Cepstral Peak Prominence Smoothed (CPPS) — trk

Extracts time-series Cepstral Peak Prominence Smoothed (CPPS) via Praat's PowerCepstrogram. CPPS quantifies voice periodicity and is a robust correlate of breathiness and dysphonia. Prefer this over instantaneous CPP when temporal smoothing is desired.

Usage

trk_cpps(listOfFiles, beginTime = 0, endTime = 0, minF = 60, maxF = 333, timeStep = 0.002, maximumFrequency = 5000, preEmphFrom = 50, windowShape = "Hanning", relativeWidth = 1, subtractTilt = TRUE, timeAveragingWindow = 0.02, quefrencyAveragingWindow = 5e-04, interpolation = "parabolic", trendLineQuefrencyMin = 0.001, trendLineQuefrencyMax = 0.05, trendType = "exponential decay", fitMethod = "robust", toFile = TRUE, explicitExt = "cps", outputDirectory = NULL, verbose = TRUE)

Arguments

listOfFiles: Character vector of audio file paths. Any format supported by av is accepted; non-native inputs are transcoded automatically.
minF: Numeric. Lower quefrency bound for cepstral peak search, in Hz (as reciprocal of quefrency). Sets the minimum F0 detectable. Default 60 Hz.
maxF: Numeric. Upper quefrency bound for cepstral peak search, in Hz. Sets the maximum F0 detectable. Default 333 Hz.
timeStep: Numeric. Frame shift for the PowerCepstrogram in seconds. Sets output frame rate (1 / timeStep Hz). Default 0.002 s (500 Hz).
maximumFrequency: Numeric. Highest frequency included in the cepstrum in Hz. Default 5000 Hz. Set to 0 for Nyquist.
preEmphFrom: Numeric. Pre-emphasis onset frequency in Hz. Default 50 Hz.
windowShape: Character. Window shape applied to each analysis frame. Default "Hanning".
relativeWidth: Numeric. Relative width of the analysis window. Default 1.0.
subtractTilt: Logical. If TRUE, subtract the fitted spectral tilt trend before measuring peak prominence (gives CPPS rather than CPP). Default TRUE.
timeAveragingWindow: Numeric. Duration of the smoothing window along the time axis in seconds. Default 0.02 s.
quefrencyAveragingWindow: Numeric. Width of the smoothing window along the quefrency axis in seconds. Default 0.0005 s.
interpolation: Character. Peak interpolation method: one of "none", "parabolic", "cubic", "sinc70", "sinc700". Default "parabolic".
trendLineQuefrencyMin: Numeric. Minimum quefrency (s) for trend line fitting. Default 0.001 s.
trendLineQuefrencyMax: Numeric. Maximum quefrency (s) for trend line fitting. Default 0.05 s.
trendType: Character. Shape of the fitted trend: "straight" or "exponential decay". Default "exponential decay".
fitMethod: Character. Regression method: "robust", "least squares", or "robust slow". Default "robust".
toFile: Logical. If TRUE, write SSFF output files and return the paths written (invisibly). If FALSE, return an AsspDataObj. Default TRUE.
explicitExt: Character. Output file extension. Default "cps".
beginTime: Start time for the extracted portion in seconds. Default: NULL (beginning of signal). Note: uses beginTime/endTime (seconds) matching DSP function conventions, unlike read_audio() which uses begin/end.
endTime: The end time of the section of the sound files that should be analysed (in seconds). Use 0 for end of file.
outputDirectory: The directory where the slice file should be stored. If not defiled (NULL), the sparse slice file will placed in the same folder as the media file.
verbose: Logical. Show a progress bar (sequential path) or a progress-aware parallel apply (pbapply/pbmcapply, if installed).

Value

If toFile = FALSE: an AsspDataObj with track:

cpp: REAL32, dB, n_frames x 1. Cepstral Peak Prominence Smoothed. Higher values indicate more periodic (healthier) phonation.

Frame rate: 1 / timeStep Hz (default 500 Hz). If toFile = TRUE: character vector of output file paths, returned invisibly.

Details

CPPS is computed via Praat's PowerCepstrogram. Each frame's cepstral peak prominence is measured relative to a fitted trend line (removing spectral tilt), then smoothed over timeAveragingWindow and quefrencyAveragingWindow. Typical values: 15–25 dB for normal voice; below 10 dB for breathy or dysphonic voice.

References

(Hillenbrand et al. 1994)

(Hillenbrand and Houde 1996)

(Heman-Ackah 2003)

Examples

if (FALSE) { # \dontrun{
# Extract CPPS from audio file
result <- trk_cpps("speech.wav", toFile = FALSE)

# Plot CPPS over time
plot(result$cpp, type = "l", main = "CPPS", ylab = "CPP (dB)", xlab = "Frame")

# Custom pitch range for female speaker
result <- trk_cpps("speech.wav", minF = 100, maxF = 400, toFile = FALSE)

# Batch process multiple files
trk_cpps(c("f1.wav", "f2.wav", "f3.wav"), toFile = TRUE)
} # }