OPTIMAL TELEVISION SCANNING FORMAT FOR

CRT-DISPLAYS

  Erwin B. Bellers, Inge de Weerd, Gerard de Haan , and Ingrid E.J. Heynderickx Philips Research Laboratories, Briarcliff Manor, USA

Philips Research Laboratories, Eindhoven, The Netherlands

ABSTRACT at the cost of a lower refresh rate, or a change in the inter- Current consumer-level priced and state-of-the-art scan- lace phase. Note that this recently obtained additional free- rate converters enable a spatio-temporal decoupling of the dom enables a variety of pixel distributions in both space received video and the displayed video. This paper presents and time, for every chosen pixel rate. the results of a subjective assessment indicating the pre- THE EXPERIMENTS ferred CRT-display format. The optimal balance between the various parameters can only be found by conducting a subjective assessment, as INTRODUCTION defining a reliable objective metric is still a major challenge. Interlaced video with 525 or 625 lines and a 50 or 60 Hz In our experiments, we selected the following scanning picture rate has been the television broadcast standard for formats for the subjective assessment,(see also Figure 1): quite some time. Modern bright television screens, how- ever, require a modified display format to prevent annoying • 50 Hz, progressive (1:1), 625 scanning lines (50p) large area flicker and/or interline flicker. Moreover, matrix • 60 Hz, progressive (1:1), 525 scanning lines (60p) displays require format conversion as they cannot directly • 50 Hz, interlaced (2:1), 1250 scanning lines (50i) cope with interlace. Finally, new image sources, such as the • 75 Hz, interlaced (2:1), 833 scanning lines (75i) Internet, benefit from an increased spatial pixel density to • 100 Hz, interlaced (2:1), 625 scanning lines (100i) improve the legibility of displayed textual and graphical in- formation. As such, the optimal scanning or video format Note that the above formats differ in the number of scanning requirements may differ per application. lines, the interlace phase and the refresh rate, butall use the The techniques for high quality video format conversion same pixel frequency. have recently reached a price / performance ratio that en- A complication, when comparing different scanning for- ables application in the consumer domain [1, 2, 3]. As a re- mats on a single Cathode Ray Tube (CRT), is that the spot sult, we can choose the displayed number of scanning lines, dimensions cannot be simultaneously optimal for all for- the interlace factor and the picture rate at will. Given this mats. Clearly a fine spot is required to exploit the highest new freedom, the question arises how to optimally choose vertical resolution resulting from the scanning format with a display format for the current applications. Increasing the the highest number of lines. However, for the scanning for- number of scanning lines increases the vertical resolution. mats with a lower line count, such a fine spot may lead to Modifying the interlaced scanning to the progressive format an annoying visibility of the line structure. To prevent the reduces or eliminates any of the possible interlace artifacts choice of the spot dimensions leading to a bias in the opti- like line flickering and line crawl that may appear. Finally, mization of the scanning format, we optimized the spot size an increase in the refresh rate reduces or eliminates the large per scanning format as a balance between perceived sharp- area flicker. Consequently, one might expect optimal per- ness and the visibility of line structure in a first session of formance by using both the highest number of scanning the subjective test. lines and at the highest refresh rate possible. Obviously, To realize a variable spot size without changing the dis- there is a cost increase associated with such an increase in play, we emulated a relatively low resolution display on a overall quality. high resolution monitor. A single line was mapped to a In order to have a fair optimization of the television dis- number of scanning lines of the monitor, using a Gaussian play format, i.e. comparing options with approximately the filter to represent the spot dimensions of the emulated CRT. same cost1, we may increase the number of scanning lines The viewing distance in the subjective assessment was se- lected in accordance with the emulated low resolution dis- 1the same pixel frequency play. display format From the subjective evaluation, we draw the conclusion that the video format of 75i is superior to the alternative scanning formats2. A remarkable observation is that the preference is most impressive in comparing the 75i format 625(1:1)@50Hz with the two progressive display formats (50p and 60p). In fact, all the interlaced formats are preferred over the pro- gressive formats. The loss of vertical resolution of the pro- gressive formats is apparently recognized as the most dis- 525(1:1)@60Hz input format tinguishing element. compare Another interesting observation is that the preference of Video Format the 75i format over the 100i format indicates that the ob- Conversion servers clearly notice the difference in vertical resolution, e.g. 625(2:1)@50Hz 1250(2:1)@50Hz and probably hardly observe any difference in large area flicker or line flicker. The difference of the 100i with the 50i format was not considered to be significant, i.e. it was found difficult to chose between a picture with significant

833(2:1)@75Hz line / large area flicker and a high vertical resolution, and a picture with a significant lower resolution and no visible interlace artifacts. CONCLUSIONS

625(2:1)@100Hz Recent progress in scan-rate conversion technology en- ables a decoupling of the received video format and the dis- Figure 1: Decoupling of the input video format and the dis- play format. As a result, we can choose the displayed num- play format. ber of scanning lines, the interlace factor and picture rate at will. Our subjective evaluation revealed that, for CRT 50p 60p 50i 75i 100i television displays and for the tested formats, our viewers 50p 40.8 83.5 97.4 84.2 always preferred interlaced scanning over progressive scan- 60p 59.2 84.2 96.7 89.5 ning with the same pixel rate. From the range of picture 50i 16.5 15.8 71.1 44.7 rates that we tested, the 75i format turned out to provide 75i 2.6 3.3 28.9 21.0 the best balance between flicker and resolution. Finally, our 100i 15.8 10.5 55.3 79.0 (European) viewers preferred the 100i high-end television format over the 60p high-end format. Of all tested formats, Table 1: Overall results of the subjective assessment. we conclude that the interlaced format at 75 Hz, 833 lines After this initial optimisation of the spot size, the differ- is optimal for CRT displays with a 27 MHz luminance pixel ent scanning formats were ranked in order of general pref- rate. erence by expert and non-expert viewers (19 in total) in a multiple comparison experiment. Critical image material, REFERENCES originating from either television cameras, or obtained from [1] G. de Haan, J. Kettenis and B.D. Loore, ‘IC for motion- Internet pages has been selected for this subjective assess- compensated 100 Hz TV with natural-motion movie- ment (4 different pictures in total). mode ’, IEEE Tr. on Consumer Electronics, May 1996, pp. 165-174. RESULTS The results of the experiments are presented in Table 1. [2] G. de Haan, ‘IC for Motion-Compensated De- This table shows the percentage that the scanning format interlacing, Noise Reduction, and Picture-Rate Conver- indicated at the column head is preferred over the format sion’, IEEE Tr. on Consumer Electronics, Vol. 45, Au- indicated at the row head, e.g. there is a 97.4% preference gust 1999, pp. 617-624. of the 75i format over the 50p display format. A shaded [3] M. Schu, G. Scheffler, C. Tuschen, and A. Stolze, ‘Sys- table cell indicates that the corresponding preference is sig- tem on silicon-IC for motion compensated scan rate nificantly different from 50%, which would be the result conversion, picture-in-picture processing, split screen when subjects were randomly selecting just one of the for- applications and display processing’, IEEE Tr. on Con- mats without having a real preference. The ranking of the sumer Electronics, Vol. 45, August 1999, pp. 842-850. formats resulting from these percentages is: 1) 75i, 2) 50i 2 and 100i, 3) 50p and 60p, where the combined formats into We also found (not shown in the table) that this preference was most noticed for the pictures that contain highly textured regions (mainly found one rank indicates that the difference in preference between in graphical or Internet related pictures), which was expected as interlace these formats is not statistically significant. artifacts start to become most pronounced in highly detailed pictures.