How To Use Ellipse Templates In Perspective Drawings?,

#1

Hi all I have a couple of questions regarding the use of Ellipse templates in Perspective Drawing

  1. Is it true that ellipse templates are primarily Axonometric-ally measure circles? Since it doesn’t take into account foreshortening, we won’t be able to use it so freely?

  2. Is how we use templates by aligning the minor axis to the shorter marks on teh template ?

  3. Is it true that the minor axis is always aligned with the main/thrust axis , even in perspective? (ie the wheel axis)

  4. ATTACHMENT 1 :Attached below is an example,
    e.g. Hmm, say i’m drawing the front left wheel of a car on the ground and the orange line is the shaft(to 40 deg VP) of the front wheels. And since i know the wheel will be perpendicular to the shaft, what degrees should i use this time? (assuming wheel not tilted and is perpendicular to shaft)

  5. ATTACHMENT 2 : Say, when a square/rectangle recedes into distance it foreshortens. so i’m wondering does the Ellipses
    (along the same receding plane) actually Change its TILT when it recedes or does it remain the same but just gets scaled down. I understand in reality its tilt doesn’t change because its still the same VP, but im referring to whether it changes in Perspective drawing .i.e we must use an ellipse template of a different degree e.g. the front and back wheels of a v long vehicle

  6. Anyone came across any digital program/method to calculate the minor:major axis ratio, acting like a digital ellipse guide?

  7. say i use a 40degrees template, how is the 40 degrees ellipse actually constructed? is the degrees calculate based on the minor:major axis ratio
    Any other instructions/opinions is welcomed too

Thanks

#2

I’m not quite sure what you are doing but I’ll try to answer a couple of things anyway.

  1. The shape of an ellipse is not changed by foreshortening in perspective. The center point will change accordingly.

  2. Don’t know what you are asking here.

  3. Yes. If the ellipse represented the end of a cylinder of infinite length, it would reach the vanishing point on the horizon line.

4)You can not determine the ellipse angle because that angle line (besides the known 40°)
is not known. Under the circumstances, any angle would be correct.

5)View each ellipse as a cylinder end. all the cylinder lengths will follow lines that join at a common vanishing point on the horizon. Therefore they will ALL have different degrees and tilt angles.

6)Any 3d app can produce what you need. Some may give you the sine value of your ellipse.

7)The ratio of an ellipse is represented as it’s trigonometric sine. The sine of a 40° ellipse = 0.64278
. This means the minor axis is 64.278% of the major axis. Here is one of many links on how to construct an ellipse manually: http://www.uwgb.edu/DutchS/MATHALGO/Ellipses.HTM

—You should look into 3d software.

#3

Correction for (4).

any ange but 360° would work. Ideally.

#4

thanks for your reply :slight_smile:

do you mind showing me a picture of how the minor:major is constructed? I know sin 40 = opp. / hyp , but i can’t visualize how to form a triangle with minor axis as opp. and major axis & hyp. how do u form a triangle inside to achieve that?

#5

You don’t draw any triangles. If you look straight at a circle and tilt it toward you at 40° so the top part is closer to you and the bottom is farther from you, it visually creates an ellipse. At 40° the horizontal dimension (major axis) will not change in length but the vertical line (minor axis) will be 62.478% shorter (sine 0.62
) in appearance than the major horizontal line. That’s what I was describing.

Here’s another link that may help in your examination/project:

http://www.khulsey.com/perspective_ellipse.html

Good luck in your quest.

#6

and yet another correction.
The minor axis will be 64% of the length of the major axis, not 64% smaller.

I’m going to see ellipses in my sleep now. I think my explanations are getting elliptical.

#7

haha, thanks for being so elliptical to help me out here. i read the site you gave me, some questions:

  1. So is it impossible to just use ellipses from the template & just draw circles in perspective without modifying? ( the website guy did modify a little)

  2. do you mean the ellipse are aligned to the wheel axle in 2d plane, rather than wheel axle in 3d world? Thus the different axles though in 3d world are parallel but in 2d Perspective is not due to convergence to VP, so i can’t use that as a basis for measuring ellipses?

