PDA

View Full Version : Should 2d artists use traditional perspective, or 3d tools to generate perspective?


Gord-MacDonald
05-19-2005, 07:59 PM
I have seen some incredible works with very complex use of perspective at cgtalk and elsewhere. While this is a remarkable acheivement, fact is the rules of perspective have been established since the 15th century.

That being the case, is perspective not just a means towards an end? - and if so does the method used to establish perspective in the artwork really matter?

Should 2d artists use traditional perspective, or 3d tools to generate perspective for them?

Gord

lotaH
05-19-2005, 08:05 PM
Great question!!... :rolleyes:

And...

How this is work?

[]'s

paperclip
05-19-2005, 08:06 PM
Do what works for you I guess... you should really know the rules of traditional perspective though, to draw on and to have in your mind while drawing. You won't always have maya with you after all. (This isn't just for you Gord, it's also for all the people who visit this thread.) There's an excellent link in the art tutorials sticky about perspective- the 'handprint' site.

Gord-MacDonald
05-19-2005, 08:42 PM
Do what works for you I guess... you should really know the rules of traditional perspective though, to draw on and to have in your mind while drawing. You won't always have maya with you after all. (This isn't just for you Gord, it's also for all the people who visit this thread.) There's an excellent link in the art tutorials sticky about perspective- the 'handprint' site.

handprint is a great site - I downloaded all of their material on perspective some time ago!

Personally, I think there is definitely an advantage to having built all three types of perspective with and without the use of elevations using traditional methods. I think it gives the student of perspective a deeper comprehension of just what is going on. After that, I think that workflow and productivity becomes the major concern.


Gord

eparts
05-19-2005, 09:11 PM
it sure has been tempting to use 3d blocks for perspective.. sounds excellent for landscapes and city paintings.. however I have been too lazy to try it. I think its quicker to just put out some perspective guidelines.. And its easier to think perspective when you see the lines in front of you. and have perspective lines in mind while you draw.. and not make a scene at first in 3d.

I dont know.. some people argue about that digital art isnt REAL art.. so whats the problem really.

jmBoekestein
05-19-2005, 11:19 PM
What I basically miss in 2d art as to perspective is the fact that all perspective turns out linear instead of bolic/parabolic, if you take a panorama picture for instance you'll see every building swirve around some parabolic curves. Nothing is in a straight line when it comes to perspective/parallax.

That alone should be reason enough to revise the methods used. There are simple ways to do this even on paper, I'd have to look it up but I have it in a book somewhere. But as for the question, I'd guess go for the middle way, find the balance in all the trade offs for the puepose. :shrug:, If it's a landscape, or wideangle shot, you'll certainly have to at least think about it a while, but mostly I wouldn't even bother using them. But I'm inexperienced :D. lol.

Ordibble-Plop
05-20-2005, 12:27 AM
Try this test for yourselves.

In your 3D app:

-Set up your perspective view so that you are looking directly down the z-axis
-Create a cylinder at 0,0
-Duplicate that cylinder and drag it right or left along the x-axis
-Render the scene with a camera looking directly down the z-axis
-Open the render in Photoshop or similar and drag one cylinder onto the other

If your app is anything like XSI (which I tried this in) the cylinder at 0,0 will appear thinner than the cylinder moved off to the side, which, given that both cylinders are roughly the same distance from the viewing point, is patently wrong. (If anything, the cylinder to the side is further away than the one at 0,0 and should, therefore, be slightly thinner.)

The funny thing is, that if you do this on paper using standard methods you will get the same 'distortion', although there are ways to 'correct' for it.

With that in mind, I personally think that an understanding of how perspective works is a must, if just so that one doesn't take it for granted.

And once you know how it works you can play with it. I can't link to individual paintings directly but check out the pictures on the right-hand side (http://users.senet.com.au/%7Erfrancis/) of the boy on the skateboard and on the hill (and some others too). Also the technical page where he briefly describes how he uses software to generate the perspective in his paintings.

