Nonlinear methods are usually used in cardiac event detection and classification due to the fact that these methods commonly provide with a single value which can quantify the complexity (chaos theory methods), correlation properties (fractal theory methods) and irregularity (infor- mation theory methods) of the signal. Feature extraction for morphological analysis is also based on WT and statistical methods. Removal on EGM has been achieved using WT ECG signal compressions has been performed using PCA, whereas foetal extraction from maternal-foetal ECG has been developed using, preferably, ICA techniques. Vocal nodules have been seen to occur more often for the untrained singer than the trained. It is certain that, prior to undertaking large research projects into professional singing voice, previous investigation must take place using smaller groups of non-professionals in order to learn and assess any key requirements and considerations that will be essential when looking at professional groups in the future. Singing voice has been deemed to be the area of professional voice that will be examined in the second part of this chapter, although at the current stage we have only performed an exploratory analysis on adolescent music students rather than on adult professionals. Furthermore, analysis of the normal professional voice in action can provide researchers, health specialists and scientists with a more profound insight into the workings of the vocal apparatus in general. The two foci are connected in that professional voice users experience a proportion of voice pathology, as previously mentioned, and a deeper understanding of the dynamics of professional phonation can assist in prevention and treatment of the voice problems that this group experiences. The second focus of the study was on singing voice and the production of singing voice by professionally trained musicians and singers. In the first instance, the focus of the study was on pathological voice and whether or not DSP technologies can provide medics with a robust methodology for diagnosing voice samples from the voice signal alone. The current work describes two specific and separate experimental objectives that will illustrate the main focus of the research that has been carried out over the duration of the thesis study and preparation.
0 kommentar(er)
