Aligns a candidate sound to a reference sound using MFCC-based dynamic time warping.
Usage
sounds_to_dtw(
reference,
candidate,
analysis_width = 0.015,
time_step = 0.005,
band = 0,
slope = 3
)Arguments
- reference
Reference sound (prototype, y-axis)
- candidate
Candidate sound (test, x-axis)
- analysis_width
Window length in seconds (default: 0.015)
- time_step
Time step between frames (default: 0.005)
- band
Sakoe-Chiba band width in seconds (0 = no constraint)
- slope
Slope constraint: 1=none, 2=1/3-3, 3=1/2-2, 4=2/3-3/2
Examples
ref <- Sound$create_tone(frequency = 200, duration = 0.3, sampling_rate =
16000)
test <- Sound$create_tone(frequency = 220, duration = 0.3, sampling_rate =
16000)
dtw <- sounds_to_dtw(ref, test)
print(dtw$get_weighted_distance())
#> [1] 93.25283