Skip to main content
. Author manuscript; available in PMC: 2013 Aug 13.
Published in final edited form as: Am J Audiol. 2012 Aug 2;21(2):269–294. doi: 10.1044/1059-0889(2012/12-0013)

Table 6.

Clincial Question 2: Comparison of amplitude compression WDRC versus linear amplification on speech recognition outcomes at average input levels (5570 dB).

Citation Condition Compression settings Stimuli Open/closed set Presentation level/SNR Effect size [95% CI]a p
Boothroyd et al. (1988)b WDRC Threshold: NR
Attack time: NR
Release time: NR
Ratio: prelimiting gains increased by the amount by which the participant’s dynamic range of hearing fell short of 30 dB
Volume control: NR
Nonsense syllables: All contrasts Closed NR rpb = −.14 [−.77, .63]c p < .01 (favors linear)
Final consonant voice Closed NR rpb = −.26 [−.82, .54]c p < .01 (favors linear)
Vowel height Closed NR rpb = −.35 [−.85, .47]c p < .01 (favors linear)
Duration Closed NR rpb = −.04 [−.72, .68]c ns
Intonation Closed NR rpb = −.03 [−.72, .69]c ns
Vowel place Closed NR rpb = −.07 [−.74, .67]c ns
Initial consonant continuance Closed NR
NR
rpb = −.28 [−.82, .53]c p < .01 (favors linear)
Linear CL Threshold: NR
Attack time: 2 ms
Release time: preamplifier = 1 s; second compression-limiter = 20 ms
Ratio: NR
Volume control: NR
Initial consonant voice Closed rpb = .11 [−.64, .76]c ns
Initial consonant place Closed NR rpb = .11 [−.64, .76]c ns
Christensen (1999) WDRC Threshold: Low: 45–50 dB SPL
Attack time: 5 ms
Release time: 50 ms
Ratio: Variable:1.1:1–7.8:1
Volume control: listener unable to adjust during testing
Bisyllabic spondaic words Open 65 dB SPL/quiet, +5, +10 SNR NR/NC ns across SNRs
Passages Open 65 dB SPL/quiet, +5, +10 SNR NR/NC ns across SNRs
Linear CL Threshold: High: 75 dB
Attack time: 5 ms
Release time: adjustable up to 1 s
Ratio: 8:1
Volume control: listener unable to adjust during testing
Linear PC Threshold: none
Attack/release time: instantaneous
Ratio: none
Volume control: listener unable to adjust during testing
Gou et al. (2002) WDRC Threshold: 40 dB
Attack time: NR
Release time: NR
Ratio: NR
Volume control: AGC
Words (discrimination of phonological oppositions) Closed 65 dB SPL/quiet NR/NC p < .01 (favors WDRC)
Monosyllabic words Closed 65 dB SPL/quiet NR/NC ns
Bisyllabic words Closed 65 dB SPL/quiet NR/NC ns
Own linear Threshold: NR
Attack time: NR
Release time: NR
Ratio: NR
Volume control: fixed
Jenstad et al. (1999) WDRC Threshold: 45 dB
Attack time: 10 ms
Release time: 200 ms
Ratio: varied by DSL prescription
Volume control: listener unable to adjust
Nonsense words Closed 60 dB SPL/quiet (avg. speech at 1 m) NR/NC ns
Sentences Open 60 dB SPL/quiet (avg. speech at 1 m) NR/NC ns
Linear Threshold: NR
Attack time: NR
Release time: NR
Ratio: NR
Volume control: listener unable to adjust
Marriage & Moore (2003) WDRC Threshold: 50–55 dB SPL
Attack time: 5 ms
Release time: 30 ms
Ratio: varied between 1.7: 1 to 2.7: 1
Volume control: NR
Words (consonant confusion task) (severe and profound HL) Closed 60, 50 dBA/quiet, +10 SNR r = .78 [.07, .97]d p = .021 (favors WDRC)
Compression × Presentation Level: ns
Monosyllabic words (severe HL) Open 60, 50 dBA/quiet, +10 SNR r = .999d,e p < .001 (favors WDRC)
Compression × Presentation Level: ns
Linear NR
Marriage et al. (2005) WDRC Threshold: 40 dB at 200 Hz; 25 dB at 4400 Hz
Attack time: 5 ms
Release time: 30 ms
Ratio: 1.2:1 to 3.0:1
Volume control: NR
Monosyllabic words (profound HL) Closed 65 dBA/quiet NR/NC p < .002 (WDRC > linear CL)

ns (WDRC vs. linear PC)
Words (profound HL) Open 65 dBA/quiet NR/NC ns
Common phrases Open 65 dBA/quiet NR/NC ns
Sentences SNR Open In 60 dB modulated noise NR/NC ns
Linear CL Threshold: high
Attack time: 0.5 ms
Release time: 50 ms
Ratio: NR
Volume control: NR
Linear PC Threshold: high
Attack time: 0.5 ms
Release time: 5 ms
Ratio: NR
Volume control: NR
Stelmachowicz et al. (1995) WDRC NR Nonsense syllables Closed 65 dB SPL NR/NC ns
Linear NR (conditions assessed using 2 different devices)
a

Positive effect sizes indicate that the direction of the effect favors WDRC; negative effect sizes indicate that the direction of the effect favors linear amplification.

b

One outlier removed.

c

Effect size rpb (point-biserial correlation coefficient), calculated from individual participant data provided in the study.

d

Effect size r calculated using F statistic and degrees of freedom provided in the study.

e

Confidence intervals were not calculable.