Horizons. Does anyone have any idea why.......?

[Saulman]

This is not a VRAY question, but a 3D studio question.

Please see the attached image........



I have had a number of camera views taken with a precise GPS unit and coordinated the camera and target points in 3D studio.

The camera was mounted on a tripod and leveled with a spirit level.

When I level the camera with the target in 3D studio the horizon line cuts the viewport (and the image) in half so that the top and bottom are equal. When the real camera was used the camera views have the horizon nearer the bottom of the image.

On each image I think there is an approx 60m difference between the virtual horizon and the real horizon.

Does anyone have any idea why? Or does anyone know where I can find out why?

It may be something thats very obvious but Ive hit a brick wall. I know this is a bit of a wierd question, but any help would once again be greatly appreciated.

Thanks.

[Saulman]

Having had a long discussion with the surveyor, we've established that he may have cut more from the bottomn of the image than the top.

Problem solved.

Thanks.

[percydaman]

for future consideration, please dont post non vray questions in this section of the forum.

[Saulman]

Apologies, Percy.

[percydaman]

no problemo

[Tobbs]

Sorry for bumping this post, but I had to give an explanation to your horizon-problem.

The difference lies in the curvature of the earth. The horizon in Max is based on a flat infinite plane which the earth isn´t.

As you see in the attached picture the difference is quite big even when the camera is placed exactly at the earths surface.

The amount of difference also depends on your FOV.
If you figure out the altitude of your real camera you could probably match the horizon perfectly if you just make a sphere representing the earth...

/T

http://www.rithuset.se/content/tobias/earthhorizon.gif

[trixian]

Actually, Tobias, your earth is too small. If you remove all topological changes on our planet and placed a camera (with a lens simulating human vision in a 16:9 framing) at about 180 cm over the ground level, you should not be able to detect any curvature by looking at the image. You could measure it if the image was high enough res, but not manually with your eye. In theory, on a smooth planet with a camera completely levelled, the horizon should generally be just slightly under the 50% hight of the image framing with its mid-point just hitting the 50% hight. Not as extreme as you have in your gif example. The smaller the planet the more it would look like your example.

[Da_elf]

hey tobbs did you happen to read this one? seemed a no brainer to me and had absolutely nothing to do with the curviture of the earth

Having had a long discussion with the surveyor, we've established that he may have cut more from the bottomn of the image than the top.

Problem solved.

Thanks.