Skip to main content
. 2021 Jul 30;9(1):32. doi: 10.1007/s13755-021-00162-8

Table 1.

Overview of the extracted features, divided according to the domain of analysis

Region Study Feature Information
Entire signal [22] IEDCC(1–6) Vocal tract and vocal folds abnormalities [22]
[3, 10, 12] Zero crossing rateH Voice activity (Details in [1])
[10, 16, 22] DFAH Self-similarity of the voice (Details in [14])
Voiced [16] BandwidthL Frequency range
[10, 16, 43] Harmonic ratioL Ratio of signal over noise [16]
[10, 16, 43] F0L Vocal folds vibration and frequency alteration
[46] Spectral features: fluxL, skewnessL, entropyL, crestL, flatnessL, slopeL, roll offL, spreadL, centroidL, kurtosisL Spectrum shape information (Details in [1])
[6] LPC(1–3)L Formants and resonances (Details in [1])
[10] Short time energy L Energy variation among frames
[16, 45, 46] MFCC(1–13)L, Δ MFCC(1–13)L, ΔΔMFCC(1–13)L, Subtle changes in the motion of articulators (Details [32, 42])
Transition Present study PTS Ability to promptly interrupt/start vocal fold vibration
Present study ETS Ability to promptly interrupt/start vocal fold vibration
[34] MFCC(1–12), ΔMFCC(1–12), ΔΔMFCC(1–12), Ability to promptly interrupt/start vocal fold vibration
[34] BBE(1–25) Ability to promptly interrupt/start vocal fold vibration

The apex letter represents the classification between LLf and HLf subgroups