A novel efficient algorithm for music transposition

We present a novel efficient algorithm for scaling the frequency content of an audio signal by any desired factor in the range 0.5 (minus one octave) to 2.0 (plus one octave) enabling a recording to be played in any desired key without affecting the tempo. The algorithm uses an adaptive overlap-add (AOLA) technique to realise the desired frequency scaling without affecting the duration. Informal listening tests show output quality equal to that of a conventional overlap-add algorithm used in many commercially available systems, but offering significant computational saving relative to that algorithm. The algorithm is also used to simultaneously scale both the tempo and key of a recording.

[1]  Garry J. Heard,et al.  Genetic Algorithm Inversion of the 1997 Geoacoustic Inversion Workshop Test Case Data , 1998 .

[2]  Ernst Mach,et al.  Sensations of tone. , 1897 .

[3]  A. Wilgus,et al.  High quality time-scale modification for speech , 1985, ICASSP '85. IEEE International Conference on Acoustics, Speech, and Signal Processing.