
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
_version_ 1866915971096117248
author Zhou, Yang-Hao
Li, Haitian
Lin, Rexar
Huang, Heyan
Zhou, Jinxing
Yuan, Changsen
Lan, Tian
Zhou, Ziqin
Li, Yudong
Xu, Jiajun
Liao, Jingyun
Cheng, Yi-Ming
Chen, Xuefeng
Mao, Xian-Ling
Feng, Yousheng
author_facet Zhou, Yang-Hao
Li, Haitian
Lin, Rexar
Huang, Heyan
Zhou, Jinxing
Yuan, Changsen
Lan, Tian
Zhou, Ziqin
Li, Yudong
Xu, Jiajun
Liao, Jingyun
Cheng, Yi-Ming
Chen, Xuefeng
Mao, Xian-Ling
Feng, Yousheng
contents Recent advances in text-to-audio-video (T2AV) generation have enabled models to synthesize audio-visual videos with multi-participant dialogues. However, existing evaluation benchmarks remain largely designed for human-recorded videos or single-speaker settings. As a result, structural failures in generated multi-talker dialogue videos, such as identity drift, unnatural turn transitions, and audio-visual misalignment, cannot be effectively diagnosed. To address this issue, we introduce MTAVG-Bench, a failure-driven diagnostic benchmark for multi-talker dialogue-centric audio-video generation. MTAVG-Bench is built via a semi-automatic pipeline, where 1.8k videos are generated using mainstream T2AV models with carefully designed prompts, yielding 2.4k manually annotated QA pairs for fine-grained failure diagnosis. The benchmark evaluates multi-speaker dialogue generation at four levels: audio-visual signal fidelity, temporal attribute consistency, social interaction, and cinematic expression. Built on a hierarchical failure taxonomy and a targeted QA protocol, MTAVG-Bench is primarily designed to evaluate whether proprietary and open-source omni-models can reliably identify failure modes in multi-speaker T2AV outputs. We benchmark 12 proprietary and open-source omni-models on MTAVG-Bench, with Gemini 3 Pro achieving the strongest overall performance, while leading open-source models remain competitive in signal fidelity and consistency. Overall, MTAVG-Bench enables fine-grained failure analysis for rigorous model comparison and targeted video generation refinement.
format Preprint
id arxiv_https___arxiv_org_abs_2602_00607
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MTAVG-Bench: A Diagnostic Benchmark for Multi-Talker Dialogue-Centric Audio-Video Generation
Zhou, Yang-Hao
Li, Haitian
Lin, Rexar
Huang, Heyan
Zhou, Jinxing
Yuan, Changsen
Lan, Tian
Zhou, Ziqin
Li, Yudong
Xu, Jiajun
Liao, Jingyun
Cheng, Yi-Ming
Chen, Xuefeng
Mao, Xian-Ling
Feng, Yousheng
Multimedia
Sound
Recent advances in text-to-audio-video (T2AV) generation have enabled models to synthesize audio-visual videos with multi-participant dialogues. However, existing evaluation benchmarks remain largely designed for human-recorded videos or single-speaker settings. As a result, structural failures in generated multi-talker dialogue videos, such as identity drift, unnatural turn transitions, and audio-visual misalignment, cannot be effectively diagnosed. To address this issue, we introduce MTAVG-Bench, a failure-driven diagnostic benchmark for multi-talker dialogue-centric audio-video generation. MTAVG-Bench is built via a semi-automatic pipeline, where 1.8k videos are generated using mainstream T2AV models with carefully designed prompts, yielding 2.4k manually annotated QA pairs for fine-grained failure diagnosis. The benchmark evaluates multi-speaker dialogue generation at four levels: audio-visual signal fidelity, temporal attribute consistency, social interaction, and cinematic expression. Built on a hierarchical failure taxonomy and a targeted QA protocol, MTAVG-Bench is primarily designed to evaluate whether proprietary and open-source omni-models can reliably identify failure modes in multi-speaker T2AV outputs. We benchmark 12 proprietary and open-source omni-models on MTAVG-Bench, with Gemini 3 Pro achieving the strongest overall performance, while leading open-source models remain competitive in signal fidelity and consistency. Overall, MTAVG-Bench enables fine-grained failure analysis for rigorous model comparison and targeted video generation refinement.
title MTAVG-Bench: A Diagnostic Benchmark for Multi-Talker Dialogue-Centric Audio-Video Generation
topic Multimedia
Sound
url https://arxiv.org/abs/2602.00607