Concurrent information communication in voice-based interaction

Publication Type:
Thesis
Issue Date:
2019
Full metadata record
Speech-based information is primarily communicated to users sequentially; however, users are capable of obtaining information from multiple sources concurrently. This fact implies that the sequential approach is under-utilising human perception capabilities and restricting users to perform optimally. In this research, two informal studies and two experiments were carried out for investigating concurrent communication of multiple voice-based information streams. The informal studies were carried out to understand users’ interest and expectations in concurrent information communication and to examine whether users can comprehend concurrent information. In the first experiment, different designs for speech-based multiple information communication and the depth of comprehension by users in each design were tested. In the second experiment, various combinations of information streams presented concurrently and their viability regarding cognitive load were tested. The results of the first study manifested user’s interest in concurrent information communication design and supported the argument that users are able to discriminate and understand the concurrent voice streams using their selection and attention abilities. The results led to the second study, where users, including visually challenged users, expressed their expectations from such system and shared how would they prefer to interact with the systems providing concurrent information communication. Based on user’s feedback, a web-based ‘𝘝𝘪𝘯𝘧𝘰𝘮𝘪𝘻𝘦 𝘧𝘳𝘢𝘮𝘦𝘸𝘰𝘳𝘬’ is designed to allow for concurrent communication of multiple information streams to users. Findings from the third study showed that concurrent speech-based information designs, involving intermittent form and a spatial difference in sources of the streams, provide satisfying comprehension of the content. The study further showed that users could comprehend both the main information and the detailed information. The fourth study showed that the perceived cognitive workload for the listening task in baseline condition and concurrent combinations remain the same; however, users response in preference and frequently using different combinations remain significantly lower than the baseline condition. The fourth study also showed that the combinations created with music were preferred the most by the users in concurrent combinations, followed by the song. From the information types providing speech-based information (non-music/song), result shows the intermittent form of communication creates the low cognitive workload in voice-based information communication. Our research findings contribute to providing improvements in methods to communicate voice-based information efficiently under a large variety of application fields.
Please use this identifier to cite or link to this item: