[Joachim]

Thanx alot eetu !

I really appreciated everything you said, and also it made a lot of sense to me.
I'm sure there's even more mistakes to it, but of what you told me, things got a lot clearer.
I will go over and do the things you told myself as well, just to print the principals into my head. Don't think I'll do more cube pics before the next lessons, kinda get bored of doing cubes after doing it for two days

Thx again !



------------------
Joachim
web: http://home.sol.no/~jbarrum/

[iska]

heres my cube try. used wacom. the perspective is quite weird indeed ;D

[eetu]

trance-r: use a lower threshold setting for the radiosity

joachim: glad to be of assistance.. heh, cool, i thought it'd take a long time for me until i could help here and not just be helped

damn i'm a bit ashamed of my cubes, perhaps i'll try again um um

and welcome, iska!

eetu.

[Mozeman]

Adding my pic to the mix. I need these lessons big time.





------------------
Mozeman
************************
[email protected]
************************
"I like it rough..."

[Shadow]

hehe i bet u all professional artists closer yer eyes and see a bunch of 3d shapes... just like i see chess in my head after playing a few games
hhehe
I CHALLENGE j00!! MUAHAHA
heh

-S

[Sedone]

This started making my head hurt, so I figured it was a good time to stop.




------------------
http://sedone.cjb.net

[craig]

"Fred",

First thank you for all your hard work,
this is awesome! I'd also like to thank
all the others that are providing input, instruction, and sharing of knowledge
that makes this forum so cool.

Francis,
Your replies on perspective are right on!
I used a book titled "Perspective", by William Powell that was very helpful, but
it is short. Your additions helped me out!
Thank you.
I have taken a crack at the shapes, and
I also went to the local park and sketched
a slide. As a free suggestion (and it is
worth what you're payin' for it) these
playground sets are ALL comprised of the
basic shapes that we are drawing. I was paricularly struck by the real life
examples of reflected light! Drawing from
the "real world" is where it is at (in case you weren't already convinced).

[This message has been edited by craig (edited May 30, 2000).]

[This message has been edited by craig (edited May 30, 2000).]

[This message has been edited by craig (edited May 30, 2000).]

[This message has been edited by craig (edited May 30, 2000).]

[This message has been edited by craig (edited May 30, 2000).]

[This message has been edited by craig (edited May 30, 2000).]

[This message has been edited by craig (edited December 31, 2000).]

[Gimbal]

Here is a single cube I did (I'm eating this elephant one bite at a time ).



I'm making a two cube one now with a better light source direction but as you can probably see I need to get the basics covered still.

What's a good method for getting a nice smooth gradiant in value?

[micke]

Kill me.


-Micke

------------------
-Mikael Noguchi-

http://www.katode.org/noguchi/

[Affected]

Damn, I never expected cubes to be so fun.

[General Confusion]

I have included the cloud image to show how to incorporate everything that is being discussed, shadows, forms and most importantly perspective. I have seen this addressed before where it was stated that clouds are just gases why should they have form. Hopefully I haven't jumped the gun on the lesson progression but maybe you can see how to start thinking about form in the context of real applications. For those of you who will undoubtedly tear apart the rendering of the cloud or the sky, for your own sakes don't look at what I've done in that aspect. However do try to see, how what is being taught here has effects on the outcome of anyone of your final paitings and why these foundations should be drilled into your workflow.



------------------
www.geocities.com/genconfusion/

[Frost]

Here are some things that I have noticed... I hope you don't mind;

Mozeman: A few things I've noticed in your drawing (if I may) -- if you placed a line across the base of the cone going through it's center, you'd notice that the cone doesn't seem to be lying on the ground surface. The curve definitions of a circle along a flat plane are always sharper where the VP intersects the rim - making a perspective line crossing the mid of the base might help. Also the shadow cast by the cube on the base of the cone is too high... draw a line from the top edge of the cube to the edge of it's shadow, and you'll see. The Shadows seem unpersistent, following different light sources.

Sedone: The angle of the cast shadows seem pertinent but their lengths are off as the foremost cube should have a longer shadow (judging by the wide angle difference between the shadow cast by each side of the cube, the lightsource is relatively close to the cubes and the shadow should be much longer on it), and faces which are not facing the light source (faces which face a shadow) should be in the second half of the gradient as shown in Fred's first post. If a face is not facing the light, then the light should not affect that face, and should only be lit by the enviroment, such as the ground plane.

