Skip to main content
Fig. 4 | EURASIP Journal on Audio, Speech, and Music Processing

Fig. 4

From: Efficient music identification using ORB descriptors of the spectrogram image

Fig. 4

Pitch test for robustness (Shazam vs ORB). This figure involves the pitch test where audio samples were queried to be correctly identified when pitch distortions are applied. Pitch distortion would cause the frequencies in the audio sample to shift vertically. The time synchronization would remain the same, and there are no temporal changes involved. The test was done similarly to the speed where a random 20-s clip was taken from each song in the database, and the pitch distortion was applied. The typical range was tested (−10 to +10%), but the limits of the prototypes were tested (−20 to +20%). It was seen that the ORB prototype produced better results compared to the Shazam algorithm

Back to article page