PDA

View Full Version : Camera position



Integrity
08-21-2005, 08:48 PM
When setting up the camera, is the location of the camera the exact position of what would be the image plane in a real camera? Or is it the lens, virtual image, intersection of the rays, etc?

Weepul
08-21-2005, 11:00 PM
I believe it's the "nodal point" - the point where all the rays seem to converge. If it was a pinhole camera, it'd be the location of the pinhole.

Surrealist.
08-22-2005, 12:00 AM
I would say it is the location as this is the only peramiter that is actually mesuarable. Good question. Never thought about it. But I would say, if you have a camera that in the real world is 5 feet from the subject to film plane with a 100mm lens then set it up this way in Lightwave and that will match. There are no other peramiters that I can think of. With verion 9, I think things like the distance to the image plane and angle will be able to be animated, but untill then, my guess is, go by the camera position.

kjl
08-22-2005, 09:08 PM
I believe it's the "nodal point" - the point where all the rays seem to converge. If it was a pinhole camera, it'd be the location of the pinhole.

Yes, that is correct. You can test it by setting your camera to have a square aspect ratio and an FOV of 90 degrees (zoomfactor = 1) and putting a 1 meter square plane exactly 50 centimeters from the camera. The square will exactly line up with the image border.

Surrealist.
08-22-2005, 11:35 PM
The Real world example is camera reports come in with measurements of the camera hight and distance from subject, or some other reference point along with the lens settings and so on. You want to match the shot for compositing. OK where is the film plane? Right? In otherwords, when measruing the distance from the real world camera to the lens, you take a tape mesurer and stretch it from the subject to a spot on the camera that tells you where the film plane is.

Here's is a quote from a product that automates this real-world process:


Measuring distance just became easier with the new Cinematography Electronics CINE TAPE MEASURE System. This versatile ultrasonic system continuously calculates the distance between a subject and the camera's film-plane. By continuously measuring distance during a shot, small or subtle subject movements are no longer a problem. With non-stop measurements, focusing a camera lens is simplified which saves time, avoids retakes and reduces production costs.
So the question is, where is the film plane?

It's the camera position. I think that is meant here by the nodal point. Where all of the lines converge.

There is a custom object in LW called Camera mask. You apply it to a null, parent the null to the camera, open the dialogue for the costum object and give it a size. It will then give you a position readout in layout- relative to the camera lens - where this mask will line up.

EDIT: The main use for this tool I should mention is to have a way to line up a polygon that is the same aspect ratio as your camera for purposes of mapping images on a polygon when that image should fill the screen as in compositing.

Apply the above formula and you get the same result. I meter square mask(polygon), lense factor of 1 and it tells you to put it at .5m or 50 cm/500mm.

You can line up any size mask(polygon) to any camera lens this way. As to what this has to do with matching the film plane to the real world camera, I have no idea. It would be interesting, if there is some significance to the above example. I'd be interested. :) Is this based on some kind of real world physics to do with optics? I don't know.

However, the bottom line is. Where is the image being recorded at?

Wherever the camera position is of course. Why would it be anything else?

In the real world that "position" is a mark on the camera where the film plane is. In LW, it's just the camera position.

I believe the programers had in mind real world aplications and there was no need to make it anymore complicated than that.