An acoustic cry analysis using frequency and time-frequency tools is developed with the objective of obtain precise and significant information about the events that occur and to analyze the effect of noise on the signals. For hungry baby cry signals, spectrogram, coherence function and wavelet packet decomposition were used on sonorant and non sonorant segments and de-noising techniques were applied to remove the unwanted components. Results show that it can be useful to use those techniques to perform a better analysis of the signal and to eliminate the noise, but the final assessment has to come from the physician’s point of view.
Index Terms. Cry analysis, Spectrogram, Coherence Function, De-noising, Wavelet Packets
Cite as: Martínez, F.M., Azpiroz, J.J., Martínez, A.E. (2007) Analysis of noise in cry signal using frequency and time-frequency tools. Proc. Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2007), 211-214
@inproceedings{martinez07_maveba,
author={F. M. Martínez and J. J. Azpiroz and A. E. Martínez},
title={{Analysis of noise in cry signal using frequency and time-frequency tools}},
year=2007,
booktitle={Proc. Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2007)},
pages={211--214}
}