# More than a Moment: Towards Coherent Sequences of Audio Descriptions

arXiv: https://arxiv.org/abs/2510.25440 · HTML: https://mlsanity.com/paper/2510.25440/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 31 references that are in the map.
- First posted on arXiv 29 Oct 2025.

## Builds on (most cited first, 19 of 19)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [CIDEr: Consensus-based Image Description Evaluation](https://mlsanity.com/paper/1411.5726/) (2014)
- [SPICE: Semantic Propositional Image Caption Evaluation](https://mlsanity.com/paper/1607.08822/) (2016)
- [Movie Description](https://mlsanity.com/paper/1605.03705/) (2016)
- [MAD: A Scalable Dataset for Language Grounding in Videos from Movie Audio Descriptions](https://mlsanity.com/paper/2112.00431/) (2021)
- [AutoAD II: The Sequel -- Who, When, and What in Movie Audio Description](https://mlsanity.com/paper/2310.06838/) (2023)
- [MM-Narrator: Narrating Long-form Videos with Multimodal In-Context Learning](https://mlsanity.com/paper/2311.17435/) (2023)
- [AutoAD III: The Prequel -- Back to the Pixels](https://mlsanity.com/paper/2404.14412/) (2024)
- [Moving Object Segmentation: All You Need Is SAM (and Flow)](https://mlsanity.com/paper/2404.12389/) (2024)
- [AutoAD-Zero: A Training-Free Framework for Zero-Shot Audio Description](https://mlsanity.com/paper/2407.15850/) (2024)
- [Learning Video Context as Interleaved Multimodal Sequences](https://mlsanity.com/paper/2407.21757/) (2024)
- [LLM-AD: Large Language Model based Audio Description System](https://mlsanity.com/paper/2405.00983/) (2024)
- [Shot-by-Shot: Film-Grammar-Aware Training-Free Audio Description Generation](https://mlsanity.com/paper/2504.01020/) (2025)
- [DistinctAD: Distinctive Audio Description Generation in Contexts](https://mlsanity.com/paper/2411.18180/) (2024)
- [Appearance-Based Refinement for Object-Centric Motion Segmentation](https://mlsanity.com/paper/2312.11463/) (2023)
- [What You See is What You Ask: Evaluating Audio Descriptions](https://mlsanity.com/paper/2510.00808/) (2025)
- [FocusedAD: Character-centric Movie Audio Description](https://mlsanity.com/paper/2504.12157/) (2025)

## Built on by (most cited first, 5 of 5)

- [REFRAMED: Towards Realistic Audio Description Generation for Movies](https://mlsanity.com/paper/2608.09765/) (2026)
- [READ More than What You See: Reinforcement Learning for Accurate and Coherent Audio Description Generations](https://mlsanity.com/paper/2606.22766/) (2026)
- [Andha-Dhun: A First Look at Audio Descriptions in Hindi](https://mlsanity.com/paper/2607.06457/) (2026)
- [StrAD: A Streaming Method and Benchmark for Audio Description Generation for Long-form Videos](https://mlsanity.com/paper/2608.12549/) (2026)
- [What, When, and How: Audio Description as Constrained Global Optimization](https://mlsanity.com/paper/2609.30121/) (2026)

## Abstract (arXiv)

Audio Descriptions (ADs) convey essential on-screen information, allowing visually impaired audiences to follow videos. To be effective, ADs must form a coherent sequence that helps listeners to visualise the unfolding scene, rather than describing isolated moments. However, most automatic methods generate each AD independently, often resulting in repetitive, incoherent descriptions. To address this, we propose a training-free method, CoherentAD, that first generates multiple candidate descriptions for each AD time interval, and then performs auto-regressive selection across the sequence to form a coherent and informative narrative. To evaluate AD sequences holistically, we introduce a sequence-level metric, StoryRecall, which measures how well the predicted ADs convey the ground truth narrative, alongside repetition metrics that capture the redundancy across consecutive AD outputs. Our method produces coherent AD sequences with enhanced narrative understanding, outperforming prior approaches that rely on independent generations.