Skip to main content
EURASIP Journal on Advances in Signal Processing

Table 1 Proposed combination of features

From: Rate-constrained source separation for speech enhancement in wireless-communicated binaural hearing aids

SET

NFtSet

Features

SET1

3

A L , (A L -A R )2, (ϕ R -ϕ L )2

SET2

3

A L , abs(A L -A R ), abs(ϕ R -ϕ L )

SET3

4

A L , A L 2 , (A L -A R )2, (ϕ R -ϕ L )2

SET4

2

(A L -A R )2, (ϕ R -ϕ L )2

SET5

7

A L , A R , A L 2 , A R 2 , A L ·A R , abs(A L -A R ), abs(ϕ R -ϕ L )

SET6

6

A L , A L 2 , abs(A L -A R ), (A L -A R )2, abs(ϕ R -ϕ L ), (ϕ R -ϕ L )2

  1. A L and ϕ L are the amplitude of the STFT of the left signal, respectively; A R and ϕ R are the amplitude and phase of the STFT of the right signal, respectively.
Back to article page