This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. IEEE TRANSACTIONS ON BROADCASTING 1 Quality of 8K Ultra-High-Definition Television Viewing Experience in Practical Viewing Conditions Yoshiaki Shishikui , Senior Member, IEEE, and Yasuhito Sawahata , Member, IEEE Abstract—In this study, we examine the psychological effects of 8K ultra-high-definition television (UHDTV) under daily TV viewing conditions. The design of 8K-UHDTV targets the realization of the “ultimate 2D image.” 8K-UHDTV is already being used practically and the audience have reported experiencing a feeling of reality and presence. However, the benefits offered by 8K for daily TV viewing and the degree to which the design goals have been achieved have not been sufficiently verified. Hence, in this study, we verify them by investigating the popular impressions evoked by familiar images and their modulation with the viewing distance through subjective evaluation experiments and statistical analysis. The obtained results indicate that the impressions are enhanced by increasing the resolution from 4K to 8K, even at viewing distances beyond the design viewing distance, which are conventionally considered less effective. Thus, they confirm that 8K provides strong impressions over a wide viewing range. Furthermore, the relationship between higher-order impressions, such as “being-there” and “reality,” and lower-order impressions is derived through multiple regression analysis. This provides insight into the enhancement of higher-order impressions in production and post-processing. Index Terms—8K, impression, psychological effect, subjective evaluation, UHDTV. I. I NTRODUCTION UALITY evaluation is indispensable not only in system development but also in understanding the value that the developed system can offer. Although systems often undergo quantitative evaluation in their developmental stages, the value provided by the system, such as the effects and merits, are not completely evaluated. When a system provides a new value, an appropriate method is required to evaluate it. For example, 8K ultra-high-definition (UHD) TV [1], designed based on the perceptual limits of human vision, pursues the ultimate quality in 2D TV and offers more value than expected in the Q Manuscript received May 10, 2021; revised July 20, 2021; accepted July 21, 2021. This work was supported in part by Meiji University through Research Project Grant (B) by the Institute of Science and Technology. (Corresponding author: Yoshiaki Shishikui.) This work involved human subjects or animals in its research. Approval of all ethical and experimental procedures and protocols was granted by Ethics Review Committee on Research with Human Subjects, Meiji University. Yoshiaki Shishikui is with the School of Interdisciplinary Mathematical Sciences, Meiji University, Tokyo 164-8525, Japan (e-mail: sisikui@meiji.ac.jp). Yasuhito Sawahata is with Science and Technology Research Laboratories, NHK (Japan Broadcasting Corporation), Tokyo 157-8510, Japan (e-mail: sawahata.y-jq@nhk.or.jp). Digital Object Identifier 10.1109/TBC.2021.3105031 system design. However, because evaluation methods for the offered value have not been established, quantitative evaluation has been insufficient. Therefore, it has not been verified whether 8K-UHDTV offers the expected value by achieving the ultimate 2D TV quality. In this study, we propose a method for evaluating the quality of experience (QoE) for 8K-UHDTV viewing, and quantitatively verify the value of 8K under practical viewing conditions as found in a home-use situation. UHDTV design aims to create high-quality visual experiences—such as a strong sensation of reality, presence, and high-transparency to the real world—for various applications ranging from small screen (e.g., a tablet device) to medium screen (e.g., a screen in a living room) to large screen (e.g., a theater) [2]. To realize this design concept, the 8K-UHD system parameters were determined based on psychophysical evidence of the perceptual limits of human vision. Specifically, the number of pixels was determined through subjective evaluation experiments on the sense of “being-there” [3]–[5], and the frame rate was determined based on the perception of motion blur, stroboscopic effects, and flicker [6], [7]. Many 8K-UHDTV video viewers reported experiencing strong psychological effects, such as “I could almost smell and feel the fireworks at the opening ceremony” [8], [9]. In addition to the sense of “being-there” and “reality,” which were considered during the design, viewers of UHD images have often reported strong impressions such as “beautiful,” “powerful,” and “delicious.” Previous studies have suggested that the viewing environment, such as the viewing distance and screen size, contributed to these impressions [10]. As it can be assumed that the viewing distance and position vary significantly, particularly when viewing in a living room, these findings also suggest that the producer’s intentions represented in the video content may not be appropriately delivered to the viewers, depending on the viewing conditions. Most previous image quality evaluations focused on the overall fidelity, quality degradation, and preference [11]–[13]. For example, HD image quality has never reached the ultimate level, and the viewers can still perceive differences between the displayed images and real objects. Thus, evaluation scores with the fidelity criterion did not saturate, and were sufficient for comparing the performances of various developments. However, as the UHDTV video system almost achieves This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. 