2016 INTERSPEECH INTERSPEECH 2016

Analyzing the Relation Between Overall Quality and the Quality of Individual Phases in a Telephone Conversation

Abstract

Assessing and analyzing the quality of transmitted speech in a conversational situation is an important topic in current research. For this, a conversation has been separated into three individual conversational phases (listening, speaking, and interaction), and for each phase corresponding quality-relevant perceptual dimensions have been identified. The dimensions can be used to determine the quality of each phase, and the qualities of all phases, in turn, can be combined for overall conversational quality estimation. In this article we present the work that has been conducted to identify the weights of the individual phases for the overall quality. For this, we conducted an experiment that allows the participants to perceive each phase separately and to gather the overall quality as well as the quality ratings for each individual phase. The results enable to create a linear model to predict the overall quality on the basis of the three phases. This allows to draw first conclusions regarding the relation between the individual phases and the overall quality and provides a major landmark towards a diagnostic assessment of conversational quality.

🚀 Conference Pioneer — INTERSPEECH 2016
🌉 Interdisciplinary Bridge — Machine Learning and Mathematics & Optimization
🧭 Keyword Pioneer — conversational quality
🐣 Hot Topic Early Bird — quality assessment
🐝 Cross-Pollinator — Artificial Intelligence, Computer Science, Computer Vision, Data Science & Analytics, Deep Learning, Healthcare & Medicine, Interdisciplinary, Knowledge & Reasoning, Machine Learning, Mathematics & Optimization, Natural Language Processing, Reinforcement Learning, Speech & Audio