Huggins binaural pitch is a faint tone which can be detected by slightly altering a noise sample played to one of the two ears. The pitch you hear does not actually exist in the signal, but is generated in your head by neural interactions. This illusion is still not fully understood. Please see the PDF document for more details on the efficient generation of Huggins pitch using all-pass filters.
This submission is made of 2 functional files. The huggins.m function is where all the DSP resides and is where the Huggins pitch is generated. The uploaded screenshot shows the phase reversals which can be introduced fom the cascade filtering. The mario_huggins.m file is where the fun stuff is for those who are not too interested in the details.
The Super Mario Brothers theme song is used as a demo here as it has a nice alto and tenor part which coincide in timing and are in a good frequency range to be able to hear the binaural Huggins pitch. To hear the tune place the two m files in the same directory and run the mario_huggins script. Use good headphones and be prepared to wait a few seconds for processing of the song. Concentrate as the pitch is very faint! Mario is in your brain! Try one headphone at a time with a finger in your other ear to make sure and you should just hear noise. If you dont have the signal processing toolbox just knock out the Butterworth filtering at the end of the Mario theme file and all should work fine. Have fun . . .
Elisabet Molin - filter coefficient derivation
Stewart Bozarth - Song Transcription
James Humes - Original Mario song matlab-central submission