A comparison between the Dutch and American-English digits-in-noise (DIN) tests in normal-hearing listeners

Figures & data

Table 1. Properties of the NL DIN and US DIN.

	NL DIN	US DIN
Number of available triplets	120	50
Digits	0, 1, 2, 3, 4, 5, 6, 7, 8, 9	1, 2, 3, 4, 5, 6, 8, 9
Style	concatenated digits	naturally spoken three-digit sequences
Speaker	male	female
Noise tokens	fresh	time-locked
F0*	129.1 Hz	208.9 Hz
RMS level original digit-triplets (including silences) re RMS level noise	−2.36 dB	0.37 dB
RMS level digit-triplets (silences removed) re RMS level noise	0.66 dB	2.18 dB

*Praat software (Boersma and Weenink, 2013) was used to determine the F0 from the female English talker and the male Dutch talker.

Table 2. Average SRT (uncorrected SRTs, with corrected values between brackets) for the DIN tests, standard error of the mean (SE), and standard error of measurement (SEM). Data from experiment 1.

		NL DIN			US DIN
	Condition	SRT [dB SNR] (corrected)	SE [dB]	SEM [dB]	SRT [dB SNR] (corrected)	SE [dB]	SEM [dB]
Steady-state	Left	−9.3 (−8.6)	0.11	0.7	−10.7 (−8.6)	0.22	0.9
Noise	Right	−9.1 (−8.5)	0.18	0.7	−10.7 (−8.5)	0.16	0.6
	Diotic	−10.0 (−9.3)	0.17	0.5	−11.2 (−9.0)	0.16	0.9
	Dichotic	−15.3 (−14.7)	0.13	0.9	−17.1 (−14.9)	0.21	0.9
Interrupted	Left	−15.0 (−14.3)	0.25	1.0	−17.8 (−15.6)	0.26	1.0
Noise	Right	−15.0 (−14.4)	0.18	1.1	−17.9 (−15.7)	0.24	1.1
	Diotic	−15.6 (−15.0)	0.23	1.2	−19.0 (−16.8)	0.23	0.8
	Dichotic	−20.7 (−20.1)	0.17	1.1	−22.4 (−20.2)	0.26	0.9

Figure 1. Top panels show boxplots of the test-retest SRT averages of the original DIN test results (i.e. with speech and noise levels as defined in the original papers). Results for the NL DIN are shown with filled boxes, results for the US DIN with unfilled boxes. The left panel shows the results for steady-state noise and the right panel for interrupted noise. Bottom panels show the same data but with corrected SRTs.

Table 3. Slope (S) of the fitted speech recognition functions; standard error (SE) between brackets. Slope values were determined from a maximum likelihood fit to the raw data and from fit to the raw data after correction for the individual SRT.

		S [%/dB]
		NL DIN		US DIN
			Condition	Raw data	Corrected for SRT	Raw data	Corrected for SRT
Steady-state	Monaural	15.7 (0.9)	19.3 (1.1)	12.2 (0.8)	16.0 (0.9)
Noise	Diotic	14.7 (1.3)	18.4 (1.5)	12.3 (1.1)	15.9 (1.3)
	Dichotic	13.2 (1.2)	16.8 (1.4)	10.0 (1.0)	14.1 (1.2)
Interrupted	Monaural	10.1 (0.7)	13.6 (0.8)	10.0 (0.7)	14.7 (0.9)
Noise	Diotic	9.3 (0.9)	13.5 (1.2)	11.2 (1.0)	14.6 (1.2)
	Dichotic	12.0 (1.1)	15.8 (1.3)	9.4 (0.9)	12.8 (1.1)

Figure 2. Boxplots of the corrected test-retest SRT averages for the NL DIN, US DIN, and US DIN with modified stimuli (i.e. disturbed natural coarticulation and prosody). The left panel shows the results for steady-state noise and the right panel for interrupted noise.

Table 4. Average SRT (uncorrected SRTs, with corrected values between brackets) for the DIN tests and standard error of the mean (SE). Data from experiment 2.

		NL DIN		US DIN
				Original stimuli		Without coarticulation/prosody
					Condition	SRT [dB SNR] (corrected)	SE [dB]	SRT [dB SNR] (corrected)	SE [dB]	SRT [dB SNR] (corrected)	SE [dB]
Steady-state	Diotic	−9.7 (−9.1)	0.20	−11.2 (−9.0)	0.25	−11.0 (-8.8)	0.19
Noise	Dichotic	−15.4 (−14.7)	0.36	−16.8 (−14.6)	0.24	−16.2 (−14.0)	0.22
Interrupted	Diotic	−15.5 (−14.8)	0.35	−18.7 (−16.5)	0.18	−18.4 (−16.2)	0.30
Noise	Dichotic	−20.4 (−19.7)	0.30	−22.0 (−19.8)	0.32	−21.8 (−19.6)	0.42

Boersma P. & Weenink D. 2013. Praat: Doing Phonetics by Computer [Computer Program]. Version 5.3.41.,

 Google Scholar