2 the ultimate quality, fidelity-based evaluation is insufficient. Therefore, a new scheme is required for evaluating UHDTV system performance. Impressions can be categorized into two groups: higherorder and lower-order. In this study, we define lower-order impression as the feelings related to image features, such as the sharpness, feeling of depth, and vividity, whereas the higher-order impressions are defined as feelings with more cognitive qualities, such as “beautiful,” “powerful,” and the “sense of being-there.” QoE-based evaluation, which focuses on the higher-order concepts of the user experience, has received considerable attention for application in evaluating media performance. As mentioned previously, UHDTV viewers often report higher-order impressions; hence, QoE-based evaluation should focus on quantifying the quality of the viewing experience in UHDTV. Thus, evaluation using such familiar impressions should elucidate the new values delivered by UHDTV, provide better understanding of the potential of UHDTV services, and enable optimization at each stage of the UHDTV ecosystem. If the relationship between UHD image characteristics and such higher-order impressions is clarified, video producers can realize better cinematographic techniques to deliver their intentions. We previously proposed a method to evaluate the psychological effects of viewing ultra-high-resolution images on familiar impressions and conducted subjective evaluation with 4K-UHDTV images [14]. Images with different resolutions were presented to viewers and their ratings for the higherand lower-order impressions were obtained. These ratings were statistically analyzed, and the relationship between the higher- and lower-order impressions was examined assuming a hierarchical impression structure. This scheme facilitated an understanding of the latent relationship between higherand lower-order impressions. Another study [15] also followed a similar scheme and attempted to elucidate the relationships. With this scheme, we found that the increase in the resolution of the presented images enhanced the impressions, even when the viewing distance was four times more than the design viewing distance. The results of multiple regression analyses assuming a hierarchical structure for the lowerand higher-order impressions provided insights on impression enhancement. In this study, utilizing this methodology, we empirically examine the psychological effects of viewing 2K, 4K, and 8K images at a viewing distance of 1.5H (H: screen height) (1.5 times greater than the screen height). We evaluate the familiar impressions for the images and demonstrate the advantage of 8K over 4K and 2K systems in terms of the QoE (Fig. 1). Moreover, we examine the effect of the viewing distance on familiar impressions to verify the effective viewing area for 8K-UHDTV viewing. Viewing distances of 0.75H, 1.5H, and 3.0H were employed, providing field-ofviews (FOVs) of approximately 100◦ , 60◦ , and 30◦ , respectively, as shown in Fig. 2. Through these investigations, we establish that 8K-UHDTV can enhance the effect of familiar impressions, and the effect can be delivered at various viewing distances, which include the everyday viewing styles. In addition, through regression analysis, we show that the degree of IEEE TRANSACTIONS ON BROADCASTING Fig. 1. Experimental setup. Fig. 2. Geometry of the experimental setup. contribution of each lower-order impression differs depending on each higher-order impression. Particularly, we found that the sense of “being-there” and “reality” were yielded by significantly different contributions of the lower-order impressions. In this paper, we quantitatively verify the strong psychological effects of 8K viewing in terms of various familiar impressions. In addition, we present new findings on the viewing range and impression factors that elicit psychological effects. The remainder of this paper is organized as follows. Section II introduces the related work in this research area. Section III describes the experimental method. Section IV presents the experimental results, and Section V analyzes them. Section VI presents concluding remarks. II. R ELATED W ORK This section reviews previous work, based on which the resolution parameters of 8K-UHDTV were determined, and introduces various studies on the evaluation of UHDTV. The UHD system parameters were determined through psychophysical experiments that focused on the immersive aspects of the viewers’ feelings. For example, the FOV angle and screen aspect were determined based on subjective evaluation experiments. Hatada et al. [3] characterized the relationship between the FOV angle and the feeling of “being-there” as the inductive effect of a wide FOV image presentation on the This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. SHISHIKUI AND SAWAHATA: QUALITY OF 8K UHDTV EXPERIENCE viewers’ posture movements, and showed that the sense of “being-there” asymptotically increased with the FOV angle, up to 80◦ or 100◦ . Masaoka et al. [5] conducted experiments on the sense of “being-there” using images with two different shooting field angles and obtained results similar to those of Hatada et al. [3] for both angles. According to Narita et al. [16], the preferable screen aspect differs depending on the screen area, regardless of the FOV; more specifically, the sense of comfort is influenced by the screen aspect but the sense of potency is not, when the screen area remains the same. These results suggest that the viewers’ feelings of immersion require a wide-viewing-angle presentation. The resolution of the UHDTV system was determined based on human visual acuity to provide a wide-viewing angle presentation with sufficient image quality. ITU-R recommends a design viewing distance of 3H for HDTV and 6H for SDTV [17]–[19]. The image resolution at these viewing distances is approximately 30 cycles per degree (cpd) or one pixel per