A lightpen and a mouse as input devices were evaluated with respect to the amount of time required to point to a target on a CRT display. The distance to and the size of the target were treated as variables. While five subjects performed simple positioning tasks on a CRT display using a lightpen or a mouse, positioning times and errors were measured. The positioning time for a lightpen was found to be significantly shorter. The longer positioning time for a mouse was considered to be caused by the difficulty of eye-hand coordination and the difference of basic manual operation between the two devices. We recommend a lightpen for simple positioning tasks on a CRT, if the target size is large enough to compensate for the effect of parallax.

View full abstract

This paper considers the control performance of human operators in the task of optimal control. A subject is instructed to change the output of a plant from an initial point to a specified one, and to decrease simultaneously the value of a performance index shown to him after each trial. The performance index evaluates his control pattern.
A subject can decrease the index value after many trials even if he dose not have any knowledge about the dynamics of a plant and the optimal control pattern. If a subject knows the optimal control pattern, he can decrease the index value easily; his control pattern becomes close to the optimal one.
Such ability of a human operator to find a sub-optimal control pattern will be applicable to many fields, especially to automatic control systems.

View full abstract

A new evaluation method, which can detect the small differences between the quality of the synthesized and natural voices, was investigated using a short-term memory paradigm. Seventeen subjects listened to a series of synthesized or natural nonsense monosyllables whose quality were estimated to have almost same on the ordinary articulation score test, and recalled them immediately after the presentation of a series. The length of a series was composed of 6 or 8 monosyllables.
The recall errors were compared between two types of voices. Some synthesized syllables (eg. ‘mo’, ‘za’, ‘be’) caused much more errors than the corresponding natural syllables. Furthermore, such tendency became very clear when these syllables were presented at the latter portion of a serial position.
These results would suggest that the short-term memory paradigm is very effective to evaluate the small defects in the quality of voices.

View full abstract

According to the development of computers and the software technology, VDT works such as programming, word-processing and so on have been rapidly increasing. It is supposed, in these VDT works, that the degree to which it is easy to recognize the presented information on a CRT display has a large effect on efficiency, accuracy and fatigue in performing VDT works. This research clarifies some characteristics on a recognition process for the dotted character patterns on CRT displays, with the purpose of submitting some basic standards that are useful to design and evaluate the dotted patterns. In particular, following effects are clarified through three experiments taking up the upper-case letters as visual stimuli, that is, the effect on the accuracy in performing works (the rate of correct responses as a characteristic) by the number of dotted points that consists of a letter and by the different dotted patterns, and the effect on efficiency (processing time) by the different dotted patterns. In addition, we propose a human recognition model for dotted character patterns to discuss the above effects, and compare this model with other customary models. Finally, we suggest some guidelines for the dotted character patterns from these results.

View full abstract