Skip to contents

PowerCepstrogram

Source: R/powercepstrum.R
PowerCepstrogram.Rd

A Praat PowerCepstrogram: a power cepstrum computed at every time frame of a sound, forming a quefrency-by-time surface.

Arguments

.xptr

Not for direct use. External pointer to the underlying C++ PowerCepstrogram object; set internally when a method returns a new PowerCepstrogram.

Value

A PowerCepstrogram object.

Details

It's the standard input for CPPS (cepstral peak prominence, smoothed), a widely used measure of voice quality: high CPPS means a strong, regular harmonic structure, low CPPS means a noisy or aperiodic voice. Create one with sound$to_powercepstrogram(), then either read get_cpps() directly or slice out a single frame as a PowerCepstrum for closer inspection.

Usage


cepstrogram <- sound$to_powercepstrogram(pitch_floor = 60)

Query methods

  • get_cpp_at_time(time, interpolation, qmin, qmax, fit_method, tolerance) - cepstral peak prominence at a single time

  • get_mean_cpp(from_time, to_time, qmin, qmax, fit_method, tolerance) - mean cepstral peak prominence over a time range

  • get_cpps(subtract_tilt, time_averaging_window, quefrency_averaging_window, pitch_floor, pitch_ceiling, delta_f0, interpolation, qstart_fit, qend_fit, trend_type, fit_method) - CPPS, smoothed over time and quefrency

Transformation and export

  • get_power_cepstrum_at_time(time) - single frame as a PowerCepstrum object

  • smooth(time_averaging_window, quefrency_averaging_window) - smoothed copy as a new PowerCepstrogram

  • to_matrix() - values as a Matrix object

  • as_matrix() - values as a plain numeric matrix (quefrency x time)

See also

[PowerCepstrum], [Sound], [get_cpps_fast]

Examples

sound <- Sound$create_tone(frequency = 150, duration = 0.3)
cepstrogram <- sound$to_powercepstrogram(pitch_floor = 60)
mean_cpp <- cepstrogram$get_mean_cpp()