An experimental study describing the effects of carbon and electret telephone transducers on automatic speech recognition (ASR) performance is presented. It is shown that telephone based ASR performance on a connected digit task actually improves when speech is spoken through the carbon transducer. This surprising result is explained by a study of the differences in acoustic characteristics between carbon and electret telephone handsets. An initial attempt is made to devise a simple procedure for obtaining a parametric transformation which emulates the properties of the carbon transducer. The parameters of this transformation are trained automatically from speech spoken simultaneously through carbon and electret telephone handsets. When telephone speech data is transformed according to this procedure, a significant improvement in ASR performance is obtained. These results are interpreted and future research directions are discussed.
Bibliographic reference. Potamianos, Altxandros / Lee, Li / Rose, Richard C. (1995): "A feature-space transformation for telephone based speech recognition", In EUROSPEECH-1995, 1533-1536.