Investigation into the relationship between phonetic segments and their acoustic values demands complex data analysis. This paper describes a software environment called aps designed to aid this task, aps provides a means of indexing segments of continuous speech in a general way. These segments may be obtained from a hand-produced phonemic transcription of the speech according to a phonemic context specified by the user. Alternatively, segments may be generated on the basis of acoustic values produced by a variety of signal processing operations. Functions are available that measure the acoustic values within segments. Such operations are commonly performed on thousands of speech segments simultaneously, aps has therefore been integrated with S, a comprehensive data analysis environment.
Bibliographic reference. Watson, Gordon S. (1989): "APS: an environment for acoustic phonetic research", In EUROSPEECH-1989, 2300-2303.