Track cepstrally-smoothed spectrum

Computes a short-term power spectrum smoothed by liftering in the cepstral domain, using the libassp C library (Scheffers 2012) . Cepstral smoothing suppresses spectral fine structure (harmonics), leaving the vocal tract envelope; prefer trk_css_spectrum over trk_dft_spectrum when a smooth spectral envelope is needed without LP assumptions.

Usage

trk_css_spectrum(listOfFiles, beginTime = 0, centerTime = FALSE, endTime = 0, resolution = 40, fftLength = 0, windowShift = 5, numCeps = 0, window = "BLACKMAN", toFile = TRUE, explicitExt = "css", outputDirectory = NULL, assertLossless = NULL, logToFile = FALSE, keepConverted = FALSE, convertOverwrites = FALSE, verbose = TRUE)

Arguments

listOfFiles: Character vector of audio file paths. Any format supported by av is accepted; non-native inputs are transcoded automatically.
numCeps: Integer. Number of cepstral coefficients used for liftering. Default 0 sets this to sample rate in kHz + 1 (minimum 2).
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.
centerTime: Numeric or logical. Single-frame analysis time point in seconds; overrides beginTime, endTime, and windowShift. Default FALSE.
endTime: The end time of the section of the sound files that should be analysed (in seconds). Use 0 for end of file.
resolution: Numeric. Target FFT frequency resolution in Hz; the FFT length is set to the smallest power-of-2 meeting this target. Default 40.0.
fftLength: Integer. Explicit FFT length in points; overrides resolution. Default 0 (use resolution).
windowShift: Numeric. Frame shift in milliseconds; sets output frame rate (1000 / windowShift Hz). Default 5.0 ms (200 Hz). Must be strictly less than 32 ms (the 512-sample analysis window at 16 kHz). Values other than the training default (5 ms) may slightly reduce accuracy.
window: Character. Analysis window function type. Default "BLACKMAN". See AsspWindowTypes for supported types.
toFile: Logical. If TRUE, write SSFF output files and return the count written. If FALSE, return an AsspDataObj (single file only). Default TRUE.
explicitExt: By default, a character "d" will be prepended to the file name suffix when writing the output to file. The user can also specify an explicit extension which will be used instead.
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.
assertLossless: Character vector of additional file extensions to treat as losslessly encoded.
logToFile: Logical. Write processing log to a file in outputDirectory rather than the console. Default FALSE.
keepConverted: Logical. Retain intermediate transcoded files. Default FALSE.
convertOverwrites: Logical. Allow transcoding to overwrite existing files. Default FALSE.
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:

CSS[dB]: REAL32, dB, n_frames x (FFT_length/2 + 1) columns. Cepstrally-smoothed power spectral amplitude from 0 Hz to the Nyquist rate.

Frame rate: 1000 / windowShift Hz (default 200 Hz). If toFile = TRUE: integer count of files written, returned invisibly.

Details

numCeps controls the number of cepstral terms retained before back-transforming; lower values give a smoother envelope. The FFT size is governed by resolution (or overridden by fftLength).

Author

Raphael Winkelmann

Lasse Bombien

Fredrik Nylén

Examples

# get path to audio file
path2wav <- list.files(system.file("samples","sustained", package = "superassp"), pattern = glob2rx("a1.wav"), full.names = TRUE)

# calculate cepstrally smoothed spectrum
res <- trk_css_spectrum(path2wav, toFile=FALSE)
#> Applying `method(trk_css_spectrum, class_character)()` to 1 recording
resolution <- attr(res,"origFreq") / ncol(res[[1]])

# plot spectral values at midpoint of signal
plot(y=res[["CSS[dB]"]][400,],
    x=seq(1,ncol(res[[1]]),1)* resolution,
    type='l',
    xlab='Frequency (Hz)',
    ylab='Amplitude (dB)')