We examined the image quality, image clarity, and viewing comfort of 2-D images rendered on an autostereoscopic display. Opinions on daily use of the content was gathered. Two different stereo displays were used in the experiments-a lenticular lens stereo display employing fixed 3-D stereo and a display with a switchable lenticular lens structure. Images were rendered on the displays with three different rendering schemes. Photos of natural scenes, artificial content, and content containing textual elements were used as the test stimuli. When images with natural scenes or artificial content were categorized into clusters according to the amount of details, significant differences in image quality, image clarity, and viewing comfort scores were observed. When two of the schemes were compared using the images containing textual elements, a significant difference in the viewing comfort and a significant increase in perceived stereoscopic depth impression were found with one of the schemes. Furthermore, image quality and viewing comfort were better with the 2-D display mode than with the 3-D mode. The use of the 2-D text content in the 3-D display mode seemed to be acceptable in general, but for longer term and repeated use, improvements in text quality should be considered. The results indicate that an increase in detail levels may decrease the evaluated image quality, clarity, and viewing comfort. Moreover, for all experimental conditions, better image quality, increased image clarity, and a more comfortable viewing experience had a positive influence on decisions of daily use.
A stereoscopic 3-D version of the film Avatar was shown to 85 people who subsequently answered questions related to sickness, visual strain, stereoscopic image quality, and sense of presence. Viewing Avatar for 165 min induced some symptoms of visual strain and sickness, but the symptom levels remained low. A comparison between Avatar and previously published results for the film U2 3D showed that sickness and visual strain levels were similar despite the films' runtimes. The genre of the film had a significant effect on the viewers' opinions and sense of presence. Avatar, which has been described as a combination of action, adventure, and sci-fi genres, was experienced as more immersive and engaging than the music documentary U2 3D. However, participants in both studies were immersed, focused, and absorbed in watching the stereoscopic 3-D (S3-D) film and were pleased with the film environments. The results also showed that previous stereoscopic 3-D experience significantly reduced the amount of reported eye strain and complaints about the weight of the viewing glasses.
We examined the image quality, image clarity and viewing comfort of the 2D images rendered on a fixed 3D
autostereoscopic display by using three different rendering schemes. Furthermore, opinions on the daily use of the
content were asked. The display used in the study had a lenticular lens as a stereo structure, and both photos of natural
scenes and artificial content were used as the test stimuli. Statistical analysis showed no significant differences between
the different schemes. When images were divided into subgroups according to the amount of details, significant
differences between image quality, image clarity, and viewing comfort scores were found. Moreover, the results indicate
that increase in detail levels may decrease evaluated image quality, clarity and viewing comfort. Finally, for all
experimental conditions better image quality, image clarity and more comfortable viewing experience had some positive
influence on decisions of daily use. In conclusion, amount of details in the content seem to affect on the user experiences
of 2D content shown on a 3D display.
3D or autostereoscopic display technologies offer attractive solutions for enriching the multimedia experience. However,
both characterization and comparison of 3D displays have been challenging when the definitions for the consistent
measurement methods have been lacking and displays with similar specifications may appear quite different. Earlier we
have investigated how the optical properties of autostereoscopic (3D) displays can be objectively measured and what are
the main characteristics defining the perceived image quality. In this paper the discussion is extended to cover the
viewing freedom (VF) and the definition for the optimum viewing distance (OVD) is elaborated. VF is the volume inside
which the eyes have to be to see an acceptable 3D image. Characteristics limiting the VF space are proposed to be 3D
crosstalk, luminance difference and color difference. Since the 3D crosstalk can be presumed to be dominating the
quality of the end user experience and in our approach is forming the basis for the calculations of the other optical
parameters, the reliability of the 3D crosstalk measurements is investigated. Furthermore the effect on the derived VF
definition is evaluated. We have performed comparison 3D crosstalk measurements with different measurement device
apertures and the effect of different measurement geometry on the results on actual 3D displays is reported.
Maturation of the technologies enabling stereoscopic image capture and display offer attractive opportunities for
enriching the user experience in hand-held devices, thus expanding possibilities of stereoscopic content creation
from experienced stereoscopic practitioners to the general public. Content with inappropriate disparities or
artifacts can cause negative experiences. There is therefore a need to ensure the quality of content created with
a stereoscopic camera system to ensure positive user experiences and applicability of stereoscopic techniques in
the mobile domain.
We have investigated how the properties of a stereoscopic camera system can be characterized, and how a range
of aspects in the imaging chain and display properties affect perceived disparities and artifacts in the content.
Design considerations arising from adapting stereoscopic imaging to hand-held devices have been identified. We
propose a methodology for disparity range optimization by basing camera geometry calculations on the limitations
in viewing space extremities. A stereoscopic camera testing system with controllable camera parameters and
a test scene for generating disparities has been developed to assist in evaluating contributing factors and to
confirm the results. As a result, methods and equations for the stereo image calculations are assessed. This
paper presents several techniques for reducing unwanted artifacts in a captured scene and improving the quality
of the user-created stereoscopic content.