User preference towards HDR10 playback strategies by video player applications
Author(s)
Pooshpanjan Roy Biswas | DXOMARK
Adrien Carmone | DXOMARK
Thibault Cabana
Eloi Zalczer
Paul Bernardi
Frederic Guichard
Abstract
Modern HDR video playback strategies often involve adjustments by industry stakeholders, such as applying different Electro-Optical Transfer Functions (EOTFs), modifying peak luminance, or altering color characteristics, to enhance visual appeal. These modifications, however, can deviate from the original artistic intent. Understanding observer preferences for such alterations is crucial, especially for smartphone playback under diverse ambient lighting.
This study aims to pilot an assessment of user preferences for different HDR playback strategies on smartphones, focusing on EOTF, luminance, and color reproduction. The challenge lies in balancing
these factors to enhance visual comfort and overall experience without compromising visual quality.
A psychophysical experiment employed three HDR10-capable Android video players on identical smartphones within controlled ambient settings (dark, indoor, outdoor). Observers evaluated video playback based on seven criteria, such as overall quality, brightness, skin-tones, contrast etc. . To ensure unbiased results, three-level randomization of video, application, and question order was implemented. Additionally, objective colorimetric data were acquired using an imaging colorimeter.
Preliminary findings from eight observers revealed no significant differences in player ratings in reference conditions. However, differences emerged under low and bright lighting for brightness and contrast preferences, with one player outperforming others in overall quality and color rendering. Further data from 15 observers and deeper colorimetric analysis are forthcoming.
User preference towards HDR10 playback strategies by video player applications
Description
Date and Location: 2/4/2025 | 12:20 PM - 12:40 PM | Grand Peninsula BPrimary Session Chair:
Susan Farnand | Rochester Institute of Technology
Session Co-Chair:
Rafal Mantiuk | University of Cambridge
Paper Number: HVEI-212
Back to Session Gallery