October 2009 | NTIA Technical Report TR-10-462
Abstract: When bit errors are introduced between a speech encoder and a speech decoder, the quality of the received speech is reduced. The specific relationship between speech quality and bit error rate (BER) can be different for each speech coding and channel coding scheme. This report describes a subjective experiment concerning the relationships between BER and perceived speech quality for the TIA Project 25 Full Rate (FR), Enhanced Full Rate (EFR), and Enhanced Half Rate (EHR) speech codecs. Using the FR codec with 2 % random bit errors as a reference, we sought to characterize the BER values for which the EFR (or EHR) codec produces speech quality that is equivalent to the reference. We used an adaptive paired-comparison subjective testing algorithm to efficiently adapt BER values for the EFR and EHR codecs to quickly locate the BER values where listeners found the speech quality to be the same as the reference. The results from sixteen listeners reveal ranges of BER values that were judged to produce speech quality equivalent to the reference. When these ranges are reduced to central values, those values indicate that on average, the EFR and EHR codecs are more robust to bit errors than the FR codec. We provide a set of additional results from a popular objective speech quality estimator for comparison purposes.
Keywords: speech coding; speech quality; subjective testing; bit errors; listening tests
For technical information concerning this report, contact:
Andrew A. Catellier
Institute for Telecommunication Sciences
Disclaimer: Certain commercial equipment, components, and software may be identified in this report to specify adequately the technical aspects of the reported results. In no case does such identification imply recommendation or endorsement by the National Telecommunications and Information Administration, nor does it imply that the equipment or software identified is necessarily the best available for the particular application or uses.