Pre-allocated the XArray - that might speed things up a bit.
Profile the code - where’s the slow down? (I’d guess it’s still the reprojecting?)
Multi-process / multithreaded. If decompressing is slow (I don’t think it is) then have one thread which just decompresses and leaves uncompressed data on disk ready for reprojecting process.
Pre-allocated the XArray - that might speed things up a bit.
Profile the code - where’s the slow down? (I’d guess it’s still the reprojecting?)
Multi-process / multithreaded. If decompressing is slow (I don’t think it is) then have one thread which just decompresses and leaves uncompressed data on disk ready for reprojecting process.