Audio supplementary material
for the paper entitled
"Neuromimetic Sound Representation for Percept Detection and Manipulation"
by D. N. Zotkin, T. Chi, S. A. Shamma, and R. Duraiswami,
University of Maryland, College Park, MD 20742 USA.
The paper is to appear in the EURASIP Journal on Applied Signal Processing in 2nd quarter, 2005.
You can click on any spectrogram on this page to hear the corresponding sound.
See also the important footnote at the bottom of this page.
1. Simple auditory spectrogram manipulations.
2. Audio illustration to Fig. 5: Cortical representation is obtained by decomposing sound into individual sound ripples of different scale and rate.
How do the ripples with variable scale sound?
- rips2r03.wav: Sound ripple, scale = 2 cycle per octave, rate = 3 Hz.
- rips4r03.wav: Sound ripple, scale = 4 cycle per octave, rate = 3 Hz.
- rips8r03.wav: Sound ripple, scale = 8 cycle per octave, rate = 3 Hz.
How do the ripples with variable rate sound?
- rips2r03.wav: Sound ripple, scale = 2 cycle per octave, rate = 3 Hz.
- rips4r06.wav: Sound ripple, scale = 2 cycle per octave, rate = 6 Hz.
- rips8r12.wav: Sound ripple, scale = 2 cycle per octave, rate = 12 Hz.
3. Audio illustration to Fig. 10: Timbre-preserving pitch shift of the speech signal.
4. Original auditory spectrogram of the sound and selective removal of "neurons" corresponding to high scales, fast rates, and slow rates, respectively.
5. Synthesis of a "talking" musical instrument by patching up the timbre of female voice and the pitch of oboe in the cortical representation.
6. Audio illustration to Fig. 11, 12: few music samples of synthesized instrument TRUMTAR, which timbre is in-between the guitar and the trumpet.
Footnote: Please note that some audio samples on this page are purposefully made to sound NOT like a normal speech. Instead, they show what happens to the speech sound if the spectrum is modified in a somewhat arbitrary way. These samples include sounds obtained with spectrum compression and spectrum dilation, which do not preserve the harmonic structure of the speech, and the sound obtained with spectral smearing.
Created October 2 2002, last modified March 7, 2005.
