QUAD-BASED AUDIO FINGERPRINTING ROBUST TO TIME AND FREQUENCY SCALING
Reinhard Sonnleitner, Gerhard Widmer · 2014
We propose a new audio fingerprinting method that adapts findings from the field of blind astrometry to define simple, efficiently rep-resentable characteristic feature combinations called quads. Based on these, an audio identification algorithm is described that is ro-bust to large amounts of noise and speed, tempo and pitch-shifting distortions. In addition to reliably identifying audio queries that are modified in this way, it also accurately estimates the scaling factors of the applied time/frequency distortions. We experimen-tally evaluate the performance of the method for a diverse set of noise, speed and tempo modifications, and identify a number of advantages of the new method over a recently published distortion-invariant audio copy detection algorithm. 1.