Proxy mode is great. Use it all the time.
However, in dealing with scientific data, I often have images containing numeric IDs.
Proxy mode seems to use "cubic" as the downsampler, which will improperly 'blend' IDs. I'd like to be able to change the downsampling to "impulse" so that I don't get incorrect IDs when using proxies.
Is there a way of setting what the global proxy filter is?
Even better, is there a way of setting the filter on a READ node, or in a tree (like how concatenated transforms use the filter of the bottom-most node) so that different portions of the tree can use different filter types for the proxy filtering?