Personal reflections on 20 years of research with Manfred R. Schroeder: From concert hall acoustics to code-excited linear prediction
Bishnu Saroop Atal · The Journal of the Acoustical Society of America · 2006
I joined Bell Laboratories, Murray Hill in 1961. Bell Labs then had an excellent environment for challenging researchers to find creative solutions to difficult problems in telecommunication. Coming from India, I was fortunate to have Manfred as my mentor. He and I collaborated on many interesting problems. Stereo sound systems in those days kept the sound confined between two speakers, and we wondered why we could not spread the sound beyond the speakers. We invented a system for creating signals at the human ears that removed this limitation, producing the beginning of surround sound. Vocoders could not synthesize natural-sounding speech, and Manfred always felt that the speech synthesis models used in vocoders were too rigid and inflexible. This kind of thinking led us to the path of adaptive predictive coding and we presented a paper on this topic at the first IEEE meeting on speech communication in Boston in 1967. We continued working together producing better and better speech quality on predictive coders. Our research led to the approach of code-excited linear prediction, which forms the basis of world-wide digital cellular standards.