Yeah, they are very versatile. I once constructed high dimensional toy data sets with many uninformative dimensions where having the same widths for each dimension didn‘t really work well, but those were designed to give an edge to my own method :)