praat_sauce: Call the 'praat_sauce' analysis bundle to generate SSFF...

In humlab-speech/superassp: A speech signal processing using various framworks using a wrassp-like interface

Call the 'praat_sauce' analysis bundle to generate SSFF tracks

Description

This function applies the praat_sauce bundle of Praat scripts on a single file and puts the resulting data as tracks in an SSFF file. By default, all analyses are applied to a windowed portion of the signal every 5ms.

Usage

praat_sauce(
  listOfFiles,
  beginTime = NULL,
  endTime = NULL,
  windowShift = 5,
  windowSize = 25,
  minF = 50,
  maxF = 300,
  formantTracking = TRUE,
  numFormants = 5,
  maxFormantHz = 5000,
  nominalF1 = 500,
  nominalF2 = 1500,
  nominalF3 = 2500,
  preEmphFrom = 50,
  useBandwidthFormula = FALSE,
  channel = 1,
  resample_to_16k = TRUE,
  toFile = TRUE,
  explicitExt = "psa",
  outputDirectory = NULL,
  verbose = FALSE,
  praat_path = NULL
)

Arguments

listOfFiles	A vector of file names
beginTime	The time in the sound file where analysis should start.
endTime	The last time point to be included in the analysed sample. If zero (0), the sound file will be included until the end.
windowShift	The time shift to the next analysis window. Defaults to every 5ms.
windowSize	The analysis window length (in ms).
minF	The minimal f0 to search for.
maxF	The maximum f0 to search for.
formantTracking	Boolean; Should the formant tracking abilities of Praat be used? Defaults to TRUE. If disabled, the raw window-by-window formant values will be used.
numFormants	The number of formants to be found within the frequency space.
maxFormantHz	The cutoff frequency used when finding the numFormants formants.
nominalF1	The nominal F1 used in formant tracking.
nominalF2	The nominal F2 used in formant tracking.
nominalF3	The nominal F3 used in formant tracking.
preEmphFrom	The frequency from which pre-emphasis will be applied.
useBandwidthFormula	Should the bandwidth calculation metod of \insertCiteHawks.1995.10.1121/1.412986;textualsuperassp be used, instead of Praat's internal algorithm. Defaults to TRUE (use Hawks & Miller's method).
channel	Which channel to use analyse. Defaults to the first channel.
resample_to_16k	Resample the signal to 16000 Hz before processing? Defaults to TRUE.
toFile	Should the SSFF track file be written to disk? Defaults to true, which by default means that it will be written as a file with a explicitExt extension next to the sound file.
explicitExt	The default file extension to use for the SSFF file.
outputDirectory	A path to an alternative output directory.
verbose	Verbose output. Not currently used.
praat_path=NULL

Value

This function builds an SSFF track object and writes it to disk, or returns it (toFile==FALSE). The track object will contain tracks with these fields:

f0

A track of the fundamental frequency computed by Praat from the signal as part of the PraatSauce analysis.

fm

Computed formant frequency measures, optionally tracked, one column per formant.

bw

Computed formant bandwiths, one column per formant.

H

The amplitudes of the first three harmonics, computed without accounting for the influence of adjecent formants (uncorrected).

Hc

The amplitudes of the first three harmonics, in which the influence of adjecent formants has been accounted for (corrected).

A

The amplitudes of the three harmonics that are closest to the center frequency of the first three formants, computed without accounting for the influence of the adjecent formants (uncorrected).

Ac

The amplitudes of the three harmonics that are closest to the center frequency of the first three formants, in which the influence of the adjecent formants has been accounted for (corrected).

HH

The differences between the (uncorrected) amplitudes of the first and second (column 1) and second and fourth (column 2).

HHc

The differences between the amplitudes of the first and second (column 1) and second and fourth (column 2), computed with correction with regards to neighbouring formants.

HA

The differences between the (uncorrected) amplitudes of the harmonics closest to the center frequencies of the first and second (column 1) and second and fourth (column 2) formants.

HAc

The differences between the amplitudes of the harmonics closest to the center frequencies of the first and second (column 1) and second and fourth (column 2) formants, corrected for the influence of the adjecent formants themselves.

cpp

A track of containing the smoothed Cepstral Peak Prominence \insertCiteFraile.2014.10.1016/j.bspc.2014.07.001,Hillenbrand.1994.10.1044/jshr.3704.769superassp across the acoustic signal.

hnr

Harmonic-to-noise ratios as defined by \insertCiteKrom.1993.10.1044/jshr.3602.254superassp for frequencies 0-500 Hz, 0-1500 Hz, 0-2500 Hz, and 0-3500 Hz, respectively. (four columns)

References

\insertAllCited

humlab-speech/superassp documentation built on May 8, 2024, 2:27 p.m.