Skip to contents

Harvest Pitch Tracking (C++ implementation)

trk_pitch_harvest.Rd

Extract F0 (fundamental frequency) using the Harvest algorithm from WORLD vocoder (via SPTK). This is a high-performance C++ implementation that is 2-3x faster than the Python version and requires no Python dependencies.

Harvest is designed to be robust and accurate for speech analysis, with good performance even on noisy signals.

All input media formats are supported via the av package, including video files from which audio will be automatically extracted.

Usage

trk_pitch_harvest(listOfFiles, beginTime = 0, endTime = 0, windowShift = 10, minF = 60, maxF = 400, voicing_threshold = 0.1, toFile = TRUE, explicitExt = "f0", outputDirectory = NULL, verbose = TRUE, parallel = NULL, n_cores = NULL)

Arguments

listOfFiles

Character vector of audio file paths. Any format supported by av is accepted; non-native inputs are transcoded automatically.

windowShift

Frame shift in milliseconds (default: 10.0)

minF

Minimum F0 in Hz (default: 60.0)

maxF

Maximum F0 in Hz (default: 400.0)

voicing_threshold

Voicing threshold (default: 0.1, valid range: 0.02-0.2 for WORLD/Harvest)

toFile

Write results to file (default: TRUE)

explicitExt

Output file extension (default: "f0")

parallel

Logical. Use parallel processing for multiple files. NULL (default) enables automatically for 2+ files.

n_cores

Integer. Number of cores for parallel processing. NULL (default) uses detectCores() - 1.

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=TRUE, returns the number of successfully processed files. If toFile=FALSE, returns AsspDataObj or list of AsspDataObj objects.

Examples

wav <- system.file("samples", "sustained", "a1.wav", package = "superassp")

# \donttest{
f0 <- trk_pitch_harvest(wav, toFile = FALSE, verbose = FALSE)
head(as.data.frame(f0))
#>   frame_time f0
#> 1       0.00  0
#> 2       0.01  0
#> 3       0.02  0
#> 4       0.03  0
#> 5       0.04  0
#> 6       0.05  0

# Custom F0 range
trk_pitch_harvest(wav, minF = 75, maxF = 300, toFile = FALSE, verbose = FALSE)
#> In-memory Assp Data Object
#> Format: SSFF (binary)
#> 404 records at 100 Hz
#> Duration: 4.040000 s
#> Number of tracks: 1 
#> 	 f0 (1 fields)
# }