Miguel-LlamasLanza
commented
1 year ago Likewise, I would like to know if there is any option to extend the signal with more data: I am already truncating my data to be of length 2^L, so I wonder whether the data left outside of the interval corresponding to length 2^L could be used to extend the signal when computing the SWT.
I am using the Stationary Wavelet Transform to perform the "a trous algorithm" on timeseries data. Even if it works as expected when the variance of the data is relatively uniform for all the length of the data, is does not when the variance changes significantly throughout the data. I think that it could be an issue due to edge effects, thus due to the signal extension mode used. However, there is no option to change this mode (or to choose to pass extra data on the edges). Why is this? Is there something conceptually that does not allow to change the mode in the SWT but in the DWT yes? Thank you