The Generation of Synthetic Speech Sounds by Digital Coding

The feasibility of representing human speech by serial digital codes was investigated by exercising specially constructed digital logic coupled with standard audio output equipment. The theories being tested represent a radical departure from previous efforts in the field of speech research. Therefore, this initial investigation was limited in scope to a study of unconnected English language speech sounds at the phenome level. The experiments were conducted in two parts, with the first being the development of serialized digital codes, for selected speech sounds, derived from actual human speech. The second part was to synthesize these sounds using the specially constructed digital synthesizer, and have human listeners analyze them for intelligibility. The results seem to indicate that this is a viable scheme for speech synthesis