Skip to contents

Extract Dysprosody Prosodic Features

lst_dysprosody.Rd

Extracts 193 prosodic features using pladdrr-based dysprosody pipeline including F0 analysis (MOMEL/INTSINT), intensity, formants, and spectral tilt measures with Iseli-Alwan harmonic correction.

Usage

lst_dysprosody(listOfFiles, beginTime = 0, endTime = 0, minF = 60, maxF = 750, windowShift = 10, toFile = FALSE, return_jstf = FALSE, explicitExt = "dyp", outputDirectory = NULL, verbose = TRUE)

Arguments

listOfFiles

Vector of file paths to audio files

beginTime

Start time in seconds (default: 0.0)

endTime

End time in seconds (default: 0.0 = full duration)

minF

Minimum F0 in Hz (default: 60)

maxF

Maximum F0 in Hz (default: 750)

windowShift

Window shift in milliseconds (default: 10)

toFile

Write output to .dyp files (default: FALSE)

explicitExt

Output file extension (default: "dyp")

outputDirectory

Output directory (default: NULL = same as input)

verbose

Show progress (default: TRUE)

return_jstf

Logical. Return JsonTrackObj instead of data.frame? Default FALSE. When both toFile and return_jstf are TRUE, the file is written AND the object returned.

Value

If toFile=FALSE: list of data frames (one per file) with 193 features. If toFile=TRUE: invisibly returns vector of output file paths.

Details

Requires: pladdrr >= 4.8.23

Features Extracted (193 total):

  • F0 analysis: MOMEL quadratic spline modeling, INTSINT tonal coding

  • Intensity measures: statistics of intensity contour at INTSINT targets

  • Spectral tilt: 7 measures including Iseli-Alwan harmonic correction (L2L1, L2cL1c, L1cLF3c, L1LF3, SLF, C1, Spectral Balance)

  • Time-series statistics: mean, std, variation, IQR, max, min for all measures and their first differences

  • Global measures: duration, pitch key, pitch range, INTSINT concentration

Performance: ~10-12 seconds per file (optimized via batch queries)

Output Format: When toFile=TRUE, writes JSON Track Format (.dyp) files with all 193 features organized by time slice.

References

(Villarubia and others 2025)

Examples

if (FALSE) { # \dontrun{
# Extract dysprosody features from audio
result <- lst_dysprosody("speech.wav", toFile = FALSE)

# Access specific features
cat("Duration:", result$Duration, "seconds\n")
cat("Pitch mean:", result$momelpitchtmean, "Hz\n")
cat("Pitch key:", result$PitchKey, "Hz\n")
cat("Pitch range:", result$PitchRange, "octaves\n")

# Batch processing with file output
files <- c("speech1.wav", "speech2.wav")
lst_dysprosody(files, toFile = TRUE, outputDirectory = "output")
} # }