Transform filter memory usage

I have a (rather large) set of segmented imagedata. I’d like to use the Transform filter to explode out the segmented layers, similar to this image:

When I apply the Transform filter, Paraview suffers a crash. It’s possible I’m running out of memory, but I would have expected Transform to be a very light operation. Is it making a copy or something?

It appears that the Transform converts an Image data set to a Curvilinear structured data (makes sense). I think I know how to change the origin point in a Programmable filter. Is there an easier way to do it?

For each individual source/filter, you can use the “Translation” property on the Display properties tab to move it.

Following shows a similar viz as yours (except with sillier dataset :))

Apparently this does weird things when volume rendering:

I think I’m likely to need to get creative to either extract a bunch of surface contours, or combine everything into a single transformed dataset, or something else.

I was surprised that changing the origin of the Image data set didn’t work as I expected. It was almost like it was a translation relative to the viewpoint of the camera, rather than relative to some global coordinate system.

I think the weirdness you are experiencing has nothing to do with the translation. The problem is that ParaView just does not render multiple volumes together correctly. Volumes and solid surfaces work well. Multiple opaque surfaces work well together as well. But with multiple volume renderers, ParaView sort of gives up and just renders them in some arbitrary order. Thus, they are not blended correctly and some just appear on top of each other. This makes the inner shell, for example, look like it is moved closer to the camera, but it is really just blended incorrectly on top of the others.

fyi @LucasGandel

I agree with your assessment. I had hoped to handle the image with volume rendering, but it’s looking like I’ll need to extract surfaces for everything, instead. In that case, the Transform filter issue won’t matter much anymore.