I am doing research in ECG compression using wavelets. I usually takes samples from MIT arrhythmia data base, maximum 2000 samples. During one of my seminar one person asked me that "what is the frequency of the samples that I am considering". MIT data is sampled at 360 samples per second but he was asking about the samples that I am considering. He told that there is a relation between number of levels of wavelet decomposition and the particular frequency of the signal.. Can any one tell me how the frequency of a signal is related to the levels of decomposition..
2
votes
1 Answers
2
votes
At each level of decomposition, the signal is downsampled by a factor of 2. To satisfy the Nyquist's sampling theorem, your sampling freq should be greater than equal to twice the maximum freq of your input signal. Else aliasing occurs and corrupts your signal. This puts a limit on the factor of downsampling and thus on the number of stages of decomposition. Hope this helps!
