Dtw transition sample #2

Dynamic Time Warping (DTW) is an algorithm used to measure the similarity between two temporal sequences that may vary in speed or timing. It aligns the sequences in a way that minimizes the distance between them, allowing for flexible matching of points even if they are out of phase. DTW is commonly applied in various fields, including speech recognition, data mining, and time-series analysis, particularly in applications involving time-dependent data. I use this algorithm to find optimal matches in terms of similarity between points in two different time-series or musical sequences. By identifying the most similar points, I can create a linear interpolation between them, enabling a smooth transition between two musical states over an arbitrary number of steps. Here is e demo, a transition between an initial musical state, which is a sequence of incrementing frequencies, and a second state that is the reverse of the beginning sequence, achieved over 50 steps.

Demo Dtw_transition spectrogram x 25 (Soundcloud removed this one, because it 'might contain or be a copy of "149th Street Groove" by Invisible Man, which is owned by Brooklyn Music Limited'! :-D)

Amtework · dtw_sin

Demo Dtw_transition x 50

Amtework · twelve_session

Demo Dtw_transition x 12

Amtework · dtwsix_session

Demo Dtw_transition x 6

Demo Dtw_transition spectrogram x 50

Demo Dtw_transition spectrogram x 25

Demo Dtw_transition spectrogram x 12

Demo Dtw_transition spectrogram x 6

okavango 03/01/2025 11:44:53