Table 1 Selected acoustic speech features
Feature | Description |
|---|---|
Long pauses | The number of unfilled pauses (silences) longer than 2Â s divided by the audio length in seconds. |
Medium pauses | The number of pauses of 1–2 seconds, divided by the audio length in seconds. |
Pause duration | The duration of segments without a speech signal divided by total number of segments without any speech signal in seconds. Includes all segments without any speech signal (including < 150 milliseconds). |
Pause-to-word ratio | The number of segments without any speech signal longer than 150 milliseconds divided by number of segments with a speech signal. |
Phonation rate | The number of segments with a speech signal (in 50 milliseconds windows) over the total number of speech segments, irrespective of audio duration. |
Audio duration | The total length of the audio sample in seconds. |
Fundamental frequency | The mean of the sequence of fundamental frequency values extracted from the audio file in Hertz, using the Parselmouth library (equivalent to Praat method for computing fundamental frequency). The cutoff range is 70–620 Hz. |
Intensity | The mean of the intensity curve (i.e., loudness), relative to 2*10− 5 Pascal (normative auditory threshold for a 1000-Hertz sine wave) in decibel. |
Intensity variance | The variance of the intensity curve (i.e., loudness), relative to 2*10− 5 Pascal (normative auditory threshold for a 1000-Hertz sine wave) in decibel. |
Local shimmer | The average absolute difference between the amplitudes of consecutive periods, divided by the average amplitude, in percentages. |
Local jitter | The average absolute difference between consecutive periods, divided by the average period, in percentages. |