Need some help with Convolution

[Evilduck]

I've found the problem,

the 1D convolution operation leaves a number of unncesseary (and extra) entries in the beginning of the post convolution array as a result of the algorithm used. the number of extra terms is equal to the size of the kernel (in this case 1,-1) divided by 2. If I had convolved with a 4 term wavelet, I would have to remove 2. The best way to do it is to use the split array, and split with the size of wavelet/2. the second subarray is equal in size to the original and has no initialization issues.