# Revisiting Ground-Truth Synthesis from High-Speed Video: Exact Validity Conditions and an Audited Consumer Capture Corpus

arXiv: https://arxiv.org/abs/2610.05298 · HTML: https://mlsanity.com/paper/2610.05298/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 34 references that are in the map.
- First posted on arXiv 4 Oct 2026.

## Builds on (most cited first, 20 of 20)

- [Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurring](https://mlsanity.com/paper/1612.02177/) (2016)
- [Video Enhancement with Task-Oriented Flow](https://mlsanity.com/paper/1711.09078/) (2017)
- [Noise2Noise: Learning Image Restoration without Clean Data](https://mlsanity.com/paper/1803.04189/) (2018)
- [Pervasive Label Errors in Test Sets Destabilize Machine Learning Benchmarks](https://mlsanity.com/paper/2103.14749/) (2021)
- [Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation](https://mlsanity.com/paper/1712.00080/) (2017)
- [Quadratic video interpolation](https://mlsanity.com/paper/1911.00627/) (2019)
- [XVFI: eXtreme Video Frame Interpolation](https://mlsanity.com/paper/2103.16206/) (2021)
- [Zoom To Learn, Learn To Zoom](https://mlsanity.com/paper/1905.05169/) (2019)
- [Learning to Extract a Video Sequence from a Single Motion-Blurred Image](https://mlsanity.com/paper/1804.04065/) (2018)
- [Realistic Blur Synthesis for Learning Image Deblurring](https://mlsanity.com/paper/2202.08771/) (2022)
- [Bringing Alive Blurred Moments](https://mlsanity.com/paper/1804.02913/) (2018)
- [AceVFI: A Comprehensive Survey of Advances in Video Frame Interpolation](https://mlsanity.com/paper/2506.01061/) (2025)
- [Self-supervised Learning to Bring Dual Reversed Rolling Shutter Images Alive](https://mlsanity.com/paper/2305.19862/) (2023)
- [Efficient video integrity analysis through container characterization](https://mlsanity.com/paper/2101.10795/) (2021)
- [GS-Blur: A 3D Scene-Based Dataset for Realistic Image Deblurring](https://mlsanity.com/paper/2410.23658/) (2024)
- [Benchmarking Video Frame Interpolation](https://mlsanity.com/paper/2403.17128/) (2024)
- [MIORe & VAR-MIORe: Benchmarks to Push the Boundaries of Restoration](https://mlsanity.com/paper/2509.06803/) (2025)
- [IR Motion Deblurring](https://mlsanity.com/paper/2111.11734/) (2021)
- [SloMoDeblur: A Large-Scale Smartphone Image Deblurring Dataset](https://mlsanity.com/paper/2506.19445/) (2025)
- [iPhoneBlur: A Difficulty-Stratified Benchmark for Consumer Device Motion Deblurring](https://mlsanity.com/paper/2605.05990/) (2026)

## Abstract (arXiv)

Motion-deblurring datasets are commonly synthesised by averaging N consecutive high-frame-rate frames and labelling the result with the central frame. We show that this label is unbiased for every capture timing only when the window is odd and the frames' sample durations are equal. An even window shifts every label by a fixed fraction of the blur length, even under perfect timing. Unequal durations are subtler: their mean misalignment is zero, so dataset statistics cannot reveal them, yet when the offset cannot be read from the blur they convolve the supervision rather than adding noise to it. Auditing 51 smartphone clips recorded at a nominal 240 fps, we find 19 captured near 176 fps, a behaviour recorded only in the container's timing tables. In a controlled test the parity choice costs a fitted linear deblurring filter far more than these timing irregularities do, and interpolation labels, unlike blur labels, can be repaired with the true frame times. We provide a tool that checks both conditions without decoding, and will release the clips and their timing tables.