(Color online) TFR of a Beluga sound (called “Creak whistle” by the biologists). The arrows indicate the location of the pulsed and tonal components.
(Color online) The DCT basis for a transformation of size 8 × 8. The indexes along the top and left side of the image represent the vertical and horizontal spatial frequency coefficient indexes. The arrows indicate the horizontal and vertical coefficients used for textural analysis.
(Color online) Evolution of the estimated PTR when the theoretical PTR increases. PTR estimates with the proposed method in the case where TFR satisfies marginal properties (DCT based), and with the mixed derivative thresholding (CIFderiv based) are shown. Signal length: 0.5 s, T pulsed = 5 ms, , Nh = 1. The value changes in the range A = [1.5,…,10] and ρ pulsed = 70 in one case and in the range A = 2.5 and ρ pulsed = [2,…,250] in the other.
(Color online) Simulated sound of 1.5 s. Spectrogram (top); PTR evolution (bottom). The first 0.5 s contain a pulse density of 20 pulses/s, the next 0.5 s contain a pulse density of 50 pulses/s, and the third 0.5 s contain a pulse density of 100 pulses/s.
(Color online) Simulated sound of 2 s. Spectrogram (top); PTR evolution (bottom). The number of 10 kHz harmonics increases by one every 0.5 s.
Whistle creak (top) and Creak whistle (bottom).
TFR of Beluga sounds s 1 (top) and s 2 (bottom).
(Color online) Box and Whisker plot (median ±25th and 75th percentiles) for the Whistle creak and Creak whistle vocalizations.
(Color online) Sound recording composed of a Creak whistle followed by pulsed bursts (top) and sound recording composed of a mix of flat whistles, with different number of harmonics, and pulsed bursts (bottom).
Marginal constants K 1 and K 2 for some discrete-time, discrete-frequency TFRs found in the bibliography.
Measured PTR of the different sounds.
Article metrics loading...
Full text loading...