Starkey Research & Clinical Blog

Does Expansion Decrease Low Level Speech Understanding?

Effects of Expansion on Consonant Recognition and Consonant Audibility

Brennan, M., & Souza, P. (2009). Effects of expansion on consonant recognition and consonant audibility. Journal of the American Academy of Audiology 20, 119-127.

This editorial discusses the clinical implications of an independent research study. The original work was not associated with Starkey Laboratories and does not reflect the opinions of the authors. 

The primary goal of a hearing aid fitting is to improve audibility and availability of speech sounds while maintaining comfort and loudness tolerance.  A linear hearing aid fitting may provide audibility for average speech sounds but may result in discomfort or for loud sounds and inaudibility for soft sounds. The use of wide-dynamic range compression (WDRC) has addressed these issues, helping maximize the useful dynamic range of hearing for individuals who require amplification for quiet and moderate sounds, yet have limited tolerance for loud sounds.

One potential issue with WDRC has been the increased audibility of very soft environmental sounds, which may be unwelcome for individuals who have adjusted to long-term hearing loss and are not used to perceiving these sounds. An additional problem for individuals with good residual hearing at some frequencies is that they will hear circuit noise from the hearing aid itself. Both of these issues can be unpleasant for the listener, possibly resulting in rejection or limited use of the hearing aids.

Expansion makes hearing aids quieter at low input levels. This is done in almost all modern hearing aids in order to reduce annoying environmental or circuit noise.  There is concern, however, that if too aggressive, expansion can have a detrimental effect on speech intelligibility (Plyler et al., 2005). It has been proposed that reduced speech recognition ability with expansion is due to reduced audibility of speech cues (Walker et al, 1984; Plyler et al, 2007).

Previous examinations of expansion have measured its effect on audibility of room noise (Plyler et al, 2005) or the long-term average speech spectrum or LTASS (Zakis and Wise, 2007) but did not directly measure the effect of expansion on audibility of the speech signals.  The current authors sought more specific insight into the effect of expansion on speech recognition by studying the relationship between expansion and consonant audibility.

Though there may be other related parameters warranting examination, the primary variables of interest relating to expansion are the ratio and the kneepoint.  In theory, a high expansion kneepoint should have a negative effect on speech recognition, because gain for stimuli below the kneepoint is reduced, resulting in decreased audibility.  Speech presented above the expansion kneepoint should be less affected by the expansion.

Therefore, the hypotheses for Brennan and Souza’s study were as follows:

1. A high expansion kneepoint will significantly reduce consonant-vowel (CV) recognition.

2. A high expansion kneepoint will significantly reduce CV audibility.

3. The effect of expansion on speech recognition and audibility will be reduced for increased speech input levels.

4. There will be a significant positive correlation between CV recognition and audibility for each condition.

Thirteen hearing-impaired individuals participated in the experiment. Nine were experienced hearing aid users; the remaining four did not use hearing aids. Subjects were fitted monaurally with a multi-channel, digital, behind-the-ear hearing aid.  Venting was 3mm for most subjects, but was reduced to 1mm for two subjects and plugged for one subject. 

The hearing aids were set to DSL 4.1 targets and had three separate programs:

1. Multichannel WDRC with an expansion kneepoint of 50dB SPL (high kneepoint condition)

2. Multichannel WDRC with an expansion kneepoint of 30dB SPL (low kneepoint condition)

3. Linear amplification with output compression limiting (control condition)

Expansion ratio was constant at 0.7:1, which represents a typcial expansion ratio currently available in hearing aids.

Eight CV nonsense syllables, four voice and four unvoiced, from the Nonsense Syllable Test (Dubno and Dirks, 1982) were presented to subjects at 50, 60 and 71dB SPL. Recordings of aided stimuli were measured at the tympanic membrane for each subject (Souza and Tremblay, 2006) and signal audibility was determined using the Aided Audibility Index -AAI (Stelmachowicz et al, 1994) using modifications for hearing-impaired subjects as described by Souza and Turner (1998).

