Auditory Testing of a Formant-Coding Speech Compression System
James L. Flanagan, Arthur S. House · The Journal of the Acoustical Society of America · 1956
An evaluation of the intelligibility of speech transmitted by a formant-coding compression system is described. The compression system codes the input speech in terms of six electrical signals, representing the frequencies of the first three formants, the intensity of voicing and of friction, and the fundamental vocal frequency. These signals occupy a total band width of approximately 50 cps and require signal-to-noise ratios of approximately 30 db for their transmission. Natural monosyllabic utterances are transmitted through the compression system and are presented to listeners for identification. Vowel and consonant articulation scores are computed and sound-confusion matrices are constructed. The contribution to intelligibility of various components of the system and the inherent limitations of the system are examined and discussed. [This work was supported in part by the Air Force Cambridge Research Center under Contract No. AF 19(604)-626.]