Skip to contents

Per-GCI voice-quality summary via voiceanalysis

lst_vq_vat.Rd

Extracts the canonical voice-source parameters (Kane and Gobl 2013) from an audio file using voiceanalysis: NAQ (Normalised Amplitude Quotient), QOQ (Quasi-Open Quotient), H1H2 spectral tilt, and HRF (Harmonic Richness Factor). Internally runs SE-VQ for GCIs, IAIF for the glottal flow derivative, then vat_voice_quality().

Usage

lst_vq_vat(listOfFiles, beginTime = 0, endTime = 0, toFile = FALSE, return_jstf = FALSE, explicitExt = "vqv", outputDirectory = NULL, verbose = TRUE)

Arguments

listOfFiles

Character vector of audio file paths.

beginTime, endTime

Analysis window (seconds). Defaults: full file.

toFile

Logical. If TRUE, persist results as JSTF and return paths. Default FALSE.

explicitExt

File extension. Default "vqv".

outputDirectory

Output directory. Default NULL = next to input.

verbose

Logical. 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 and length(listOfFiles) == 1: a named list with vectors gci_time (seconds), NAQ, QOQ, H1H2, HRF. For multiple files, a list of such lists. If toFile = TRUE: invisible vector of output paths.

Details

Returns one row per GCI. To get continuous-track equivalents, interp to a 10 ms grid via stats::approx.

References

(Kane and Gobl 2013)

See also

lst_covarep_vq