Hi folks,

I already posted this in an answer to another thread, but i feel it warrants it's own thread………

Please consider the two pictures below.

In the first, we have a cube and sphere intersecting, and i want to delete part of the cube. ………. if i use a boolean subtract on it (and change the sphere to polygon), it'll take a bite out of the cube which is roughly a quarter of the volume of the sphere.

But now take a look at the second picture. This is exactly the same scene, only the camera has been moved to a different position.

Again, I want to delete part of the box. But this time, instead of using a boolean, i want to employ a “laser” cutter or knife which fires out of our camera, and will cut away only the part where we can see the sphere and box intersecting from this camera angle. It will keep on cutting in a straight line from the camera, all the way through the cube.

So we'll end up with a lot less volume removed from the box, and with a very different shape.

And although we “know” that there is a much larger area than we can see where the sphere and box intersect, only the part we can actually see will be removed.

Is something like this possible?

And sorry if this is a totally weird question!

Yes it is possible…you just have to do a few steps which I can't describe here since it's been a while.

You can set up a construction plane to an orientation - and from that create your geometry outline on that plane - from there extrude and/or translate, etc.

To get the construction plane at the desired angle/orientation you can build it from 3 points (there's couple steps I forget while working in viewport to do that).

I miss working like this (the way you want) from Solidworks which is much more refined and easy to use. Especially good for industrial work, and designing shapes and forms based and numerical relationships.

The last time I was working this way you couldn't save the construction planes you created to persist from hip session to hip session.

I think I put a RFE for that, don't know if it got implemented.

Hi BabaJ,

Thanks for the info.

I imagined this must be something built into Houdini, but as you describe it, it sounds like you need to jump through some hoops to get it done.

Strange, i'd imagine for quicker rendering it would be sensible to have an option not to calculate any geometry that's hidden from the camera.

Anyway, after i made this post i discovered the OpenGL Driver output, which will kinda allow me to do the same thing using some masks (i'm working mostly on graphic design type stuff where i just want to export what i'm seeing in the viewport, but in different layers)

Cheers! ……….. D