arc minute, which corresponds to the separable visual angle at a normal visual acuity of 20/20 in Snellen notation. In this condition, it has been assumed that a person with normal visual acuity cannot discern the pixel structure at the design viewing distance. According to this concept, to provide a horizontal FOV angle of 100◦ , the horizontal pixel count must be approximately 8000 and the design viewing distance must be 0.75H. This is the initial design of 8K-UHD. In the case of 4KUHD, the design viewing distance is 1.5H [20]. Therefore, the viewing distance beyond the design viewing distance is considered to exceed the perception limit, implying that viewing tests under this condition have little meaning. Recent studies have demonstrated the significant psychological effects of UHD resolution under various viewing conditions. Masaoka et al. [21] investigated the relationship between the angular resolution and “sense of reality.” Their results showed that the “sense of reality” increased with the angular resolution, even though it reached saturation above approximately 60 cpd (=120 pixels/degree); i.e., the sense of reality was saturated at twice the design viewing distance. Tsushima et al. [22], [23] examined the relationship between the spatial resolution of a visual image and the depth sensation using primitive stimuli defined by gradual luminance-contrast changes [22] and Gabor patches [23]. They established that higher-resolution images (above 60 cpd) produced a stronger depth sensation, even when the enhancement in resolution was not noticed. Lachat et al. [24] investigated the relationship between the viewing distance and perceptual quality for UHD images. Their results showed that the image quality (i.e., the goodness of the presented picture) at 4K resolution was considerably better than that of the HD at a viewing distance of 1.5H, which is the design viewing distance for 4K, and slightly better even at a viewing distance of 3H [24]. Although these studies only focused on the overall image quality (goodness) and not on the other impression factors, the effects provided by UHD images are similar to those in [21]–[23]. These studies suggest that UHD resolution can have various psychological effects on the viewers, even beyond the design viewing distance. Park et al. performed paired comparison between 8K and 4K images on 65-inch displays at 3 a viewing distance of 2.74 m [15]. They established that 8K resolution images were perceived to be better with respect to all the assessment items, except “brightness” and “fatigue.” UHDTV features not only higher resolution but also a wide color gamut and high dynamic range (HDR). Hanhart et al. investigated the preference of the peak luminance level of the display and showed that among the options, the highest peak luminance was the most attractive [25]. Bist et al. indicated that brightness preference was highly content-dependent, particularly, on the number of bright pixels in the content [26]. Noland et al. attempted to derive the perceptive brightness of HDR images from the pixel values and showed that the most effective metrics tested were the mean of the pixel luminance values [27]. To the best of our knowledge, the relationship between the dynamic range and the various impressions has not been quantitatively investigated. We previously analyzed the impression enhancement effects through higher resolution and higher dynamic range, and determined that high resolution and HDR were individually or collectively effective for enhancing a certain impression [28]. An increasing number of studies have been focusing their evaluation over the range from low-level image quality to high-level experiences such as impressions [29]. Most previous studies focused on lower-order impressions. For example, Van Wallendael et al. [30] focused on a lower-order impression, namely, the feeling of sharpness, and compared it between upscale HD and UHD. In psychology studies, several works have focused on higher-order impressions, such as the aesthetics. For example, Palmer et al. [31] focused on the preferences for artworks and investigated the relationship between the visual features (color, brightness, and composition) and preferences. However, the relationships between the lower- and higher-order impressions were not investigated. Furthermore, the study did not consider viewing conditions, such as the image resolution and viewing distance. III. E XPERIMENTS We conducted subjective evaluation experiments in which the subjects viewed 2K, 4K, and 8K resolution test images at viewing distances of 3H, 1.5H, and 0.75H, and rated them with respect to the given evaluation terms. Table I lists the experimental conditions, including the test images, apparatus, room conditions, and evaluation terms. The viewing conditions conform to the ITU-R BT.500 laboratory environment [32], [33]. A. Test Images The test images included foods and outdoor scenes, as shown in Fig. 3. We selected test image subjects that would elicit viewer impressions such as “beautiful” and “delicious.” The images were captured using a camera (Canon EOS 5DS with 8688 x 5792 pixels) and recorded in RAW format. The images were then developed using the gamma and color gamut of ITU-R BT.2020 with 10 bits-per-pixel precision and trimmed into 8K format (7680 x 4320 pixels). We created 4K- and 2K-equivalent images from the 8K images by applying a low-pass filter (LPF) that mimics the modulation transfer function (MTF) characteristics of a 4K This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. 