[ceenda]

Hi there.

In some of the pictures I do, especially of scenes where the viewer is looking down a hallway, the perspective lines come from a vaninshing point _so_ high that it is near impossible to draw them on the paper accurately.

For instance, a scene where the viewer is looking down the nave of a cathedral but looking up at a slight angle. The lines of the vertical geometry converge at a viewpoint perhaps 1 meter above the paper!

Any ideas? I think I've seen spooge, Fred and lotor do pics of this kind and I wondered if anyone had any hints or tips?

I'd rather not have to do a basic wireframe version in 3DSMax as that is a bit cheaty and teaches nothing about perspective.

Nick.

[synj]

ceenda, do you have any examples you could post? I think I know what you're saying but feel silly trying to answer it since I am not sure.

btw, doing a wireframe in max first DOES teach you of perspective. not only do you see how the perspective should look, you can spin it and see why it's that way.



-synj

[This message has been edited by synj (edited July 29, 2000).]

[Matt]

If the v.p. is really that high, then just don't worry about it - limit yourself to two-point perspective and make the verticals vertical. As you become more experienced, you can allow yourself to "cheat" a bit and let your verticals converge for some three-point effect.

I would seriously suggest studying (a class) perspective rather than merely observing a wireframe - the latter will do little to teach you about the cone of vision, understanding shadows and reflections, or how to manipulate perspective to simulate the effects of different camera lenses, etc.

[Frost]

Hi.

I made this little image here... which seems to work out what you mean. The bigger fuzzy lines show the main lines I layed in at first to guide the perspective, and the other thinner lines are created from subdividing the space between between the main ones. By doing so, all the lines merge at one point off the image/canvas. The lines outside the two main lines were created by copying the difference between two smaller lines and shifted along the axis in which the sample was taken... uhm, yeah.

Of course, since they are linearly subdivided, they do not reflect foreshortening and depth perspective, BUT, the point is to have the lines converge at one point, which this does.

Basically, red is depth, blue is height, and green is width.

Hope it helps a little.



[This message has been edited by Frost (edited July 29, 2000).]

[Francis]

A friend of mine named Seth Fisher has an interesting method of dealing with this too:

http://www.lllama.com/i_resources_tip1.htm

------------------
Francis Tsai
TeamGT Studios

[ceenda]

Thanks to everyone who replied! Lots of good advice in there!

synj - yeah, you're right. by moving an object around I guess you 'learn' why objects look as they do in perspective.

BTW - good grief! what's going on there???

Matt - I'm currently making my way through the perspective section in a book I'm reading right now. All the examples so far had local vanishing points.

Frost - Thankyou, you've kinda hit the nail on head there. Fantastic stuff!

Francis - That link went out of the thread and straight in the 'Favourites' menu.

Thanks everyone!

[EDIT] - Am I over-using this smilie?



[This message has been edited by ceenda (edited July 29, 2000).]

[ceenda]

quote

Quote:

You don't look hard enough

You think I'm some kinda wuss?

... only joking, he he.

[spooge demon]

Yes, 3-d programs are really great tools for learning the concepts of perspective, esp lenses. Before, it was kinda difficult.

The problem is it's too easy. You don't look hard enough.