Table 7 Common acoustic features obtained by applying ANOVA, RF and mRMR feature selection methods on phone and web-based recordings

MFCC 5, 11, 12 (mean)

MFCC 6, 9 (std)

ΔMFCC 11, 13 (mean)

MFCC 3 (skew)

ΔMFCC 0, 3, 6, 9, 10 (skew)

MFCC 3, 5 (kurt)

Δlog Mel freq 0, 5, 6 (skew)

ΔMFCC 0 (std)

Voicing prob. (kurt, std)

LSP freq 7 (mean)

ΔVoicing prob. (kurt, mean, std)

LSP freq 2, 3, 4 (skew)

LSP freq 0 (kurt)

LSP freq 1 (kurt)

F0 (skew)

ΔLSP freq 3 (mean)

Jitter local (kurt, skew)

ΔLSP freq 5 (skew)

ΔJitter local (kurt)

log Mel freq 2 (skew)

ΔJitter DDP (kurt)

Δlog Mel freq 1, 2, 3 (std)

ΔShimmer local (kurt)

Voicing prob. (kurt, std)

Loudness (kurt)