speech-transformed-1.wav

The original sound from freesound.org https://www.freesound.org/people/xserra/sounds/317745/ was transformed using sms-tools transformation package to demonstrate the frequency transpose scaling to distort the sound.
The attached sound was first transformed using the sms-tools HPS (Harmonic plus Stochastic Model). The parameters used are as shown below:-
Window = Hamming, M=1101 (window size), N=2048 FFT size, t = -100db threshold, minSinDur = 0.07s, nH = 80 (Max harmonics), minf0 = 120, maxf0=260, foet=7,(Harmonic selection parameters) an stocf = 0.5.
The resulting sound was then transformed using only frequency scaling.
The sound fragment is 4s which was divided into 3 segments. The first segment between 0 and 1.3 scaled the frequencies by 0.5 i.e made one octave lower, the last segment had frequencies scaled by 2.0 i. made one octave higher. The middle segment between 1.3s and 2.7s linearly varied the frequencies from 0.5 to 2.0. The parameters for the transformation model were [0.0,0.5,0.33,0.5,0.67,2.0,1.0,2.0] Note time-value pairs where 1.0 is maxTime = 4.0s