Three of Brennan and Souza’s hypotheses were confirmed:  high expansion kneepoints significantly reduced signal audibility for speech at all presentation levels and consonant-vowel recognition scores were significantly lower for the high kneepoint condition, especially at presentation levels of 50dB and 60dB SPL. Subsequent regression analyses revealed that CV syllable recognition scores were significantly associated with audibility. The authors’ presumption that the effect of expansion on audibility and speech recognition would decrease with increasing speech presentation levels was not confirmed.  Instead, expansion had negative effects on CV recognition and audibility at all presentation levels.  This was in contrast with previous work reporting that expansion did not affect speech recognition above certain levels (Walker et al, 1984; Bray and Ghent, 2001; Plyler et al, 2005a, 2007), but this discrepancy may be explained by differences in presentation level, speech materials, expansion ratio or time constants or other hearing aid settings.

Despite some variability in results across studies, it is clear that high expansion kneepoints result in decreased speech recognition scores, presumably due in part to decreased audibility. Other potential explanations involve degradation of temporal cues and disruption of the intensity relationships between consonants and vowels, which provide important cues for consonant recognition (Walker et al, 1984; Hedrick and Younger, 2007).

Expansion is a feature of modern hearing aids that is often misunderstood; because it is characterized in terms of a ratio and kneepoint, it may be easily confused with compression.  Essentially the opposite of compression, expansion results in less amplification for softer sounds than louder sounds.  The expansion kneepoint is often the same as the compression kneepoint, indicating that expansion occurs below the kneepoint level and compression occurs above it. Alternatively, the input/output function of a circuit might show a region of linearity between the expansion and compression kneepoints.

Regardless of its various characteristics, expansion may help reduce the perception of unwanted environmental and hearing aid circuit noise, resulting in improved subjective hearing aid performance.  However, because audibility and speech recognition are two primary goals of amplification, it is essential to ensure that expansion does not result in decreased objective performance.  Brennan and Souza suggest that the use of active noise reduction for lower level stimuli may provide the benefits of expansion without the negative effect on speech recognition, but more research on this topic is warranted. Because expansion is commonly used in current hearing instruments, it is important for audiologists to understand the principles of compression and expansion so that appropriate settings can be selected to maximize audibility and comfort for individual hearing aid users.  And as always, counseling is essential for preparing both new and experienced hearing aid users for adjustment to new hearing aid technology and the perception of normal environmental sounds.

References

Brennan, M. & Souza, P. (2009). Effects of expansion on consonant recognition and consonant audibility. Journal of the American Academy of Audiology 20: 119-127.

Dubno, J.R. & Dirks, D.D. (1982). Evaluation of hearing-impaired listeners using a Nonsense-Syllable Test: I test reliability. Journal of Speech, Language and Hearing Research 25: 135-141.

Hedrick, M.S. & Younger, M.S. (2007). Perceptual weighting of stop consonant cues by normal and impaired listeners in reverberation versus noise. Journal of Speech, Language and Hearing Research 50: 254-269.

Plyler, P., Hill, A., & Trine, T. D. (2005a). The effects of expansion on the objective and subjective performance of hearing instrument users. Journal of the American Academy of Audiology, 16, 101-113.

Plyler, P.N., Lowery, K.J., Hamby, H.M. & Trine, T.D. (2007). The objective and subjective evaluation of multichannel expansion in wide dynamic range compression hearing instruments. Journal of Speech, Language and Hearing Research 50: 15-24.

Souza, P.E. & Tremblay, K.L. (2006). New perspectives on assessing amplification effects. Trends in Amplification 10: 119-143.

Souza, P.E. & Turner, C.W. (1998). Multichannel compression, temporal cues and audibility. Journal of Speech, Language and Hearing Research 41: 315-326.

Stelmachowicz, P.G., Lewis, D.E., Kalberer, L. & Creutz, T. (1994). Situational Hearing Aid Response Profile User’s Manual (SHARP, v 6.0). Omaha: Boys Town National Research Hospital.

Walker, G., Byrne, D., & Dillon, H. (1984). The effects of multichannel compression/expansion amplification on the intelligibility of nonsense syllables in noise. Journal of the Acoustical Society of America 76: 746-757.

Zakis, J.A. & Wise, C. (2007). The acoustic and perceptual effects of two noise suppression algorithms. Journal of the Acoustical Society of America 121: 433-441.