3DTV Work Group VQEG meeting Krakow 28 June – 2 July 2010

Goals of 3DTV work group Support standardization activities in terms of subjective assessment of 3DTV QoE: – ITU-R Question 128/6: Digital three-dimensional (3D) TV broadcasting – ITU-T Study Group 9, Question 12 Link with other consortiums promoting the development of 3DTV

3D Video Quality “3D video quality of experience” is more than just visual quality in the 2D sense New types of artifacts compared to 2D: – Issues in capture of stereoscopic signals Geometric distortions, camera miscalibration, … – Binocular rivalry Differences in color, contrast, asymmetric coding, … – View reconstruction Interpolation of occluded areas, … – 3D Rendering issues Crosstalk, color and brightness deficiencies, …

3DTV QoE How can we reliably assess the different quality and physiological aspects of QoE in 3D? – Visual quality – Depth perception – Discomfort – Fatigue – Naturalness

Access to 3D source material Identification of 3D content – Several cameras and producers necessary – IRCCyN has identified second sequences from freely downloadable sequences which would be interesting for subjective testing – Contact these parties and ask if they could open the use to VQEG activities – Identify other potential contributors of 3D content

Short-term goal of the 3DTV group Towards an equivalent of ITU-R BT.500 for 3D video Conduct studies concerning a reliable way to assess the different aspects of 3DTV quality Survey of activities related to human perception/factors in standards-related bodies and international consortiums

Going forward Identify parties that can actively contribute on one or several points: – Content sharing – Tools to process 3D content re-formatting depth/disparity map computation disparity adaptation – 3D equipment and subjective testing

Yonsei´s Contribution Recently, 3D video programs and cinemas are widely available and 3D video quality assessment emerges as an important issue. Furthermore, eye fatigue and safety problems are also important issues. In 3D video, crosstalk is a major factor in determine video quality and eye fatigue problems. Consequently, it should be considered in 3D video quality assessment. To develop 3D video quality assessment methodologies, there are several issues to be addressed, which were not problems in 2D images. Thus, it is proposed that the VQEG 3DTV project should study the following issues to aid the development of 3D video quality assessment methodologies: – 3D display requirements – Subjective assessment methods for 3DTV – Test designs

Proposal Collaborative study on crosstalk Advantages: – Error condition (HRC) is easy to understand – Only a small subset of perceptual degradations (quality of experience) – Different displays can be evaluated – A minimum requirement might be derived for a ITU standard

Proposal Subjective testing using the following methodology: – Adopt the viewing conditions of ITU-R BT.500 as far as possible lighting, viewing distance etc. – Training phase using DSCQS – Test using ACR and binary decision for visual fatigue

Issues to solve on the way Creating a testplan – Editors: Yonsei, FUB, IRCCyN – List of recommended monitors – Screening of subjects (left and right eye, stereo acuity) – Extract as much information from subjects as possible (questionnaires, free-form reports, …) Reproducibility of subjective experiment – Redoing the same setup across labs Optimal 3D viewing conditions – Changing the setup inside the same lab across labs

Call for participation Work items/Schedule: – Selection of SRCs – Preparation of PVSs (next meeting) – Setup of testplan document Who is willing to contribute to this effort? – T-Labs, Intel, Yonsei, AGH, NTT, NIT, Orange-Labs, FUB, Technicolor, IRCCyN, Acreo Wiki, conference calls, mailing list