  3. So is it true that using Ellipse templates can only be used for Axonometric drawing because their wheel axle doesn’t converge? Since in perspective the same receding ellipse would change in tilt?

  4. This was my initial concept, pls tell me why it is wrong. SEE ATTACHMENT

  5. My question is like say there is a row of ellipse with the same degreet tilt lying horizontally in front of me, would the minor:major axis ratio of the ellipse of both ends be different despite them being equal in degrees to viewer in reality? ( equivalent to a row of cylinders tilted acorss a perspective camera in max & the ellipse we’re considering here is the end nearer to the camera; i am trying to learn max so will show picture if i manage to create what i mean)

#8

You are a bottomless well of ellipse questions. Curiosity is good, but unfortunately I’m at the end of my ellipse rope. You should do more web searches to satiate your intense hunger for ellipse knowlege. You will have to take the reigns now, *Grasshopper.

Also, I think you will benefit greatly by using 3d software.

*Grasshopper, from the tv series, Kung Fu

Good luck.

#9

oh no =( , i don’t need luck… i need you!!! haha… since you seem to be the most helpful to my questions around here.
thanks for helping man :slight_smile:

one more attempt to tap your knowledge:

  • If i simply create a perfect circle in a perfect square then use ‘free transform tool’ in photoshop to Bring the 2 backends of the square together while keeping them aligned to lines extended frm VP, is that an accurate ellipse?
    'compressing the two ends of the bounding box to create illusion of
    Perspective. I was wondering in this case( attached picture, not drawn to
    scale(esp edge length receding towards vp) :

(i read that this digital method of ‘squeezing’ isn’t very accurte?) See Attachment

#10

The answer is no, the ellipse won’t be accurate by using that method.
Your homework is to figure out why.
Hint: You have all the information at your disposal.

#11

Hi,

i did my homework ( hopefully Correctly). the answer is: because the receding edges of the square was just brought together, the circle appeared elongated. the receding edges of the square should have got shorter due to foreshortening but there is no way that could happen by just bringing the 2 end points together.

is that correct? please correct me if im wrong and show me the correct ans pls?

#12

Yes, you are correct.
What I’d like to know is, why are you so fascinated with ellipses and perspective? Not to say there is anything wrong with it. I was obsessed with everything perspective when I was in high school, this was before cg and 3d apps were mainstream.

#13

haha, i’m still very much into 2d illustration & perspective is like a big mystery, it feels so good to understand something new about it each time. i’m starting to pickup Zbrush too which i see you mentioned in your signature.

  1. btw, is is faster to model a character in 360 in Zbrush than in like 3dxmax? also, is there a feature that allows wireframe view of our models in zbrush?

  2. back to our discussion about ellipses, i hope to verify this with you:

    so am i right to say if i were to create a perfect circle in a square, group them in photoshop, then later manage to adjust the 2 handles of the square; aligning them to the correcting vanishing line & also adjust them to the correct length.
    i will automatically get a circle in correct perspective? ie as accurate as plotting many points in a perspective square then joining them to form the perspective circle/ellipse?

  3. If the above is valid, doesn’t that also mean i can draw a profile view of any object, and present it in correct perspective by enclosing it in a square/rectangle and then using the same above method to adjust it, as long as the receding lengths of the square/rectangle is of correct length in perspective & along the correct vanishing line leading to its VP?

Thanks dude :applause:

#14

ZB: I use ZBrush for creating detailed displacement maps for organic models I start in C4D or Modo. ZB is not a general-purpose modeling app.

Ellipse accuracy: I guess if you use your measuring method you seemed to have ironed out in your follow up thread, you should be able to produce an accurate ellipse. As far as drawing everything else based on a measured cube in perspective, I would say it would help as a grid reference.

#15

I see. thanks for the tip.

what about the questions :

  1. back to our discussion about ellipses, i hope to verify this with you:

so am i right to say if i were to create a perfect circle in a square, group them in photoshop, then later manage to adjust the 2 handles of the square; aligning them to the correcting vanishing line & also adjust them to the correct length.
i will automatically get a circle in correct perspective? ie as accurate as plotting many points in a perspective square then joining them to form the perspective circle/ellipse?