What I basically miss in 2d art as to perspective is the fact that all perspective turns out linear instead of bolic/parabolic, if you take a panorama picture for instance you'll see every building swirve around some parabolic curves. Nothing is in a straight line when it comes to perspective/parallax.

I guess there is this idea that traditional perspective is the proper way to represent the way we see the world in depth but really that is all it is, a representation. Looking through a camera lens is another representation. It amazes me that we accept both (unless there is radical distortion like with a fish-eye lens) as reality so readily when neither is really like the way we look at the world; in rapid saccades (glances) focusing on individual objects with the rest relegated to peripheral vision and partially ignored. Maybe it is because we look at pictures the same way we look at the world?

Thinking about these things does my nut in a bit, but I think that is where technique and artistry combine to create an image using these tools that engages and captivates.

Gord-MacDonald
05-20-2005, 12:44 AM
(http://www.cgtalk.com/member.php?u=56159)Ordibble P. Lop - you have touched on some really good points - it is amazing that we so easily forget that neither the image from the camera lense or perspective are binocular. The issues of peripheral vision and saccades are very facinating.

Gord

CIM
05-20-2005, 12:51 AM
I don't think you should if you don't know and understand perspective well. If you do, then go ahead, I guess. :shrug:

Lunatique
05-20-2005, 05:11 AM
If you don't know how to generate perspective with traditional methods, you should learn it, because it's one of the most fundamental tools an artist uses. But once you learn it and know how to use it fluently, it's fine to use 3D to save time.

Qslugs
05-20-2005, 04:34 PM
I would think using computers to generate perspective would save you from creating a bunch of extra lines on your paper :) Actually when my wife was in school for medical illustration, I would mock up stuff all the time for her to use for shading and perspective. If what you are doing is for money, and after all time is money, using the computer for all sorts of outside the box things is worthwhie.

I would argue, if youre using a computer for perspective and layout, you at least understand the principals enough to know the importence in using them. If youre a really foreward thinking outside the box kind of guy (or girl), you can even print that stuff out for reference for your manual drawings.

This brings to mind scientists and astronomers using photoshop for filtering purposes for viewing scientific research. My brother in law that goes to school for very small things (mems) had a long discussion with me. He wanted to know how to filter out all but the blue channel of something. That was so that he could view a specific component of his image. I stil laugh about that. An artist telling somone going for a doctorate about how photoshop. I assure you he can program a vcr or work a toaster.

Also, I went to school for animation on the tail end of film, right when the internet exploded and CG suddenly became attainably cheap. I remember doing a audio project that we were supposed to do on 16mm mag track. I had a mac at the time with Sound Edit 16, I went and did my project in SE16, outputting that to normal cassette tape, then transfering it over to the mag track. I think I had 24 or 32 tracks all playing in real time on the mac. Way beyond the capiablities of the mixing board at the school without doing multiple passes and mixing down manually. I actually got in trouble for that and was forced to do it over manually with the mixing board. Also, I failed to mention I got it done first way ahead of everyone else. In the end I handed it in last, getting reprimanded for handing it in late. At which point I mentioned to my instructor that he should remember that this is my second attempt, I already did this once :)

Another story... I also had to learn the optical printer. I made mattes in photoshop and printed out a positive and a negative. Everyone else did theirs by hand.

Also... I used photoshop once to lay out floor tiles in an interresting pattern :)

In the end I think its all about the tools and the style you are going for. Another thing to think about, if you need camera distortion for your drawing or painting. You can dial in any lens focal length in your 3d program. That in fact would take a bit of time to work out manually.

Gord-MacDonald
05-20-2005, 07:34 PM
Cim, Lunatique - good advice!

Qslugs - interesting stories! forced to redo a project? - darn those luddites!

rebuilder
05-20-2005, 11:06 PM
One simple trick for linear perspective is to create a pattern of lines originating in a single point, like an asterisk but with more lines. Just place the center of this pattern at your vanishing point or points and scale it to fill your canvas. This gives a rough guideline which is good enough for most drafting of scenes.

