pc_envelope(patch) not considering Z values

[strk]

should it ? is Z dimension mandatory ?

[pramsey]

Not to my mind. The primary "spatial" thing about pointclouds is x/y. The amount of variability in Z is minimal, so it's not important to add to the envelope.

[Remi-C]

I think precisely the opposite !
I have a workaround to filter based on Z, but it is cumbersome (at best).
Common, outputting Z coordinate in the enveloppe won't hit the perf, and this make pgpointcloud usefull with other pointcloud than classical aerial Lidar data.
If you are really concerned about compatibility, why not fabricate an optional argument to output or not Z in the geom?

(Sorry @pramsey , but for terrestrial Lidar data, Using Z to filter makes you gain a 10factor.)
Cheers!

[pramsey]

I defer to the person w/ actual experience.