Beverly H. Williges

The intelligibility of synthesized speech in data inquiry systems

Recent improvements in speech technology have made synthetic speech a viable form of system feedback. However, little is known about the effects of various speech parameters on speech intelligibility and system performance. Experiment 1 examined the effect of situational context clues and speech rate on synthesized speech intelligibility. Subjects who received pragmatic context information prior to each message had transcription error rates 50% lower than those who received no context information. Speech rates of 250 words per minute (wpm) yielded significantly more transcription errors and longer response latencies than rates of 180 wpm. In Experiment 2 the effects of speech rate, message repetition, and location of information in a message were examined. Transcription accuracy was best for messages spoken at 150 or 180 wpm and for messages repeated either twice or three times. Words at the end of messages were transcribed more accurately than words at the beginning of messages. Subjective ratings indica...Ninety male and female college students reclining on their backs in the dark were disoriented when positioned on a rotating platform under a slowly rotating disk that filled their entire visual field. Half of the disk was painted with a brighter value (~69% higher luminance level) of the color on the other half The effects of red, blue, and yellow were examined. Subjects wearing frosted goggles viewed the illuminated disk for three rotations. The disk was stopped when the subjects felt that they were right side up. A significant proportion of subjects selected the disk position in which the brighter side of each of the three colors filled their upper visual field. These results suggest that color brightness as well as lighting variation could provide space station crew members with body orientation cues as they move around.

Effects of recognition accuracy and vocabulary size of a speech recognition system on task performance and user acceptance

The purpose of this study was to determine the effects of recognizer accuracy and vocabulary size on performance of a speech recognition system. Subjects ranging in age from 20 to 55 years performed a data-entry task using a simulated speech recognizer that simulated three accuracy levels and three levels of available vocabulary. The accuracy level at which the recognizer was performing significantly influenced the time to complete the task as well as the users acceptability ratings but had only a small effect on the number of errors left uncorrected. Available word-level vocabulary size also significantly affected the task completion time; however, its effect on the average number of uncorrected errors and on the acceptability ratings was negligible. Age of the subject was found to influence both objective and subjective measures. Older subjects required more time to complete the tasks, though they consistently rated the speech input systems more favorably than did the younger subjects. Minimal experience with computer text editors may account in part for the speed differences between older and younger subjects.

User feedback requirements with automatic speech recognition

The optimal type and amount of secondary feedback for data entry with automatic speech recognition were investigated. Six feedback conditions, varying the information channel for feedback (visual or auditory), the delay prior to feedback, and the amount of feedback history, were compared to a no-feedback control. In addition, the presence of a dialogue requiring users to confirm a word choice when the speech recognizer could not distinguish between two words was studied. The word confirmation dialogue increased recognition accuracy by about 5% with no significant increase in the time to enter data. Type of feedback affected both accuracy and time to enter data. When no feedback was available, data entry time was minimal but there were many errors. Any type of feedback/error correction vastly unproved accuracy, but auditory feedback provided after a string of data was spoken increased the time to enter data by a factor of three. Depending on task conditions, visual or auditory feedback following each word ...

Modeling the Human Operator in Computer-Based Data Entry:

Meaningful human factors applications to the design of human-computer tasks require a quantitative data base that describes operator behavior as a function of various independent variables. Three classes of metrics (operator satisfaction ratings, work-sampling procedures, and embedded performance measurement) are described as important measures in evaluating a human-computer interface used to enter and update personnel records. Polynomial regression procedures were used to generate functional relationships between each of these metrics and four independent variables representing system delay, display rate, keyboard echo rate, and keyboard buffer length. Each of the 22 separate dependent variables showed different functional relationships among the four systems variables, but overall system delay and keyboard echo rate were the major determiners of operator behavior. Additionally, the three classes of metrics were combined into three underlying interface dimensions relating to operator production, waiting, and planning activities.