If you're interested in curvilinear perspective, there's some good info in this thread over at the sijun forums:
http://forums.sijun.com/viewtopic.php?t=38954

The handprint site dismisses curvilinear perspective as a curiosity, but it has applications especially in CG. Most interesting is the nodal point pan method described in the thread I linked to, combined with curvilinear paintings.

lotaH
05-20-2005, 11:32 PM
HI!

I think the idea that: " the more we know, best " is here in all the opinions. I agree.

I think to know the lineal perspective (traditional) it can help to organize the image in my head, before placing it in the paper or in the monitor. Even if you don't use the technique to draw.

One of the best things that learned (here in CGTalk) is that to study art it is so (or more) important that to learn to use a software. The soft without the art doesn't say anything.

Summarizing: I think is important to know on the whole perspective type.

The largest subject here in the art topics, it is the abandonment (or not) of traditional methods in the Computer. I think should not forget of them.

[] ' s
Good discussion. :thumbsup:

NOOB!
05-21-2005, 11:29 AM
heh,this kinda syncs with the *are we getting too digital thread*

i answered yes in that thread.now that i read this thread i say yes even more.

Kanga
05-22-2005, 04:09 AM
It doesnt really matter. You could spend lots of time constructing the perspective traditionaly or lots of time making 3d models. One thing tho is that in 3d you can duplicate shapes easily for say sky scrapes etc. The correctness of the perspective is flawed in the traditional and 3d approach as straight lines are curved so mechanical perspective is only an aproximation.

Another advantage to 3d tools is you can rotate your stuff around and choose the best viewpoint,... where as if you use a grid or rulers you have to draw it from scratch if there is a change.

If your composition is bad niether approach will work.

NoSeRider
05-24-2005, 09:32 AM
I just find that building a 3D scene in a 3D Software is more time consuming then just knowing the rules of perspective and drawing it.

Drawing is suppose to be less time consuming then 3D.....and drawing I mean sketching. Not rendering a masterpiece.

But even making an illustration should take less time then making a 3d scene....when you taking into account UV Mapping, Rigging, Lighting, Shaders. Generally, you should be able to make a reasonable drawing in about a day, maybe two, but my experiiences are 3D takes about a week for a good character model: Uv Mapping, Rigging, Lighting, Shaders........

default-rol
05-24-2005, 10:20 AM
My take on this issue is that perspective has to fit the purpose and means you are using it for. By this I mean that learning traditional methods of perspective on paper should be developed all the time so you actually know what it is you are looking for when issues like "distortion" occur.

Early in my 3D learning I was shown an image on a monitor of a cube primative that looked as though it was distorting on its corner furthest from the viewing plane. My knowledge of 2D perspective was not brilliant at that time, (I know a little more now) but it just "looked" wrong in the default camera set-up. That is "mathematical" accuracy for you.

I am always trying to re-address my perspective knowledge on 2D surfaces to further understand the complex issues involved, and I feel sure that any insight I gain from this process will help me in the 3D world. Afterall, anything done in 3D can only ever be rendered in 2D - with the "3rd" dimension coming from time in the case of animation.

MDurante
05-24-2005, 05:53 PM
I used Maya to help in my latest acrylic painting, which was an entirely imagined scene (landscape) so I didn't have a lot of reference.

For composition it was very helpful. I used stand-ins for elements and could easily move them around. Also, by grouping linear curves together I was able to build golden rectangles, and adjusting them (facing the camera) allowed me judge and change my composition (as an exercise in "dynamic symmetry"); this was alot easier than using tracing paper.

Maya was also helpful for judging shadow-shapes. I had a curved doorhandle that cast a shadow on the door, and I modeled and rendered it because I wasn't sure what that shadow would really look like.

Here's a question: does anyone know what camera settings best approximate the human eye? I just played with the fov until this looked "okay", but I wonder if there's a best setting.

Thanks,

- Matthew Durante

Gord-MacDonald
05-24-2005, 06:54 PM
Here's a question: does anyone know what camera settings best approximate the human eye? I just played with the fov until this looked "okay", but I wonder if there's a best setting.

