Load Sound Window from File with Optional Resampling

Extracts a time window from a sound file without loading the entire file. Optionally resamples the window to a target sampling rate. Uses LongSound for lazy loading - only the requested window is loaded from disk.

Usage

sound_load_window(path, start, end, resample_to = NULL, preserve_times = FALSE)

Arguments

path

Path to sound file (WAV, AIFF, FLAC, MP3, etc.)

start

Start time of window in seconds

end

End time of window in seconds

resample_to

Target sampling rate in Hz (optional). If NULL, no resampling.

preserve_times

If TRUE, keep original time domain. If FALSE, shift to start at 0 (default: FALSE)

Value

Sound object containing the windowed (and optionally resampled) audio

Details

A traditional workflow loads the entire file, resamples the entire file, and then extracts the window of interest — reading and processing far more samples than needed. This function instead opens the file as a `LongSound` (lazily, reading only the header), extracts just the requested window from disk, and resamples only that window.

**Use cases:** - Pharyngeal analysis: extracting short vowel windows from long recordings - Formant tracking: analyzing specific time points - Batch extraction: processing windows from multiple files - Large file processing: working with hours-long recordings

See also

Sound, LongSound

Other batch-ops: textgrid_merge()

Examples

# Write a short synthetic recording, then load only part of it
sound <- Sound$create_tone(frequency = 220, duration = 2.0)
path <- tempfile(fileext = ".wav")
sound$save(path)

window <- sound_load_window(path, start = 0.5, end = 0.6)

# Extract and resample (for spectral analysis)
window_10k <- sound_load_window(path, start = 0.5, end = 0.6, resample_to =
 10000)

# Preserve original time domain (window starts at 0.5, not 0.0)
window_timed <- sound_load_window(path, start = 0.5, end = 0.6,
 preserve_times = TRUE)

unlink(path)