what would it look like with nearest, or gaussian filters? Would be interesting to see the same sample graphs with different kernels.