Consonant perception with linear and compression amplification

The consonant perception of 15 subjects with mild to moderate sensorineural hearing loss was evaluated using linear amplification and two different types of compression amplification. A specially modified hearing aid was used which allowed for variation of the amplifier input/output function in three steps, such that the compression ratio could be set to 1 (linear), 1.3 or 1.8. The Nonsense Syllable Test (NST) was recorded through the aid in quiet and in two different noise conditions (four-talker babble and a background noise with sharp intermittent sounds), and replayed to the listeners through headphones. No differences in consonant perception were found between the different types of amplification in the quiet condition. In the babble condition, consonant perception was significantly better with linear amplification than with either form of compression. In the sharp noise condition, there was no difference in performance between linear amplification and compression amplification with the ratio of 1.8. Consonant perception was adversely affected, however, by the compression and amplification with the ratio of 1.3 in this condition. Overall NST results and results for particular classes of consonants are discussed.