Thanks,

- Matthew Durante

I can't give a 100% honest no-bs answer, but my understanding is that portait photographers use 80mm lens as a rule of thumb, to produce the least amount of distortion.
(there are people on this forum who are alot more qualified than me on this one)

Gord

MDurante
05-25-2005, 05:23 AM
I can't give a 100% honest no-bs answer, but my understanding is that portait photographers use 80mm lens as a rule of thumb, to produce the least amount of distortion.
(there are people on this forum who are alot more qualified than me on this one)

Gord

Thanks Gord. I remember hearing that one of the reasons wide-screen cinema is superior to the traditional 35mm square-shape is that widescreen 70mm better approximates how we see. So 80mm sounds like it would be about right.

- Matthew Durante

rebuilder
06-09-2005, 06:22 PM
MDurante:
There is no single lens angle that will best approximate human vision. Firstly, the human field of vision is over 180 degrees for most people with both eyes functioning. (also the field is a bit oddly shaped thanks to the nose , eyebrows etc.) What matters, if you intend to have the perspective of the image match up with the viewers perspective is the size of your image and how far the viewer is from it. If your image has a 60 degree horizontal FOV and the viewer is situated so the image fills a 60-degree horizontal slice of their FOV, then there should be little distortion. (If the image is on a planar surface, that produces some distortion, which will become more apparent the more of the viewer's FOV the image it fills. Mostly this isn't significant though.)

Anyway, the point is, lenses do not distort. Not in the sense of making noses look too big or whatever. It's just that with a wide lens you have to get pretty close to make a head fill the frame, and the perspective gets "exaggerated" since we're usually not looking at people 30-40 cm from their faces. I'm guessing portrait photographers use 80mm lenses because it lets them take photos at a distance at which we are usually comfortable with looking at people.

MDurante
06-10-2005, 04:12 PM
Thanks Rebuilder, that's a good explanation and makes sense. I have to think about it more to understand all the consequences but this is the start I needed. And it's probably not something I have to be overly obsessed with when my goal is just to compose landscapes that look "natural".

- Matthew Durante

kraal
06-10-2005, 04:45 PM
scott robertson addresses this in his Q&A.... I agree with him. To copy perspective from a photo or 3d mock up will make a good pic but you will cheat yourself by not learning the rules. learn what you need to create from your mind out will benifet you more than just learning to draw what you see....

.:ZRDwD:.
06-10-2005, 05:30 PM
Excuse me for probably being the only one being "outside the box" and such a downer, but.... why all this science mumble-jumble and "rules"? I'm not dissing anyone, but when can art be ...... art? Must someone know the formula of lightwaves to accurately layout the placement of linear mackle-frackle to bring out the color in Charlie Brown's eyes?? I mean, come on.

In my 12 years of schooling, I was the only model for all my drawings. I think of something, look in the mirror and/or whatever body part, and place myself as the different characters in each piece. Now, I have 3D means of doing this (and still look in the mirror). I can't afford models, so I do the cheapest possible way. Does it take away the "rules"... I'm sure it does. But, who cares? If I want to do a photorealitic picture... then, uhhhhh... shouldn't I just take a photo if I want this light exactly like it is in "real life"? Who cares about these "rules". "Rules" are set for this "real world"... not the imagination. And I see art as that: the imagination. All rules are forfeit. If anyone critiques that the proportions of Cthulhu are unrealistically aligned and the lighting isn't shadowing where it needs to be under his tentacles, I can only close my eyes and shake my head. Where have these people seen an actual beast named Cthulhu and in which realm? Nowhere, because it came out of the individual's head.

Should 2D artists use traditional perspective or by 3D means?
I say use whatever. A naked women photo or Maya render.... drawn or painted with a flowing gown on top of a fiery dragon.... the winner will come out being what the imagination had portrayed, not the means or "science" of fundemental rules of realism.

Sorry for that burst of analism. Great topic, Gord, and I'm glad I got that off my chest. :scream:

CGTalk Moderation
06-10-2005, 05:30 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.