Welcome to the LIVE Paired Comparison HDR vs. SDR Database
LIVE Paired Comparison HDR vs. SDR Database
Introduction
Although both High Dynamic Range (HDR) and Standard Dynamic Range (SDR) video formats are commonly used today, HDR is generally believed to offer a more engaging viewing experience due to its broader luminance range and expanded color gamut. However, it's still uncertain whether HDR consistently delivers higher visual quality than SDR across typical, real-world display devices. Research exploring the specific conditions under which HDR outperforms SDR in perceived quality remains limited, as does our understanding of how different display characteristics influence the final viewing experience.
To address this gap, we developed the LIVE Paired Comparison HDR vs. SDR Database—a comprehensive collection of 960 videos that includes both HDR and matched SDR versions, each subjected to various levels of compression artifacts. We also carried out a large-scale subjective quality assessment, involving 145 participants who viewed and rated the content across six distinct HDR-capable televisions with diverse performance characteristics.
Our findings reveal that viewers significantly preferred HDR content in scenes with high brightness and vivid colors. Conversely, SDR was sometimes favored for darker, less colorful content—largely due to tone mapping techniques and hardware limitations in HDR displays. We believe that this new dataset will be a valuable asset for training and evaluating video quality assessment (VQA) models designed for HDR and SDR in realistic settings.
Due to copyright restrictions, actual frames from Amazon's live sports videos are not shown. We use screenshots from publicly available open-source sports videos as substitutes.
We are making the open-sourced part of the LIVE Paired Comparison HDR vs. SDR Database available to the research community free of charge. If you use this database in your research, we kindly ask that you reference our papers listed below:
- Chen, B., Lee, C.H., Chen, Y., Shang, Z., Wei, H. and Bovik, A.C., 2025. HDRSDR-VQA: A Subjective Video Quality Dataset for HDR and SDR Comparative Evaluation. arXiv preprint arXiv:2505.21831. [ArXiv]
Download Link Here! Please fill the Google Form to get access to this database.
Database Description
In the released part of the dataset, we provide both reference and encoded videos for open-sourced content in HDR10 and SDR formats. The dataset includes 31 open-sourced videos, totaling 558 video clips, each with corresponding scaled JOD scores.
The dataset also includes subjective scores for 10 Video-on-Demand (VoD) contents and 10 live sports contents. However, due to copyright restrictions, we are unable to make these videos publicly available. The scores for these videos are still included in the released dataset.
Folder Structure:
- Open-sourced HDR10/: Contains open-source videos in HDR10 format.
- Open-sourced SDR/: Contains open-source videos in SDR format.
- JOD_separate.csv: Contains scaled JOD scores with separate anchoring, where both the HDR10 and SDR reference videos are independently assigned a JOD score of 10.
- JOD_together.csv: Contains scaled JOD scores with shared anchoring, using the HDR10 reference video as the sole anchor with a score of 10.
Investigators
The investigators in this research are:
- Bowen Chen ( bwchen@utexas.edu ) -- Graduate student, Dept. of ECE, UT Austin.
- Cheng-han Lee ( chenghan.lee@utexas.edu ) -- Graduate student, Dept. of ECE, UT Austin.
- Yixu Chen ( yixchen@amazon.com ) -- Applied Scientist, Amazon.com, Inc.
- Zaixi Shang ( zaishang@amazon.com ) -- Applied Scientist, Amazon.com, Inc.
- Hai Wei ( haiwei@amazon.com ) -- Principal Research Scientist, Amazon.com, Inc.
- Alan C. Bovik ( bovik@ece.utexas.edu ) -- Professor, Dept. of ECE, UT Austin / Visiting Professor, Dept. of ECE, CU Boulder
Copyright Notice
-----------COPYRIGHT NOTICE STARTS WITH THIS LINE------------
Copyright (c) 2025 The University of Texas at Austin
All rights reserved.
Permission is hereby granted, without written agreement and without license or royalty fees, to use, copy, modify, and distribute this database (the videos, the results and the source files) and its documentation for any purpose, provided that the copyright notice in its entirety appear in all copies of this database, and the original source of this database, Laboratory for Image and Video Engineering (LIVE, http://live.ece.utexas.edu) at the University of Texas at Austin (UT Austin, http://www.utexas.edu
), is acknowledged in any publication that reports research using this database.
The following paper/website are to be cited in the bibliography whenever the database is used as:
- Chen, B., Lee, C.H., Chen, Y., Shang, Z., Wei, H. and Bovik, A.C., 2025. HDRSDR-VQA: A Subjective Video Quality Dataset for HDR and SDR Comparative Evaluation. arXiv preprint arXiv:2505.21831. [ArXiv]
IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT AUSTIN BE LIABLE TO ANY PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OF THIS DATABASE AND ITS DOCUMENTATION, EVEN IF THE UNIVERSITY OF TEXAS AT AUSTIN HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
THE UNIVERSITY OF TEXAS AT AUSTIN SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE DATABASE PROVIDED HEREUNDER IS ON AN "AS IS" BASIS, AND THE UNIVERSITY OF TEXAS AT AUSTIN HAS NO OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
-----------COPYRIGHT NOTICE ENDS WITH THIS LINE------------
Back to Quality Assessment Research page