New viewpoint about perspective and drawing

#1

This a short and interesting note about thinking and drawing, this is the link:
http://www.viewpoint.cl/perception.html

#2

This is an interesting POV. I like the thinking and the more-less analitic approach, but yet very understandable.
Thanks for the link!

#3

I will upload the second part soon.
Thank you.

Ruben.

#4

This very much reminds me of Escher’s self portrait in the metallic sphere (the way the perspective lines are bent). Its an interesting idea, but seems more along the lines of a method of exaggeration than “thinking in 4d”. THanks for the article :slight_smile:

#5

Since you can notice this phenomenon, also you can control it or play with the distortion. If you increase the effect, in this way you can exaggerate correctly.

The sphere of Escher or full sphere images works like a fish eye lens. You can get these images combining images from two rotating fish eyes lens. But these lens don’t create the visual effect, they exagerate it. Because it is like If our field of view was greater. It was described there.

Also this article tries to show how to get more freedom when thinking about perspective and we want to draw it on the paper.

Thank you by read the article.

#6

This is hardly a new viewpoint and certainly not 4D, it’s called curvilinear perspective and it been around since the renaissance. It places the viewer on the inside of a sphere and uses 6 vanishing points, one at either end of each special dimension (3 x 2 = 6) of course the sixth point is behind the viewer so it’s purely theoretical. Oddly however it turns out that it is exactly the same as one-point perspective just mapped onto a hemisphere. So if you have a ruler and computer you’re laughing. Curvilinear perspective despite being a more accurate representation of how we actually see never really caught on very well for two reasons, one is that without a computer it’s a nightmare to lay out, and two that it raises an interesting philosophical problem, being that if we see things curved and then we draw them curved and then we see what we have drawn, we find it has been curved twice. So it’s only really seen as useful for very wide-angle pictures, or like storyForge said for exaggerating. It does prove however that there is only one true method of perspective drawing and that is either one-point or five-point (6-1) depending on how you feel about it.

#7

The six point of perspective is a method to circumscribe the world in a sphere, it works like the Escher’s sphere.

My method is an observation of real world. The real world is curved, this effect is perceived a bit by our limited field of view. Our brain corrects this phenomenon. We don’t think about distortion.
My viewpoint works like a mix beetween 2 points (most used) and 4 points(distorsioned world) of perspective. But these methods use straight or curved lines. The real world doesn´t work in this way. The curved lines are soft arcs.

It is for open our minds to real world. It seems like a individual force that we (or anything that it exists) exert around us. And push the light. And we can`t push the far things. It depends of our distance to the objets. You can rotate a cam and see it. Also It can be a philosofical problem, but it is ahead us.

You can’t perceive a plane drawing curved twice, unless you put it 2 inches of your eyes. Our vision isn’t very distorsioned.

The main viewpoint is: if we draw our sketches with rulers, we are limiting our representation.

#8

I already covered this in my last post. Let me try to clarify. Technically speaking there is only ONE vanishing point, and it’s situated bang in the middle of the viewer’s field of vision. I know a little about the technical side of this because I have scratch built 3D engines. This is of course one-point perspective and it’s the system all 3D computer programs draw with.

  People however aren’t as apt at number crunching as computers and don’t cope very well with drawing objects edge on using only one vanishing point. So they add extra points to aid them. This is what, off centre, two, three, and multipoint perspective are. They are however still one-point perspective, just a more human friendly representation.

  To get back to the point, if you were to draw something to scale (say a matchbox) which is made up of straight edges, and you used straight lines to draw it, and then viewed the object and the drawing from the same distance (be it 2 inches or 200 feet) the distortion would of course be the same for both. Therefore there is no need to draw the curve. If you do draw the curve the matchbox will be curved double when you look at it, which is exaggeration.

  This not only applies to objects drawn to scale. But also to larger objects that are the same size as the drawing, when viewed from the distance depicted within the drawing, when the drawing is at the distance it is meant to be viewed from (say arms length)

  The only time adding the curve is necessary is when the drawing depicts an unrealistically large field of view, which is of course wide angle. So I say again that curvilinear is only useful for exaggeration and wide angle (and wide angle is a form of exaggeration anyway)

  Your viewpoint method is the same as curvilinear perspective (4 point is the same as 5 btw just like 2 and 3 are the same as 1) you will find soft arcs eventual turn into big circles. The further away you move the points the bigger the sphere gets and the shallower the curve seems.

  I’m not saying there is anything wrong with exaggeration in drawing, but curvilinear perspective or whatever you choose to call it. Is by definition exaggeration as I have just demonstrated.
#9

I respect your knowlegde in this theme.

I think that my viewpoint is about this methods and another things, It´s oriented to understand it and how we can use it like a resource in our graphic.

My method works mainly when we drawing a sketch. It shouldn’t be limited with a ruler. Mainly If we are drawing a human figure.
Also when our brain sees some exaggerated form, It interprets this exaggeration like something very near of him and with punch effect. Nowadays this effect is very used in impact images, you can see exaggerated (more or less) images in advertising or cinema.