4 IEEE TRANSACTIONS ON BROADCASTING TABLE I E XPERIMENTAL C ONDITIONS Fig. 4. Frequency response and impulse response of the LPF. Fig. 5. Presentation of the test images. domain. Fig. 4 displays the MTF and impulse response of the used LPF. B. Experimental Procedure Fig. 3. Test images. and 2K camera, respectively. The original 8K images as well as the 4K- and 2K-equivalent images were then displayed on an 8K display. Specifically, we designed an LPF for generating 2K-equivalant images based on the MTF characteristics of an actual 2K camera reported by Nakamura et al. [34]. The LPF was designed to satisfy the condition that the MTF response remains greater than 0.35 at the 800 TV line, complying with the condition that must be satisfied by TV cameras in professional productions. Similarly, we designed an LPF for generating 4K-equivalent images based on MTF characteristics twice that of the 2K camera in the frequency A total of 25 individuals (aged 20–22 years) with normal or corrected-to-normal visual acuity participated in our experiment. Eighteen subjects participated in each session of the trial series performed with different viewing distances. The number of subjects in each session included some randomly chosen subjects who participated in multiple sessions. No postscreening was conducted. Note that ITU-R BT.500 [32] states that “at least 15 observers should be used.” The visual acuity was tested using the Tumbling E eye chart. Informed consent was obtained from the subjects prior to the experiment. For the 1.5H and 3H viewing distance conditions, two subjects viewed the presented images simultaneously, whereas one subject viewed the images for a viewing distance of 0.75 (Fig. 2). In one of the trial series, images with eight different content types and three different bandwidths were randomly presented to prevent the image presentation order from affecting the ratings. In addition to the eight (image content) × three (bandwidth) stimuli, two dummy stimuli were included and placed at the beginning of the sequences to remove initial instability. A mid-gray screen was presented 5 s before each trial as shown in Fig. 5. During the 30-s presentation, the subjects viewed This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. SHISHIKUI AND SAWAHATA: QUALITY OF 8K UHDTV EXPERIENCE 5 Fig. 7. Fig. 6. Score sheet. the images and input the scores based on nine impressions (Fig. 6) with a seven-grade scale that ranged from “strongly disagree” to “strongly agree.” An iPad and an Apple pencil were used as score input devices. For the food images, “beingthere” was excluded, and for the outdoor scene images, “fresh” and “delicious” were excluded. In typical quality assessments in picture coding studies, participants are asked about the overall image quality. Contrastingly, we did not ask about the image quality because the focus of our study is the effects of different video systems on everyday impressions. We adopted a seven-grade scale for evaluation because the subjects in our preliminary experiment reported having difficulties in expressing their impressions using five grades and preferred seven grades to describe the absolute scores. This study investigates the extent to which the common impressions in our daily life are altered by the differences in the image resolution (i.e., whether the impression can be strengthened by increasing the resolution). We selected terms such as “looks beautiful,” “looks delicious,” “looks fresh,” “looks real,” and “being-there” as the evaluation terms for expressing popular values. These are evaluation terms of higher-order impressions, which were obtained by combining various image features. The evaluation terms for lower-order impressions, such as high resolution, color vividity, glossiness, and three-dimension (3D), were also selected and used in the experiments for analyzing their relationship with the physical features. C. Analysis We calculated the average values of the subjective evaluation experimental results and their 95% confidence intervals based on the t-distribution. A t-test was performed to determine whether there were significant differences in the impressions among the resolutions. To directly examine the effect of high-resolution images for each type of content and viewing distance on each impression, we employed multiple t-tests to compare the impression scores between adjacent resolutions (i.e., between 8K and 4K and between 4K and 2K) for each type of content and viewing distance. Performing Hierarchical structure of the impressions. statistical tests multiple times causes a multiple comparison problem leading to false positives because of the increase in the practical significance level; hence, we adjusted the false discovery rates (FDRs) for each p-value using the Benjamini and Hochberg (BH) method [35]. The strength of the impression may vary depending on the image content and viewing distance. To examine the effect of these factors, we performed the two-way analysis of variance (ANOVA2) process for the scores of the 8K, 4K, and 2K resolutions, respectively. We further performed multiple comparison testing on the main effects. To examine the relationships between the lower- and higherorder impressions, we executed multiple regression analyses on the obtained rating scores for cases where a significant difference was observed in the higher-order impressions, regardless of the images or viewing distances. We assumed a linear model with a hierarchical causal relationship between the lower- and higher-order impressions with respect to resolution, as depicted in Fig. 7; i.e., the increase in resolution enhances the four lower-order impressions, which in turn reinforces the higher-order impressions. The linear model is in line with psychophysical studies that suggest the human visual system has a function called “visual attention” that selects a subset from all the available visual information, such as area in visual fields and specific visual features, on a top-down basis and improves the