International telecommunication union

Download 146.19 Kb.
Date conversion08.07.2018
Size146.19 Kb.
1   2   3   4   5   6

4 Abbreviations and acronyms

This Recommendation uses the following abbreviations and acronyms:

5 Conventions


6 Selection of 3D Source Content

In order to evaluate 3D visual experience and other terms defined in this Recommendation in various circumstances, the content should cover a wide range of 3D videos. In particular, 3D content with a variety of texture, depth and motion should be used for accurate assessment.

The 3D test sequences should be selected according to the goal of the test and recorded on a digital storage system. When the experimenter is interested in comparing results from different laboratories, it is necessary to use a common set of source sequences to eliminate a further source of variation.

The selection of the test material should be motivated by the experimental question addressed in the study. For example, the content of the test sequences (sport, drama, film, etc.) and their spatiotemporal and depth characteristics should be representative of the programmes delivered by the service under study.

6.1 Visual comfort

Excessive disparity/parallax causes visual discomfort possibly because it worsens the conflict between accommodation and vergence. Therefore, it has been suggested that to minimize the accommodation-vergence conflict, the disparities in the stereoscopic image should be small enough so that the perceived depths of objects fall within a “comfort zone”. [Editor’s Note: Further studies required. In particular, how to compute DoF. Subjective tests show the “comfort zone” is delimited by the Depth of Field (DoF) of the human eye. For typical 3DTV viewing conditions, DoF=±0.2 diopters is used as a common threshold.]

The selected stereoscopic test sequences content should be normally comfortable to watch. The visual comfort of stereoscopic images depends critically upon the disparity contained in the image. The visual comfort of stereoscopic images depends also upon the subject’s inter-pupilary distance and the viewing distance.

[Editor’s Note: Further studies required. In particular, how to compute DoF. To reduce bias due to the presence or feeling of visual discomfort on 3D source test material, it would be interesting to calculate the depth range of each scene in diopter unit to know whether the scene is in accordance with the comfort zone requirement. The depth range around the screen plane can be calculated in diopter unit knowing the maximum crossed and uncrossed horizontal disparities and the viewing conditions.]

6.2 Spatial and temporal information

The selection of test scenes is an important issue. In particular, the spatial and temporal perceptual information of the scenes are critical parameters. These parameters play a crucial role in determining the amount of video compression that is possible, and consequently, the level of impairment that is suffered when the scene is transmitted over a fixed-rate digital transmission service channel. Fair and relevant video test scenes must be chosen such that their spatial and temporal information is consistent with the video services that the digital transmission service channel was intended to provide. The set of test scenes should span the full range of spatial and temporal information of interest to users of the devices under test.

6.3 Optional Subjective Methods for 3D Reference Scene Selection: Visual Experience and Visual Comfort requirements

To conduct a subjective test on 3D video, it is desirable to select a set of original scene contents (reference video) with a maximum visual comfort. Preferably, 3D original scene contents should have visual comfort similar to the 2D version of that content, for short viewing durations. In fact, the main goal of 3D video subjective tests is to evaluate the impact of 3D video technologies or image processing algorithms on viewers’ opinion in terms of visual experience, image quality and visual comfort. In case of visual discomfort issues with reference scene contents, visual discomfort may interact with the other quality scales and attenuate related results, resulting on much more difficulties to evaluate the interest of 3D technologies to guarantee an optimal visual experience. Therefore, to ensure a fair comparison of technologies as well as reliable and stable results, reference scene contents should be selected by using the following procedures:

Perform a subjective video quality experiment on the reference scenes (only). Two sessions will be performed: (1) rating visual comfort and (2) rating visual experience. Each session will include the 3D reference, and also the 2D version of that reference (e.g., left eye view).

Select reference sequences where

  • The 3D version has the same level of visual comfort as the 2D version for a short viewing duration (e.g., no significant difference from a statistical point of view); and

  • The 3D version has a higher (i.e., better) visual experience than the 2D version from a statistical point of view.

These 3D reference selection criteria can be obtained from previous subjective tests (e.g., performed by other laboratories). These 3D reference selection criteria can be gathered simultaneously with the target 3D experiment.

Preferably, select scene contents of various video complexities in terms of texture, motion and depth. In fact, to evaluate the impact of the scene content on results, it is important to select original test sequences of different depth levels as well as natural and synthetic ones.

[Editor’s note: Further studies required on threshold values. 6.4 Discrepancies between left and right images

In stereo 3D systems, a binocular 3D image is formed by presenting the left and right image to their respective eye. If discrepancies arise between these two images, they can cause psychophysical stress, and in some cases 3D viewing can fail. For example, when shooting and displaying stereoscopic 3DTV programmes, there may be geometrical, optical, electrical or temporal asymmetries, such as size inconsistency, vertical shift, rotation error, and luminance or color levels between the left and right images. For the production of natural scene contents using two independent video cameras, the main issue is to guarantee the views asymmetries are under perceptual limits. Table 1 illustrates visibility thresholds obtained from subjective experiments using an impairment scale and for a viewing distance of 4.5 times the display height [1].



Visibility threshold

Vertical disparity

Vertical shift difference (local or global)

0.4 %


Rotation difference between the 2 views

0.25 deg.

Focal length

Magnification difference

0.5 %

Black level

Black level difference between the 2 views

3 %

White level

White level difference between the 2 views

10 %


Colorimetry difference considering R, G and B signals

10 %


Temporal asymmetry (shooting or visualization)

To be tested

Table 1: Visibility thresholds related to left and right views asymmetries
[1] ChenWei, PhD Thesis report, Multidimensional characterization of quality of experience of stereoscopic 3D TV, October 2012.]

1   2   3   4   5   6

The database is protected by copyright © 2016
send message

    Main page