I wrote that you can use a major exaggeration for cartoons. Also this distortion has logic.
We can`t see the big circles or arcs, because our field of view is small. Therefore the distortion should be smooth in more realistic images. I am speaking of a smooth distortion that allows freedom us to draw with out straight lines.

For example I can undestand that an object located far of us has a tele effect, and his impact is minor in us. I explained it there.

It let us understand how it works around us. This method help us when we are representing an object. Also we can understand what camara we should select in a 3D software. Or if you want to apply some method like 2 , 3 or 5 points of view.

I have seen mega 3D structures with wrong camara selected. It looks like Godzilla between the buildings in old movies.

I’m familiar with some 3D software, sometimes I have used it with human figures.
Also It’s difficult to get nice perspectives when we are approaching the camara to the figure.
I know what is a movement of camera, movement of the complete universe, scale, etc.
I want to ask you with many respect:
¿Do you think that the curvilinear perspective from 3D software is a perfect emulation of the real world?
If you think that It’s not perfect then I can change the name. :slight_smile:

Thank you very much by your time and sorry if you have suffered with my english.

#10

You may have seen funny cartoon drawings of famous people with their features exaggerated? With big ears or silly big teeth, and the cartoons are more recognisable as the person than some photographs. In this respect I totally agree with you that the exaggerations can make a space or an object feel more realistic, than a more technically accurate representation.

   I suppose also that we are familiar with seeing these distortions frequently in photographs and on television. So why not in drawings?

   I don’t know any 3D programs that use curvilinear perspective. All 3D programs seem to use one-point straight-line perspective. 

   In response to your question: I think the one-point straight-line rendering method is a good emulation of real life perspective only when rendering very small objects or narrow fields of vision.

   For larger objects such as rooms and buildings or dramatic close up perspectives, I think it is not perfect at all and looks wrong.

  I came across a piece of work while browsing the galleries that I think partly illustrates this:

http://www.elifserencioglu.com/cc/lip_liftin_800.jpg

[img]http://http://www.elifserencioglu.com/cc/lip_liftin_800.jpg[/img]     With absolutely no disrespect to the artist, this image reminds me more of a dolls house than a large room. This is not her fault it's the program and entirely due to the fact that there is no wide-angle distortion, so my brain interprets that the space must be very small.

   (This is the proper thread btw: [http://forums.cgsociety.org/showthread.php?t=274197](http://showthread.php?t=274197) )

   Personally I think 3D programs could benefit a lot from the curved line/curvilinear perspective. And I have been considering how this might be done for some time, apart from obviously by distorting the image post render. But as of yet I don’t think it exists.

     You are of course always very welcome to my time, [b]thank you[/b] for starting such an interesting thread :)
#11

Most 3d programs render scenes like a camera does,

if you render a scene using a 35mm virtual camera, that virtual camera displays the final results much like a real camera does.

it makes no sense saying that 3d programs need new camera techniques, such as curvilinear perspective. telling Maya to render a scene using a 35mm camera will get the same physical results that a real world 35mm camera will produce, and my photos taken by my real 35mm camera look realistic without the help of curvilinear perspective.

#12

I’m afraid most 3D programs definitely do NOT in fact render like a camera. 3D programs use one-point perspective, which only measures distance along the Z-axis, so objects getting further away along x and y don’t get any smaller.

   I have created this simple web page with some images and explanations to prove this.

   [http://www.geocities.com/smackcakez/3d/3dflaw.htm](http://www.geocities.com/smackcakez/3d/3dflaw.htm)

I have never come across a 3D program or game that doesn’t use one-point perspective including 3Dmax, light wave etc. I’m afraid I haven’t seen the Maya 35mm render you’re referring to, and I can’t find any examples online, so it’s difficult for me to comment. Perhaps you could post some?

What I will say though is if you really are getting the same physical results in Maya as you would with a real camera, then this 35mm render will be the same thing as curvilinear. However I doubt this is the case since it would probably mean altering very fundamental aspects of 3D such as polygons, which as far as I know Maya still uses. UNLESS as I have mentioned before it is being distorted post (or perhaps during) render.

If 35mm is not the same as curvilinear, then I’m afraid your images are not the same as a camera, which will become more obvious if you widen your field of vision.

#13

good point, you learn something everyday, or week, or for me a month

lol

thankyou

bye

#14

I have received a link in my site about a spherical cam or distortion cam plug-ing. The company there says they are the only commercial renderer available that offers it.
I am trying to know more about it, but it looks like a post production effect, when the render is done. You can select several cams in it.

My article is about how the reality works, our vision is like a mix of cameras and It tryies to explain how it works. I hope upload the second chapter soon.

The site link:
http://www.finalrender.com/products/feature.php?UD=10-7888-35-788&PID=53&FID=489

#15

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.