Phonetic classification on wide-band and telephone quality speech

Benchmarking the performance for telephone-network-based speech recognition systems i hampered bytwo factors: lack of standardized databases for telephone twork speech, and insuffi-cient understanding of the impact of the telephone twork on rec-ognition systems. The N-TIMIT database was used in the experiments described in this paper in order to "calibrate " the effect of the telephone twork on phonetic classification algo-fithrns. Phonetic classification algorithms have been developed for wide-band and telephone quality speech, and were tested on sub-sets of the TIMIT and N-TIMIT databases. The classifier described inthis paper provides accuracy of 75 % on wide-band TIM1T data nd 66.5 % on telephone quality N-TIMIT data. Over-all the telephone twork seems to increase the error ate by a fac-tor of 1.3. 2