Speech enhancement for bandlimited speech

Throughout the history of telecommunication, speech has rarely been transmitted with its full analog bandwidth (0 to 8 kHz or more) due to limitations in channel bandwidth. This impaired legacy continues with tactical voice communication. The passband of a voice terminal is typically 0 to 4 kHz. Hence, high-frequency speech components (4 to 8 kHz) are removed prior to transmission. As a result, speech intelligibility suffers, particularly for low-data-rate vocoders. In this paper, we describe our speech-processing technique, which permits some of the upperband speech components to be translated into the passband of the vocoder. According to our test results, speech intelligibility is improved by as much as three to four points even for the recently developed and excellent Department of Defense-standard Mixed Excitation Linear Predictor (MELP) 2.4 kb/s vocoder. Note that speech intelligibility is improved without expanding the transmission bandwidth or compromising interoperability with others. INTRODUCTION In analog voice transmission, speech bandwidth is limited by the channel bandwidth. In switched public telephone circuits or high-frequency radio channels, speech bandwidth is typically less than 4 kHz. With digital speech transmission, speech bandwidth still remains less than 4 kHz because the wider the speech bandwidth, the higher the data rate required for transmission. The 4 kHz speech bandwidth is standard in virtually all digital voice terminals used by the government and industry. A speech bandwidth of 4 kHz is acceptable for the vowels spoken by a majority of speakers. The same cannot be said for consonants, particularly fricatives (/s/, /sh/, /ch/, etc.), because their spectra extends above 4 kHz (see Fig. 1). Good reproduction of consonants is vital to a proper understanding of speech because they provide important cues for sentence segmentation. To improve narrowband speech, we developed a technique to shift the fricative spectrum down below 4 kHz so that some of the fricative sounds will be heard over digital voice terminals. Our earlier effort to spread fricative spectra was by making use of the aliasing effect (i.e., spectral folding around 4 kHz) [1]. This technique is not effective if there is no fricative spectrum in the 4 to 5 kHz region. Thus, we developed a new and more effective technique, which is capable of spreading fricative spectrum even if it is present only above 5 kHz.