Synthesized Warning Messages: Effects of an Alerting Cue in Single- and Multiple-Function Voice Synthesis Systems

The present study examined the effectiveness of preceding synthesized voice warning messages with an alerting cue as a function of the amount of information presented by the voice synthesizer and the workload level in the primary task. Subjects performed a simplified air traffic control task in which they were required to monitor two visual displays and to enter commands via a standard keyboard. Emergency messages were always presented by phoneme-based synthesized speech. However, the presence of an alerting cue (light and tone) prior to emergency messages and the presentation mode of noncritical messages (visual or auditory) were varied experimentally. When synthesized speech was used only for emergency messages, the presence of an alerting cue lengthened the response time to the message. However, when computer-generated speech was used for multiple functions, more emergency messages were detected when an alerting cue was used.

Varying the type and number of adaptive variables in continuous tracking.

To investigate the effectiveness of various types and numbers of adaptive variables, 48 subjects performed a two-dimensional pursuit tracking task for five 3-min training sessions. In the factorial design resulting in eight experimental conditions, three variables (frequency of the forcing function, ratio of acceleration to rate control, and control stick sensitivity) were either fixed or adaptive. A transfer and retention task in which the tracking situation changed periodically was used to evaluate the ability of subjects to adjust to change. Each adaptive variable was analyzed separately in training. The highest rate of adaptation in frequency occurred when frequency was the only adaptive variable. The rate of adaptation in acceleration was greater early in training when frequency also adapted. More adaptation occurred in gain when another variable also adapted. During transfer subjects trained adaptively generally showed more stable performance in the changing task situation. No reliable differences among conditions appeared in retention. Results are discussed in terms of stimulus and response similarity, the optimum number of adaptive variables, and the appropriateness of a changing task to evaluate adaptive training.

Integrated Research Paradigm for Complex Experimentation

Most human factors design problems involve large data spaces with so many factors that a single experiment investigating all of these factors simultaneously is unreasonable. The design of a computer-based telephone inquiry system is presented as an example interface which required consideration of over 100 independent variables which could affect end-user performance. An integrated research paradigm involving the three major steps of selecting, describing, and optimizing these independent variables is presented. Specific methods and design implications are described for each of these steps. This overall strategy for sequential experimentation is discussed in terms of its usefulness in human factors design applications.

Software Interfaces for Aviation Systems

Publisher Summary As aviation systems become more computer dependent, the emphasis on user-oriented software design must increase to improve user acceptance, system efficiency, and safety. Human factors considerations in the design of computer-based aviation systems must address both the issues and the process of software interface design. These interfaces must incorporate fundamental principles and guidelines of the human–computer dialogue design that are tailored to the specific functions performed by pilots, controllers, and others who interact with computer-based aviation systems. Software functions that are relevant for the design of aviation systems include data entry, data display, data management, action control, feedback and user guidance, error management, data protection, language, and communication between users. In parallel with the development of human factors design guidelines and tools to use them, the software design process itself needs to be considered. The concept of layering software into computation and HCI components is central to the design of software for aviation systems. The design process for computational software for aviation systems must provide for software maintenance and must address the complexities of real-time operation, concurrency, and fault tolerance. HCI software must take into consideration the capabilities and limitations of the pilots and controllers who will use it, and iterative redesign is key to successful HCI software for computer-based aviation systems.

Learner-Centered Versus Automatic Adaptive Motor Skill Training

Three types of training (fixed-difficulty, automatic-adaptive, and learner-centered) were used to teach 18 male and 18 female students a two-dimensional pursuit-tracking task. A 7-min tracking session, in which task difficulty shifted each minute, was used to measure transfer. Although training type did not result in differences in training time, students trained under learner-centered procedures had less tracking error during transfer. Females required on the average twice as much training as mates. During transfer no sex differences were noted. The differences in training time for males and females may reflect previous experience with similar motor-control tasks.

Developing quantitative guidelines using integrated data from sequential experiments

The results of four experiments were integrated using empirical models and data bridging. These models can be used in the design of telephone-based information systems as quantitative guidelines. Ten independent variables including environmental, hardware, dialogue, and user factors were considered in four related studies. Three dependent variables (total search time, user-added keypresses, and message transcription accuracy) were evaluated in each separate study. Multiple linear regression was used to generate an integrated second-order empirical model for each of the three dependent variables. Overall, the use of integrated empirical models offers several advantages. First, both qualitative and quantitative design guidelines can be generated. Second, a broad base for generalization across several studies is provided. Third, interface design trade-offs are possible.