topic Re: Camera position in axonomtery in Archicad C++ API

Camera position in axonomtery

Tomasz Marek — Wed, 02 Aug 2023 13:23:25 GMT

I have hard time getting a position of a camera with axonometric projection.
The 3x3 Transformation gives just info on direction and axonomtery in general.
Camera->GetViewPoint() works only for perspective. How can I figure out a center of a view?

Re: Camera position in axonomtery

Achille Pavlidis — Wed, 20 Mar 2013 12:32:25 GMT

well, there actually isn't a camera position in axonometry. Axonometry is a parallel projection just like a plan or an elevation.

Re: Camera position in axonomtery

Tomasz Marek — Wed, 20 Mar 2013 12:36:57 GMT

How will you match a render with a view in 3D window?
You can pan&zoom 3d view and a Archicad rendering covers just exactly this portion of the view ( camera->GetBounds() ). Direction is not enough, I have to have a point in 3D space to "attach" this vector to.

There is no "camera" position , but a "viewer position" given that the view area doesn't match an exact extents of a model.

Re: Camera position in axonomtery

Achille Pavlidis — Wed, 20 Mar 2013 12:39:30 GMT

maybe post some screenshots, to understand what you are after

Re: Camera position in axonomtery

Tomasz Marek — Wed, 20 Mar 2013 12:49:20 GMT

I think it clarifies it well enough.
Two 400x300 renderings of an axonometry of same model.
Without knowing the position I cannot replicate the view/rendering.

Re: Camera position in axonomtery

Ralph Wessel — Wed, 20 Mar 2013 17:12:55 GMT

Tomasz wrote:
I have hard time getting a position of a camera with axonometric projection.
The 3x3 Transformation gives just info on direction and axonomtery in general.
Camera->GetViewPoint() works only for perspective. How can I figure out a center of a view?

Sorry, that's all the information you can get from the API (as far as I'm aware). It is quite limiting, but reflects the way ArchiCAD works.

Re: Camera position in axonomtery

Tomasz Marek — Thu, 21 Mar 2013 11:10:58 GMT

The ExampleRenderingAddOn supplied with SDK 16 renders the view correctly, so there must be a way of figuring out a center of the view. I am surprised GS didn't supply such a basic functionality in their SDK.

I have spent few hours analyzing the example, but haven't found a solution yet.

Re: Camera position in axonomtery

ivanfi — Mon, 25 Mar 2013 13:51:17 GMT

Calling GetBounds () on the axonometric camera tells you the bounding box of the region that should be displayed. The center of the view is the center of this bounding box.
Hope this helps,
Zoltan

Re: Camera position in axonomtery

Tomasz Marek — Mon, 25 Mar 2013 14:29:06 GMT

ivanfi wrote:
Calling GetBounds () on the axonometric camera tells you the bounding box of the region that should be displayed. The center of the view is the center of this bounding box.

Bounding box gives me just horizontal and vertical size of a visible space. It doesn't give me the center, which should be 3D point. Get Bounds() returns two 2d points!

Re: Camera position in axonomtery

ivanfi — Tue, 26 Mar 2013 10:55:14 GMT

The bounding box is 2D because it is the bounding box for clipping after the projection. There is no camera position in axonometric projection, only a direction and a 2D bounding box for clipping. It is enough to clip the result after the projection, because the unclipped result is independent of the bounding box.

Re: Camera position in axonomtery

Tomasz Marek — Tue, 26 Mar 2013 16:45:40 GMT

Thank you. I understand now. The bounding box is positioned in relation to project origin. This will help me finding the "central" axis.

For my application I have to specify a position in 3D to avoid clipping so typical in parallel projection. I will calculate it based on Model::GetBounds(). It would be nice if SDK could give such a safe "camera" position for the parallel projection.