AVT-VQDB-UHD-1 - A Large Scale Video Quality Database for UHD-1

This large-scale video quality database presents comprehensive subjective and objective quality assessment of 4K ultra-high-definition videos. Initially published at IEEE ISM 2019, the database consists of short-term videos of segment length similar to DASH segments, based on several short movies that are either publicly available or created by TU Ilmenau.

Author: Rakesh Rao Ramachandra Rao, Steve Göring, Werner Robitza (Audiovisual Technology Group, Technische Universität Ilmenau), Bernhard Feiten (Deutsche Telekom AG), Alexander Raake (Audiovisual Technology Group, Technische Universität Ilmenau)

Partner: No

Tags: , , , , , , ,

Categories:

DOI: 10.1109/ISM46123.2019.00012

Subjective scores: true

Resolution: 360p, 540p, 720p, 1080p, 1440p, 2160p

Method: Subjective quality assessment with MOS scores and confidence intervals

Description

This large-scale video quality database presents comprehensive subjective and objective quality assessment of 4K ultra-high-definition videos. Initially published at IEEE ISM 2019, the database consists of short-term videos of segment length similar to DASH segments, based on several short movies that are either publicly available or created by TU Ilmenau.

The dataset contains four subjective test datasets (test_1 through test_4) with videos encoded using three different video codecs: H.264, HEVC, and VP9. The resolutions of the compressed videos range from 360p to 2160p with frame rates varying from 15fps to 60fps. All source 4K contents use 60fps, maintaining 3840 Ă— 2160 pixel resolution (UHD-1).

The database includes encoded video segments, source videos, subjective ratings presented as Mean Opinion Scores (MOS) with confidence intervals, objective quality scores, and additional metrics including BRISQUE, NIQE, and VMAF reports. This comprehensive collection enables research on codec performance comparison and quality model validation.

This collaborative work between TU Ilmenau’s Audiovisual Technology Group and Deutsche Telekom AG has been highly influential, cited in 38 subsequent publications and downloaded over 900 times. The database has been utilized in multiple QoMEX conference publications, particularly for ITU-T P.1204.3 video quality model evaluation and related video quality assessment research. Five peer-reviewed papers have specifically utilized this dataset between 2019 and 2022.

Access

Openly available for download from https://avtshare01.rz.tu-ilmenau.de/avt-vqdb-uhd-1/ (approximately 55GB)

License

GNU General Public License v3 (code), mixed licenses for video content including Creative Commons Attribution 3.0 (Big Bucks Bunny), Netflix licensing, and CC BY-NC 4.0 for TU Ilmenau content