Extract harmonic model phase distortion features (HMPD): AE, PDM, PDD

Returns amplitude envelope (AE), phase deviation mean (PDM), and phase deviation deviation (PDD) per frame using the HMPD method (Degottex et al. 2014) . Suitable for voice quality characterization and vocoder-based resynthesis.

Usage

trk_hmpd(listOfFiles, f0s = NULL, beginTime = 0, endTime = 0, f0min = 60, f0max = 440, toFile = FALSE, explicitExt = "hpd", 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.

f0s

Optional numeric matrix with 2 columns (time in seconds, F0 in Hz). If NULL (default), a constant 100 Hz F0 is assumed with a warning; supply an accurate F0 track for reliable PDM/PDD values.

f0min

Numeric. Minimum F0 in Hz for sinusoidal analysis. Default 60 Hz.

f0max

Numeric. Maximum F0 in Hz for sinusoidal analysis. Default 440 Hz.

toFile

Logical. If TRUE, write SSFF output files and return the paths written invisibly. If FALSE, return an AsspDataObj. Default FALSE.

explicitExt

Character. Output file extension. Default "hpd".

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 tracks:

ae

FLOAT, Mel-frequency-warped cepstral amplitude envelope, n_frames × 25 (coefficients 0–24). Dimensionless (log-amplitude units).

pdm

FLOAT, phase deviation mean on a log-harmonic scale, n_frames × 25. Units: radians.

pdd

FLOAT, phase deviation deviation (aperiodicity), n_frames × 13. Larger values indicate more noise-like phase scatter.

Frame rate: 100 Hz (10 ms grid, resampled from variable-rate sinusoidal analysis). If toFile = TRUE: character vector of output file paths, returned invisibly.

Details

HMPD captures both spectral shape (AE) and aperiodicity-related phase scatter (PDM, PDD). Phase deviations are computed relative to a deterministic sinusoidal model at each harmonic; the mean and standard deviation across harmonics give PDM and PDD.

When f0s is NULL the function falls back to a constant 100 Hz F0 and emits a warning. AE is computed by Mel-cepstrum (method 2 in COVAREP). PDM uses a log-harmonic scale with 8 harmonics in the linear region (Bezier interpolation above). PDD is derived from circular phase variance.

References

Degottex G, Kane J, Drugman T, Raitio T, Scherer S (2014). “COVAREP: a collaborative voice analysis repository for speech technologies.” In Proceedings of the 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 960–964. doi:10.1109/ICASSP.2014.6853739 . Open-source toolkit for voice analysis including HMPD, GCI detection, formants, pitch, and voice quality measures, https://github.com/covarep/covarep.

Examples

if (FALSE) { # \dontrun{
# Single file, return object
hmpd <- trk_hmpd("speech.wav", toFile = FALSE)

# Batch process multiple files
files <- c("file1.wav", "file2.wav")
hmpd_results <- trk_hmpd(files, toFile = TRUE, outputDirectory = "output/")
} # }