Skip to main content

Table 5 Top three LLD with their highest ranked functional for CC-based ranking and CC-based ranking after singer normalisation of features (SPKSTD-CC) as well as CDCC-based ranking; Pearson correlation coefficients given in parentheses for each feature; EmoFt (top) and ComParE (bottom) feature sets

From: Emotion in the singing voice—a deeperlook at acoustic features in the light ofautomatic classification

Arousal

Valence

SPKSTD-CC, EmoFt

 

Jitter, fifth percentile (0.55)

F 0, 95th percentile (−0.21)

Shimmer, mean (0.52)

Jitter, range (−0.19)

Loudness, pos. mean (0.49)

Voice prob., mean (0.19)

Weninger-CDCC, EmoFt

 

Jitter, fifth percentile (0.56)

Voice prob., fifth percentile (0.20)

Shimmer, mean (0.53)

F 0, fifth percentile (0.18)

F 0, range (0.49)

Loudness modulation, max

 

freq. (0.16)

CC, EmoFt

 

Jitter, fifth percentile (0.46)

F 0 δ, pos. mean (−0.18)

Shimmer, mean (0.44)

Jitter δ, 95th percentile (−0.17)

Loudness δ, pos. mean (0.43)

F 0, 95th percentile (−0.16)

SPKSTD-CC, ComParE

 

Jitter, first quartile (0.60)

MFCC 5, third quartile (−0.322)

Log. HNR, pos. mean (−0.59)

log. HNR, skewness (−0.24)

Shimmer, first quartile (0.59)

RASTA f. band 10, LP gain (−0.24)

Weninger-CDCC, ComParE

 

Jitter, first quartile (0.61)

MFCC 5, third quartile (0.26)

Shimmer, first quartile (0.59)

log. HNR, skewness (0.26)

Jitter DDP, second quartile (0.56)

MFCC 13, LPC2 (0.24)

CC, ComParE

 

Loudness δ, pos. mean (0.51)

RASTA f. band 10, LPC 3 (−0.249)

Spec. centroid δ, IQR 2–3 (0.50)

Spec. variance, seg. len σ (0.20)

Sharpness δ, first quartile (−0.50)

log. HNR, qaud. reg. err. (−0.20)