related perceptional sensitivities [36]. Therefore, asking about higher-order impressions drove the attention function in viewers, and then the related lower-order visual features were selected based on viewers’ prior knowledge and experiences and used in the subsequent higher-order cognitive processing. To compute the degree of contribution of each lower-order impression to each higher-order impression, we used multiple regression analyses in which each higher-order impression was explained by the weighted sum of the four lower-order impressions. The estimated weight values (partial regression coefficients) for each lower-order impression can be interpreted as the degree of contribution to the corresponding higher-order impression. IV. R ESULTS A. Impression Strength for Various Image Resolutions Here, we present the impressions for 8K images at viewing distances of 1.5H and 3H. Because the results at a viewing This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. 6 IEEE TRANSACTIONS ON BROADCASTING Fig. 8. Comparisons of the lower-order impressions for each image at different resolutions (1.5H). Fig. 9. Comparisons of the high-order impressions for each image at different resolutions (1.5H). distance of 0.75H had a similar tendency, which is statistically confirmed further in Section IV-B, they are not depicted in this section. Fig. 8 shows the results of the lower-order impressions (e.g., 3D effect, resolution, color vividity, and glossiness) for 8K, 4K, and 2K images at a viewing distance of 1.5H, wherein an asterisk (*) indicates a difference at the 5% significance level (FDR corrected), observed between images with adjacent resolutions. Fig. 9 displays the results of the higherorder impressions (e.g., beautiful, delicious, fresh, reality, and “being-there”) at a viewing distance of 1.5H. The results show significant differences between the 8K, 4K, and 2K images in many cases, among combinations of the impression terms and image content. Between 8K and 4K, approximately 40% of the cases showed significant differences. The following observations were derived from Figs. 7 and 8: This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. SHISHIKUI AND SAWAHATA: QUALITY OF 8K UHDTV EXPERIENCE Fig. 10. 7 Comparison of the higher-order impressions for each image at different resolutions (3H). 1) Differences among the 8K, 4K, and 2K images were observed both for lower-order and higher-order impressions. 2) Effects of the image resolutions on the impressions were image-dependent: greater for “fruit” and “hanabi” but weaker for “nikuman” and “lady.” 3) The average impression scores for 8K images reached the “agree” level. Almost the same observations as those mentioned above were obtained for the results at a viewing distance of 0.75H. Fig. 10 shows the results for the higher-order impressions at a viewing distance of 3H. Although 3H has as much as four times the design viewing distance of 8K and the number of cases with significant differences decreased, we still observed a similar tendency to which the average scores improved as the image resolution increased, as found at the viewing distances of 0.75H and 1.5H. B. Effect of Viewing Distance We determined whether significant differences occurred due to the viewing distance and image content, and their interaction for each image and impression term through the ANOVA2 process; the results are summarized in Table II, where “*” indicates p <0.05 and “**” indicates p < 0.01. The p-values were corrected for multiple comparisons. Fig. 11 summarizes the effect of the viewing distance on “resolution,” “real,” and “being-there.” For 8K resolution, the image content factor was effective for all the impressions, except “delicious” and “being-there,” whereas the viewing distance was effective only for “beingthere,” as shown in Table II(a). No interaction between the two factors was observed. The results of the multiple comparison test indicated that the sense of “being-there” at Fig. 11. Results of the multiple comparison test for the effect of the viewing distance. 3H was significantly lower than those at 0.75H and 1.5H (Fig. 11 (a)). For 4K resolution, the image content factor was effective for all the impressions, except “delicious” and “being-there,” whereas the viewing distance was effective for “threedimensional,” “high-resolution,” “beautiful,” and “real,” as shown in the “Content” column in Table II(b). No interaction was observed. The results of the multiple comparison test indicated that the score at 0.75H was significantly lower than the others for “high-resolution” and “real” (Fig. 11 (b)). This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. 8 IEEE TRANSACTIONS ON BROADCASTING TABLE II R ESULTS O F ANOVA2 Fig. 12. Degree of contribution of lower-order impressions to the determination of higher-order impressions (ratio of the absolute value of the t- statistics obtained for each partial regression coefficient). For 2K resolution, the image content factor was effective for all the impressions, except “real” and “being-there,” as shown in Table II(c). The viewing distance was effective for all the impressions, except “glossy,” “fresh,” and “delicious.” No interaction was observed. The results of the multiple comparison test indicated that the scores at 1.5H and at 0.75H were significantly lower than that at 3H (Fig. 11 (c)). C. Regression Analysis Based on the multiple regression analyses using the hierarchical impression model shown in Fig. 7, considerable differences were observed in the degree of contribution of each lower-order impression to each higher-order impression, as shown in Fig. 12. The contributions were computed based on the absolute values of the t-statistics with 5% significance level, obtained through multiple regression analysis for each higher-order impression term, as shown in Table III. The following observations were derived from Table III and Fig. 12: 1) The coefficient of determination, R2 , was approximately 0.7 for “beautiful,” “delicious,” “fresh,” and “real,” whereas it was slightly below 0.5 for “being-there.” These indicate the precision of the regression model for each higher-order impression. 