Judder Visual Quality Assessment Dataset (JVQD)

The JVQD dataset is created to examine the visual quality of videos in presence of judder effect. The dataset consists of 15 reference and 45 compressed omnidirectional video sequences. The compression is performed using ffmpeg (x265 HEVC). The reference videos are named as [videoName]_org.mov and the compressed sequences are named [videoName]_[Bitrate]M.mov.

A single stimulus quality scoring strategy is used for the test to which a total number of 20 naive test subjects participated. Two experiments were designed for quality assessment of the generated videos. The first experiment only focused on the presentation quality evaluation (in terms of compression artefacts). The second test considered both compression and judder effects for quality evaluation. The MOS scores of two subjective experiments are provided in a Microsoft excel worksheet. The test videos are viewed using Oculus Rift HMD with 2160x1200 resolution @ 90 Hz.

• Download the Mos scores (.xlsx)
• Download the dataset (.zip)

The dataset is available to the research community free of charge. If you use this database in your research, we kindly ask to cite the following paper in the bibliography:

• Saeed Mahmoudpour and Peter Schelkens, Visual Quality Analysis of Judder Effect on Head Mounted Displays, 27th European Signal Processing Conference (EUSIPCO), 2019.
• Saeed Mahmoudpour and Peter Schelkens, Omnidirectional Video Quality Index Accounting for Judder, in IEEE Transactions on Circuits and Systems for Video Technology, doi: 10.1109/TCSVT.2020.2971357.

More Details of the subjective testing can be found in the above paper.