Telephone speech enhancement for elderly hearing-impaired listeners

Abstract

Many elderly persons with high-frequency hearing loss find telephone use frustrating due to lower intensity levels and reductions in acoustical information that can be useful in deciphering speech. The purpose of this project is to pre-process the speech signal before it is sent over the phone line and provide speech enhancement without the use of amplifying handsets or hearing aids at the receiving end. The enhancement technique takes into account the limited bandwidth of the phone line as well as the hearing characteristics of the user. Two pre-processing schemes, a single channel and a double channel approach, used to increase the intelligibility of speech in these situations are discussed. The single channel method performs amplitude compression of the entire signal. The two-channel method filters the incoming signal into high-frequency and low-frequency channels and performs independent compression on each before recombination. Results comparing the two speech enhancement schemes against no processing for a group of elderly hearing-impaired subjects are presented. [Work supported by a grant from the Franklin County Office on Aging.]

Age as a factor in ethnic accent identification in Singapore

“…As such, it was perhaps not their inability to perceive ethnicity of the callers, but circumstances made it impossible for these informants to do so. More importantly, telephone recordings lack accuracy because higher frequencies of speech, which are known for distinguishing between speakers, are lost through the telephone signals (Peterson 1952;Sheffield et al 1999), and this is especially so given the telephone technologies of the 1980s. The entire study was therefore in fact not a perception test with designed stimuli, nor was it one with some control of variables such as the ethnic group or age group of the informants and the callers.…”

Evaluation of a Telephone Speech-Enhancement Algorithm Among Older Adults With Hearing Loss

“…The Departments of Speech and Hearing Science and Electrical and Computer Engineering at The Ohio State University have collaborated to develop a computerbased telephone speech-enhancement algorithm (TSEA; Komattil, 2004;Natarajan, 2002;Poling, 2004;Sheffield, 2000). The TSEA is based on a hearing aid compressionprocessing algorithm aimed at preserving spectral contrast within the speech signal (Tejero-Calado, Rutledge, & Nelson, 2001) that was chosen, in part, to amplify the signal without introducing peak clipping.…”