Recent developments of speech analysis tools in the form of computer software enable us to visualise and observe in a more efficient manner than ever the acoustic characteristics of speech. The pedagogical applications of the analytic technique have been made to a visual-aided teaching method of English pronunciation. This paper, comparing the phonetic characteristics of English pronunciation by native speakers of English and those by Japanese speakers, attempts to clarify the causes of the so-called 'unEnglish' features by Japanese speakers in the following order: segments, rhythm and weakening, assimilation, and intonation. In so doing, teaching methods and learning materials are suggested in order for Japanese learners of English to overcome the inappropriate elements in their pronunciation.