Skip to main content
. Author manuscript; available in PMC: 2019 Apr 2.
Published in final edited form as: IEEE/ACM Trans Audio Speech Lang Process. 2017 Feb 13;25(5):952–964. doi: 10.1109/TASLP.2017.2667879

TABLE II.

Average GPE Performance Gain (%): HarFeature Method vs. the Second Best Performance Method

SNR level − 10 dB − 5 dB 0 dB 5 dB 10 dB 15 dB 20 dB Avg
Keele database 7.97 10.98 11.67 7.94 3.93 1.71 0.84 6.43
CSTR database 4.15  8.87 12.13 7.51 2.43 1.14 0.37 5.23