Many new processing features in hearing aids have their primary effects on information located in the high frequencies. Speech perception tests that are optimized for evaluating high-frequency processing are needed to adequately study its effects on speech identification. The goal of the current research was to develop a medium for evaluating the effects of high-frequency processing in hearing aids.
A list of 115 consonant-vowel-consonant-vowel-consonant nonsense syllables with American English consonants in all word positions was created in an open-set phoneme identification format. The source material was spoken by a male and a female speaker. A custom computer program was developed for administration of the test and automatic analysis of the test results. Nine normal-hearing listeners were employed in the collection of the normative data. The test was presented to the listeners in quiet (at 68 dB SPL), in noise at five signal-to-noise ratios (SNRs; −10, −5, 0, 5, and 10), and in a low-pass filter condition with cutoff frequencies at 500, 1000, 1500, 2000, and 4000 Hz. The data were examined to evaluate the psychometric properties of the test for different phoneme positions and phoneme classes. In addition, a shortened version of the test was developed based on the data from normal-hearing listeners. The test–retest reliability was verified at 0 dB SNR. The full and shortened versions of the test were repeated in 10 hearing-impaired listeners at their most comfortable listening level in quiet and in noise at various SNRs.
The availability of high-frequency output was verified with acoustic analysis. The performance intensity functions for both versions of the test (i.e., male and female speakers) showed expected monotonic growth with SNR and cutoff frequencies. High reliability was seen between test and retest identification scores in normal-hearing and hearing-impaired listeners.
The current nonsense syllable test provided a reliable and efficient means for phoneme identification testing.
The goal of the research was to develop a nonsense syllable identification test that provides a tool for evaluating the effects of high-frequency processing in hearing aids. A set of desirable features for the test was first defined and incorporated into the design. A list of 115 consonant-vowel-consonant-vowel-consonant items with all American English consonants in all word positions was created in an open-set phoneme identification format. The source material was spoken by a male and a female speaker. A custom computer program was developed for administration and automatic data analysis of the test. Various performance intensity functions were provided on normal-hearing subjects. A shortened version of the test with 32 items was also generated. Verification of the full and short versions of the test with subjects of high-frequency hearing loss was included. SUPPLEMENTAL DIGITAL CONTENT IS AVAILABLE IN THE TEXT.
Widex Office of Research in Clinical Amplification (ORCA), Widex Hearing Aid Company, Lisle, Illinois.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.ear-hearing.com).
Received November 3, 2009; accepted May 8, 2010.