  1. If the above is valid, doesn’t that also mean i can draw a profile view of any object, and present it in correct perspective by enclosing it in a square/rectangle and then using the same above method to adjust it, as long as the receding lengths of the square/rectangle is of correct length in perspective & along the correct vanishing line leading to its VP?egarding the use

if i can actually get accurate perspective that way, i could save lotsa time as i like to start from profile views. do you think it can be done, for circles & other curves & such?( ie by doing that i just need to use the measuring point method to get the correct length of the plane and the details on the plane will automatically be in correct perspective too. the planes will be profile views of the object)

i think it will work, but whats your say on it?

#16

if you want to understand perspective drawing top to bottom, get Scott Robertson’s DVD

#17

Sethellic

I’d say yes to part 2 and 3 if you keep your main elements close to the horizon line.
Things get distorted and confusing when you extend far above or below the HL as you can see in the linked sample:

http://billmelvinart.com/pictures/box/persptest.jpg

The square below the HL is distorted because the viewer’s line-of-sight is not on it. It is along the HL. The square below the HL needs 3-point perspective, which would not work in the same image with the square above it.

The reality is perspective drawing is an idealized approximation of what is really going on within an image. This problem occurs in 3d apps as well that do not have a true lens distortion camera feature based on focal length.

The bottom line is you have to cheat, here and there, to keep things believable.

I haven’t seen the DVD but heeding Mile Dream’s suggestion may be a good idea:
“if you want to understand perspective drawing top to bottom, get Scott Robertson’s DVD”

#18

thanks for the reply & effort to even illustrate it with a picture :slight_smile:

i would like to check to see if i understood what you said above:

  1. isn’t the square below the HL distorted because its very far from the viewer’s line on sight( ie far out on the circle of view) but it can still be in 2 pt perspective isn’t it?

  2. So i’m curious why you said, “The square below the HL needs 3-point perspective, which would not work in the same image with the square above it.” and by your words ‘would not work’, do you mean it will look distorted OR? can using 3pt for this square solve the distortion?

  3. if i do use 3pt perspective, all i have to do is to add another VP below and thus make the verticals converge as well, if the lengths of the square are well measured to scale with the measuring point, will the enclose image/circle automatically adhere to the laws of Linear perspective as well?

  4. I went to your website & checked out your portfolio, your works are REALLY MIND-BLOWING!!! the perspective & rendering are really tight & beautiful esp. the oils. For the digital works in Anatomy sections did you render them in 3d apps or digitally painted them in 2d eg the picture with see through muscles? I’m so impressed :wip:

Thank you so much for your replies thus far. Really appreciate your knowledge :applause:

#19
  1. Yes.
    2 & 3) You have to use either 2- or 3-point perspective in an image. You can’t decide to use 2-pt on some elements and 3-pt on others.
  2. Thanks. The see-through man images were rendered and Photoshoped up from a 3d model I purchased. I use 3d apps when ever possible in my illustration work. 3d apps take care of all my perspective work.
#20

Hi QuadArt,

thanks for the answers, i went to think about them(did my HW) and formed some conclusions, could you let me know if its the same train of thought when you gave the answers :

  1. if i do use 3pt perspective, all i have to do is to add another VP below and thus make the verticals converge as well, if the lengths of the square are well measured to scale with the measuring point, will the enclose image/circle automatically adhere to the laws of Linear perspective as well? <-- My answer is: YES, it will adhere to Linear perspective laws (+ its distortions) and thus result in correct distorted forms enclose in the Square as well. Am i Right?

  2. The reason you mentioned why the object below viewer’s line-of sight would require 3pt is because to reduce distortion and to stimulate real-life seeing more, most probably we would be directing our line-of-sight at the object, and the object lies below eye level in your example thus it causes the direction of view to not be parallel with the ground and thus you said = “it would require 3pt perspective”. Is that the reason why you said it would require 3pt?

  3. i thought of learning 3d modelling for illustration as well. is 3dMAX is good tool for general modelling? the camera, lens and lighting features seem quite comprehensive inside. or do you think Maya would be better? I’m most impressed by how you handle the Smooth value gradations in the Oils, were they painted from life,photo reference or your imagination?

Thank you :applause: