Lab-4 Pitch Perception

9/21/98

This is an introduction to pitch perception. William Yost and colleagues have taken up the cause of infinitely iterated rippled noise. This is a signal in which none of the components are harmonically related as they are in music and speech (vowels). In any case there are theories of how we perceive pitch. And the requirements for pitch. As you might expect there is not a complete agreement. Is it in the temporal response properties? Or in the spectral coding and left for a higher stage to resolve pitch?

There is an m-file that you are to copy to your folder: iirn.m along with mmctl6.m

cd c:\neuro730\yourfile

Type iirn to start the exercise.

You can select from several types of noise by responding to the display question.

When entering one of the four types of noise you must place the string in single quotes.

An empty figure will appear with a slider on it. Grab the button using the mouse to change the feedback delay which changes the pitch of the rippled noise.

But first size the MATLAB window to be a small portion of the screen. Click on the figure. You can change a parameter from 0 to 200 by dragging the slider. The actual delay is = tau/10 in milliseconds and it determines the pitch of the sound. For example, tau = 100 will result in a pitch of 100Hz for the irn+ condition. Tau is the delay between the noise signal and the noise signal added or subtracted back in to produce the final sound. The amplitude of the delayed sound is set to 0.95 since this value is suppose to give a good pitch sensation.

One type of noise is called rippled noise or cosine-rippled noise. Another is infinitely iterated rippled noise. The equations are straightforward and can be seen in the code.

The sound will repeat 12 times before plotting the final spectra. You can expand the spectra or autocorrelation with the axis command : axis([xlow xhigh ylow yhigh]).

At the end of the 12 presentations, an AM signal with the same pitch as the final rippled noise is presented (100% modulation).

Compare the four type of noise using approximately the same delay for all four cases. Vary the pitch over a wide range, at least, 50 to 1000 Hz. Plot the final results for one pitch for all four types of noise. Write on the plots which one had the strongest pitch sensation for you. Was a pitch percept always present? Hand in your results and impressions of the pitch percept in each case. Scale the plots so that all four spectra can be placed on a single sheet. Select: FILE/PRINT/PROPERTIES/GRAPHICS/SCALE = 50%. You could smooth the plots for presentation purposes.

Yost, W.A. Pitch of iterated rippled noise. J. Acoust. Soc. Am. 100:511-518.

Shofner, W.P. Temporal representation of rippled noise in the anteroventral cochlear nucleus of the chinchilla. J. Acoust. Soc. Am. 90:2450-2466.

Raatgever, J. and Bilsen, F.A. (1992) The pitch of an anharmonic comb filtered noise reconsidered. In: Auditory Physiology and Perception. Eds. Y. Cazals, K. Horner, L. Demany. Pergamon Press, New York. Pp 215-222.