2) The contribution of the lower-order impressions differed depending on the type of higher-order impression. 3) The sense of “high resolution” was the main contributor for “real,” “being-there,” and “beautiful.” V. D ISCUSSION The experimental results revealed that 8K-UHDTV enhanced both higher- and lower-order impressions more than 4K-UHDTV, and the impression scores for 8K were almost close to the maximum value in the evaluation range. Significant differences in the strength of higher-order impressions between 8K and 4K were observed at a viewing distance of 3H, which is four times that of the design viewing distance. This result is consistent with the previous experimental result with 4K [14] with respect to the angular resolution, although the FOVs differ from those of this previous study; i.e., the increase in angular resolution observed at 6H when the image resolution was updated from 2K to 4K, examined in the previous study, is equivalent to that observed at 3H when it was updated from 4K to 8K, examined in this study. The advantage of 8K was confirmed at a considerably greater viewing distance than previously presumed. The ANOVA2 results for 8K resolution [Fig. 11(a)] indicated that the viewing distance altered the strength of the “being-there” impression, whereas all the other impressions did not. This allows viewers to choose their preferred viewing distance without weakening the impression, which satisfies This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. SHISHIKUI AND SAWAHATA: QUALITY OF 8K UHDTV EXPERIENCE 9 TABLE III R ESULTS O F M ULTIPLE R EGRESSION A NALYSES Fig. 13. Comparison of the strength of the higher-order impressions among 8K, 4K, and 2K viewed from each design viewing distance. the original design concept of 8K-UHDTV for living room viewing [2]. These results are supported by the previous work by Emoto and Sugawara that investigated viewer behavior for the preferred viewing distance when watching an 8K program on an 85-inch display [37]. Their results demonstrated that viewers change the viewing distance according to the image content and narration, and that the viewing distance for 8K was smaller than that for 2K. The mean of the viewing distance statistics was 132.5 cm (1.25H) and the standard deviation was 53.2 cm (0.5H). At times, the viewers moved close to the display to observe the details or moved farther to view the whole image. Although viewing at 0.75H was previously believed to be the best in 8K viewing, our results suggest that the best viewing position, or “sweet spot,” is not limited but in a much wider area including the above preferred viewing distances. Therefore, the impression enhancements in the wide coverage of the preferred viewing area show the proof of concept of the 8K-UHDTV that provides a feeling of reality or presence through ultimate 2D images with various viewing styles. The impression of “being-there” for 8K images decreased at 3H, compared to those at 0.75H and 1.5H. As suggested by Hatada et al. [3], who investigated the effect of the FOV on the related impression, we believe that the FOV was too small because the distance provided only 30◦ FOV. As shown in Fig. 8, the score for the “being-there” impression for 2K and 4K was lower than that for 8K, although the FOV angle was the same. This indicates that the resolution is also an important factor for the “being-there” impression. Re-sorting all the experimental data (most of the data are shown in Figs. 9 and 10) with respect to the higher-order impressions and viewing distances suggests that the impressions decrease as the FOV is narrowed (the viewing distance is lengthened), as shown in Fig. 13. This figure compares the scores for the higher-order impressions among 8K, 4K, and 2K images, when viewed at each design viewing distance, namely, 0.75H for 8K, 1.5H for 4K, and 3H for 2K. The bar length indicates the average score for all the image contents. The angular resolutions are the same, regardless of the image resolution, but the FOVs differ. This figure shows that not only “being-there,” but all the other impressions are also affected by the FOV. Multiple regression analyses establish that the contribution of the lower-order impressions to each higher-order impression differed depending on the type of higher-order impression. Notably, the sense of “high resolution” was not always the main contributor to the higher-order impressions, even though the physical factor was the resolution. This result is consistent with that of a previous study [14], but the contribution ratio is slightly different owing to the difference in the image content. For example, “nikuman” originally had no color; hence, increasing the resolution did not enhance the sense of “vividity,” and it had negligible effect on the higherorder impressions. For the sense of “being-there,” which was not considered in the previous study, the contribution of the sense of “high-resolution” was considerable. In contrast, for the sense of reality, “three-dimensional” was a major contributor, in addition to “high-resolution.” It should be noted that the coefficient of determination, R2 , was slightly small for “being-there,” indicating the presence of another factor that contributes to the sense of “being-there.” These provide insights for emphasizing the higher-order impressions through the enhancement of the relatively operable lower-order impressions. This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. 