Skip to main content
. 2011 May;129(5):3281–3290. doi: 10.1121/1.3570957

Table 2.

Correlation coefficients (r) and standard deviations of the prediction error (σe) between sentence recognition scores and various intelligibility measures for vocoded Mandarin Chinese, vocoded English, and wideband (non-vocoded) Mandarin Chinese. The modulation rate (MR) used in the implementation of the NCM measure is indicated.

  Vocoded Mandarin Chinese Non-vocoded Mandarin Chinese Vocoded English
Intelligibility measures Band-weighting function r σe(%) Band-weighting function r σe (%) Band-weighting function r σe (%)
CSII ANSI 0.88 10.2 ANSI 0.90 12.6 ANSI 0.94 8.7
CSIIhigh ANSI 0.74 14.5 ANSI 0.97 6.5 ANSI 0.88 12.4
CSIImid ANSI 0.90 9.6 ANSI 0.95 8.8 ANSI 0.95 8.4
CSIIlow ANSI 0.59 17.5 ANSI 0.65 21.7 ANSI 0.78 16.5
CSII W2, p = 0.5 0.89 9.7 W2, p = 0.5 0.93 10.6 W2, p = 0.5 0.95 8.6
CSIIhigh W2, p = 4 0.91 9.2 W2, p = 4 0.97 7.2 W2, p = 0.05 0.88 12.4
CSIImid W2, p = 0.5 0.91 9.1 W2, p = 0.5 0.95 8.9 W2, p = 0.5 0.95 8.2
CSIIlow W2, p = 4 0.72 15.1 W2, p = 1 0.71 19.9 W1, p = 4 0.86 13.4
NCM Wi(2), p = 0.12 (MR = 100 Hz) 0.94 7.5 Wi(1), p = 1.5 (MR = 100 Hz) 0.96 8.0 Wi(1), p = 1.5 (MR = 100 Hz) 0.96 7.1
ANSI (MR = 12.5 Hz) 0.85 11.5 ANSI (MR = 12.5 Hz) 0.82 16.3 ANSI (MR = 12.5Hz) 0.86 13.4