Crispness judgement of royal gala apples based on chewing sounds

Chewing sounds were analysed to detect small differences in crispness between Royal Gala apples. Different parameters calculated from the sound waves for bite and chews included amplitude, energy, frequency content after fast Fourier transformation (FFT) combined with data reduction to 100Hz segments and fractal dimension. These parameters were compared to crispness scores given by a trained analytical sensory panel. Recordings of apples were compared with recordings of teeth clacking and chewing sounds of rice crackers. In the frequency domain, the high values in the teeth signal between 700 and 900 Hz and to a lesser extent between 1200 and 1400 Hz affected the rice cracker and apple signals. After FFT and data segmentation per 100 Hz, the amplitude-frequency data of the first bite in apples and the first chew for both apples and rice crackers contained a typical 'hump' at around 4 kHz. The minimum in the spectra at around 3 kHz could be caused by the damping properties of the soft tissues in the mouth. For both apples and rice crackers, the bite contained on average a lower amount of frequency components between 300 and 700 Hz than the consecutive chew. The energy of the first bite gave, depending on the panellist, mostly a better indication of crispness than the penetrometer firmness. The best correlation with sensory crispness could be found by combining the information from different frequency bands after data segmentation (correlation coefficients r up to 0.83). However, the significant frequency bands differed between panellists. Fractal analysis on the normalized amplitude-time data showed that chew sound waves have a higher degree of jaggedness than bite sound waves. The fractal dimension could not be correlated with sensory crispness. (C) 2002 Silsoe Research Institute. Published by Elsevier Science Ltd. All rights reserved.