10 IEEE TRANSACTIONS ON BROADCASTING In the case of a wide-field image with an FOV of more than 60◦ , the effective FOV that allows the viewer to obtain information instantly covers only part of the image. In addition, it has been reported that the gazing point is more focused at higher resolutions [38]. Therefore, the impression is not derived from the entire image, and it is necessary to reduce the gaze area to accurately investigate the relationship between the physical features of the image and the induced impression. Further investigation is needed to elucidate the relationship. In addition, its application in signal processing, such as automatic impression enhancement, is also included in our future study. VI. C ONCLUSION In this study, the psychological effects induced by 8KUHDTV images were investigated through subjective evaluation experiments and analyses of the familiar impressions in everyday television viewing experience. In addition, it was verified whether the 8K-UHDTV design concept aimed at realizing the ultimate 2D image was satisfied. The increase in resolution from 4K-UHD to 8K-UHD enhanced the lower-order impressions of familiar images, which in turn enhanced the higher-order impressions. This effect was found, even at four times the design viewing distance. 8K-UHDTV images induced strong viewer impressions at a wide viewing distance range, sufficient to cover the preferred viewing area in living room TV viewing. This indicates that 8K is the necessary and sufficient resolution for TV viewing in daily life. In addition, the analyses results provide insights on enhancing the impressions in UHDTV production, particularly, for the two major features of UHDTV, namely, the sense of reality and “being-there.” We used standard dynamic range (SDR) images as the test images in this study. However, HDR images have become popular and contribute to impression enhancement. We intend to perform further investigations with 8K-HDR images and clarify the relationship between the physical image features and the induced impressions. ACKNOWLEDGMENT The authors would like to thank M. Ando for his technical support. R EFERENCES [1] “Parameter values for ultrahigh definition television systems for production and international programme exchange,” Int. Telecommun. Union, Geneva, Switzerland, ITU-Recommendation BT.2020, 2012. [2] M. Sugawara, M. Emoto, K. Masaoka, Y. Nishida, and Y. Shishikui, “SUPER Hi-VISION for the next generation television,” SMPTE Motion Imag. J., vol. 1, no. 1, pp. 27–33, 2013, doi: 10.3169/mta.1.27. [3] T. Hatada, H. Sakata, and H. Kusaka, “Psychological analysis of the ‘sensation of reality’ induced by a visual wide-field display,” SMPTE Motion Imag. J., vol. 89, no. 8, pp. 560–569, Aug. 1980, doi: 10.5594/j01582. [4] M. Emoto, K. Masaoka, M. Sugawara, and F. Okano, “Viewing angle effects from wide field video projection images on the human equilibrium,” Displays, vol. 26, no. 1, pp. 9–14, Jan. 2005, doi: 10.1016/j.displa.2004.09.004. [5] K. Masaoka, M. Emoto, M. Sugawara, and Y. Nojiri, “Contrast effect in evaluating the sense of presence for wide displays,” J. Soc. Inf. Display, vol. 14, no. 9, pp. 785–791, 2006, doi: 10.1889/1.2358572. [6] M. Emoto and M. Sugawara, “Critical fusion frequency for bright and wide field-of-view image display,” J. Display Technol., vol. 8, no. 7, pp. 424–429, Jul. 2012, doi: 10.1109/jdt.2012.2191390. [7] K. Omura, M. Sugawara, and Y. Nojiri, “Evaluation of motion blur by comparison with still picture,” in Proc. IEICE Gen. Conf., 2008, pp. S-5–S-6. [8] Y. Shishikui, Y. Fujita, and K. Kubota, “Super Hi-vision—The star of the show!” EBU Tech. Rev., vol. 2009, pp. 1–13, Jan. 2009. [9] M. Sugawara et al., “Super hi-vision at the London 2012 olympics,” SMPTE Motion Imag. J., vol. 122, no. 1, pp. 29–38, Jan. 2013, doi: 10.5594/j18257XY. [10] M. Harasawa, Y. Sawahata, K. Komine, and S. Shioiri, “Effects of content and viewing distance on the preferred size of moving images,” J. Vis., vol. 20, no. 3, p. 6, Mar. 2020, doi: 10.1167/jov.20.3.6. [11] R. Sotelo, J. Joskowicz, R. Puddu, D. D. Giusto, and M. Anedda, “Extension to 4K resolution of a parametric model for perceptual video quality estimation,” presented at the IEEE Broadcast Symp. (BTS), Oct. 2018, doi: 10.1109/bts.2018.8551143. [12] T. Tian, H. Wang, L. Zuo, C.-C. J. Kuo, and S. Kwong, “Just noticeable difference level prediction for perceptual image compression,” IEEE Trans. Broadcasting, vol. 66, no. 3, pp. 690–700, Sep. 2020, doi: 10.1109/tbc.2020.2977542. [13] T. Li, X. Min, H. Zhao, G. Zhai, Y. Xu, and W. Zhang, “Subjective and objective quality assessment of compressed screen content videos,” IEEE Trans. Broadcasting, vol. 67, no. 2, pp. 438–449, Jun. 2021, doi: 10.1109/tbc.2020.3028335. [14] Y. Shishikui and Y. Sawahata, “Effects of viewing ultra-high-resolution images with practical viewing distances on familiar impressions,” IEEE Trans. Broadcasting, vol. 64, no. 2, pp. 498–507, Jun. 2018, doi: 10.1109/tbc.2018.2829118. [15] D. Park, Y. Kim, and Y. Park, “Cognitive effect on image quality in full ultra-high definition(8 K),” J. Inf. Display, vol. 21, no. 2, pp. 103–111, Dec. 2019, doi: 10.1080/15980316.2019.1704895. [16] N. Narita, M. Kanazawa, and F. Okano, “Assessment to determine format for a UDW television,” SMPTE Motion Imag. J., vol. 111, no. 12, pp. 594–599, Dec. 2002, doi: 10.5594/J14144. [17] “Subjective assessment methods for image quality in high-definition television,” Int. Telecommun. Union, Geneva, Switzerland, ITURecommendation BT.710, 1998. [18] “Subjective assessment of standard definition digital television (SDTV) systems,” Int. Telecommun. Union, Geneva, Switzerland, ITURecommendation BT.1129-2, 1998. [19] “Relative quality requirements of television broadcast systems,” Int. Telecommun. Union, Geneva, Switzerland, ITU-Recommendation BT.1127, 1994. [20] “General viewing conditions for subjective assessment of quality of SDTV and HDTV television pictures on flat panel displays,” Int. Telecommun. Union, Geneva, Switzerland, ITU-Recommendation BT.2022, 2012. [21] K. Masaoka, Y. Nishida, M. Sugawara, E. Nakasu, and Y. Nojiri, “Sensation of realness from high-resolution images of real objects,” IEEE Trans. Broadcasting, vol. 59, no. 1, pp. 72–83, Mar. 2013, doi: 10.1109/tbc.2012.2232491. [22] Y. Tsushima, K. Komine, Y. Sawahata, and N. Hiruma, “Higher resolution stimulus facilitates depth perception: MT+ plays a significant role in monocular depth perception,” Sci. Rep., vol. 4, no. 1, p. 6687, Oct. 2014, doi: 10.1038/srep06687. [23] Y. Tsushima, K. Komine, Y. Sawahata, and T. Morita, “Undetectable changes in image resolution of luminance-contrast gradients affect depth perception,” Front. Psychol., vol. 7, p. 242, Feb. 2016, doi: 10.3389/fpsyg.2016.00242. [24] A. Lachat, J.-C. Gicquel, and J. Fournier, “How perception of ultrahigh definition is modified by viewing distance and screen size,” in Proc. Image Qual. Syst. Perform. XII, Feb. 2015, Art. no. 93960Y, doi: 10.1117/12.2081140. [25] P. Hanhart, P. Korshunov, and T. Ebrahimi, “Subjective evaluation of higher dynamic range video,” in Proc. Appl. Digit. Image Process. XXXVII, Sep. 2014, Art. no. 92170L, doi: 10.1117/12.2076644. [26] C. Bist, R. Cozot, G. Madec, and X. Ducloux, “QoE-based brightness control for HDR displays,” presented at the 9th Int. Conf. Qual. Multimedia Exp. (QoMEX), May 2017, doi: 10.1109/qomex.2017.7965652. [27] K. C. Noland, M. Pindoria, and A. Cotton, “Modelling brightness perception for high dynamic range television,” presented at the 9th Int. Conf. Qual. Multimedia Exp. (QoMEX), May 2017, doi: 10.1109/qomex.2017.7965633. This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. SHISHIKUI AND SAWAHATA: QUALITY OF 8K UHDTV EXPERIENCE [28] Y. Shishikui, “Interaction between high resolution and high dynamic range in psychological effects of ultra-high-definition television,” in Proc. Int. Workshop Adv. Imag. Technol. (IWAIT), Mar. 2021, pp. 1–9, doi: 10.1117/12.2589060. [29] Y. Sawahata, Y. Miyashita, and K. Komine, “Estimating angular resolutions required in light-field broadcasting,” IEEE Trans. Broadcasting, vol. 67, no. 2, pp. 473–490, Jun. 2021, doi: 10.1109/tbc.2020.3047218. [30] G. Van Wallendael, P. Coppens, T. Paridaens, N. Van Kets, W. Van den Broeck, and P. Lambert, “Perceptual quality of 4K-resolution video content compared to HD,” presented at the 8th Int. Conf. Qual. Multimedia Exp. (QoMEX), Jun. 2016, doi: 10.1109/qomex.2016.7498935. [31] S. E. Palmer, K. B. Schloss, and J. Sammartino, “Visual aesthetics and human preference,” Annu. Rev. Psychol., vol. 64, no. 1, pp. 77–107, Jan. 2013, doi: 10.1146/annurev-psych-120710-100504. [32] “Methodology for the subjective assessment of the quality of television pictures,” Int. Telecommun. Union, Geneva, Switzerland, ITURecommendation BT.500-13, 2012. [33] R. Sotelo, J. Joskowicz, M. Anedda, M. Murroni, and D. D. Giusto, “Subjective video quality assessments for 4K UHDTV,” presented at the IEEE Int. Symp. Broadband Multimedia Syst. Broadcast. (BMSB), Jun. 2017, doi: 10.1109/bmsb.2017.7986225. [34] T. Nakamura, Y. Iwasaki, S. Tajima, K. Masaoka, T. Yamashita, and Y. Nishida, “MTF measurement of a super hi-vision camera,” in Proc. ITE Winter Annu. Convent., 2013, pp. 11–18, doi: 10.11485/itewac.2013.0_11_6. [35] Y. Benjamini and Y. Hochberg, “Controlling the false discovery rate: A practical and powerful approach to multiple testing,” J. Roy. Stat. Soc. B Methodol., vol. 57, no. 1, pp. 298–300, Jan. 1995. [36] N. Cowan and C. C. Morey, “Visual working memory depends on attentional filtering,” Trends Cogn. Sci., vol. 10, no. 4, pp. 139–141, Apr. 2006, doi: 10.1016/j.tics.2006.02.001. [37] M. Emoto and M. Sugawara, “Viewers’ optimization of preferred viewing distance by spatial resolution of TV display,” Displays, vol. 45, pp. 1–5, Dec. 2016, doi: 10.1016/j.displa.2016.07.003. [38] H. Nemoto, P. Hanhart, P. Korshunov, and T. Ebrahimi, “Impact of ultra high definition on visual attention,” presented at the MM ACM Multimedia Conf., Nov. 2014, doi: 10.1145/2647868.2654917. 11 Yoshiaki Shishikui (Senior Member, IEEE) received the B.S., M.S., and Ph.D. degrees in electrical engineering from the University of Tokyo, Tokyo, Japan, in 1981, 1983, and 1997, respectively. He joined NHK (Japan Broadcasting Corporation), Tokyo, in 1983. From 1986 to 2014, he worked with NHK Science and Technology Research Laboratories, where he was engaged in research on digital signal processing, picture coding, HDTV broadcasting systems, IPTV systems, advanced data broadcasting systems, and UHDTV research activities. He led the Super Hi-Vision public viewing project at the London 2012 Olympics. From 2001 to 2003, he was with NHK Engineering Services Inc., on loan, where he helped develop video archives and video-on-demand systems. In April 2014, he was appointed as a Professor with the Department of Frontier Media Science, School of Interdisciplinary Mathematical Sciences, Meiji University. Prof. Shishikui is a Fellow of IEICE Japan, ITE Japan, and SMPTE. He has been actively involved in standardization activities at SMPTE and ISO-IEC (MPEG). Yasuhito Sawahata (Member, IEEE) received the B.S., M.Sc., and Ph.D. degrees in information science and technology from the University of Tokyo in 2001, 2003, and 2015, respectively. Since 2003, he has been working with Science and Technology Research Laboratories, NHK (Japan Broadcasting Corporation), Tokyo, Japan. His work is focused on human information processing, neuroimaging, neural-decoding, and psychophysics.