Gimbal: Same as Sedone -- bright faces not facing the light source.

...I'm not Fred, Spooge or Francis -- hope I'm not stepping on anyone's feet. =) And nobody's perfect...

I didn't get a comment or answer to my post/question concerning light degradation not being linear... anyone?

Thanks.
frost.

Ed- Never mind... guess I'm not worth any comments, help or thanks... so much for participating...

[This message has been edited by Frost (edited May 31, 2000).]

[AliasMoze]

GC, thanks for the post. I've had a hard time explaining to people why their clouds "stand up" instead of being in perspective. I'll direct them to your pic in the future.

[General Confusion]

If anyone is interested I have the images much larger in case you would like better res and the guide drawing was done in Illustrator so I could provide that document also.

------------------
www.geocities.com/genconfusion/

[Francis]

I might add one tiny comment - you might want to have your clouds converge to a different vanishing point than the (man-made and man-placed) objects on the ground. If you think about it, it's probably unlikely in real life that your clouds would vanish to the same point as everything else in the environment (particularly artificial objects).

But still, great post and great graphic - don't worry about toes!

------------------
TeamGT Studios

[Francis]

quote:

---

Originally posted by Sumaleth:
Francis:

I'm still a little unsure on the two-cube perspective issue - in the image with the cube that is more distant, how did you determine the distance apart that the VP's should be in order for it to give the same "angle of view" as for the closer cube?

It seems to me that if the two distances are not somehow related, each cube will appear to be viewed through a different "lens"?
(snipped)
--

Cheers,
Sumaleth

---

True. If you're arbitrarily choosing vanishing points on the horizon, you're sort of limited to "eyeballing" your other construction lines. If you lay out your cube placement in plan and derive your construction lines from there, you can be sure that you are seeing the two cubes through the same "lens." The thing you might run into in setting up a drawing that way is that if your station point is too close to the subject, your perspective drawing ends up having a fish eye lens feel - this comes from the edge distortion, which happens a lot in one point perpsective. It has to do with the field of view of the drawing as opposed to the f.o.v. of your eye (which I don't know offhand - anyone?).


------------------
TeamGT Studios

[This message has been edited by Francis (edited May 30, 2000).]

[Francis]

quote:

---

Originally posted by Gimbal:
1- Constructing a cube that isn't aligned with the 2 perspective points. My attempt was to draw a square on the ground(in perspective)and use it to draw a circle (ellipse) within it. Then, by eye, I made an 'X' in the circle aligned in the direction I want the cube to be in. I extended the two perpendicular line out until they met up with the eye-line and where they intersected would be the two points I would use to contruct the cube.

---

?? I'm not sure why you are having to do the thing with the "x." It IS another way to pick two new vanishing points, but unless I'm misunderstanding something, you can just pick two new points without having to go through that process.

quote:

---

2- Is there another way to make sure the two cubes are on the same level surface besides using the ol' Mk-1 Eyeball?

---

As long as you have the vanishing points of the two cubes rest on the horizon line, they should appear to be on the same plane.

On that subject, (in case it isn't obvious) there is usually only one horizon line (or eye-line) in a drawing. Objects may have their v.p. above or below the horizon line, which would mean the objects are floating or resting below the ground plane (for example, a sunken roadway observed from ground level).


------------------
TeamGT Studios

[This message has been edited by Francis (edited May 30, 2000).]

[Shorty]

Hello! Thought I should use the opportunity to follow this class. I have just started to learn drawing, so this class is just great!



-sHORTY-

[Francis]

quote:

---

Originally posted by AliasMoze:
Francis,
... It's a good idea to take the shape, copy it to another layer (or use paths), and enlarge it to get the parallel lines coming from the station point (in case anyone was having a problem with it).

---

Thanks! Hadn't thought of that...


------------------
TeamGT Studios

[This message has been edited by Francis (edited May 30, 2000).]

[Francis]

quote:

---

Originally posted by Affected:


I have a couple of questions... First of all, how does one go about making the shadows in true perspective? I used a separate VP fo the shadows, as demonstrated by Francis, but this sort of fakes the perspective, doesn't it? They do not (My english, I'm afraid, fails me, not that I could really explain what I mean in Finnish, either) get smaller towards the horizon. ('course, if the light source is near the objects, they wouldn't do that anyway, but perspective should have some effect on them right?) Or does the fact that the cubes have been constructed in 'correct' perspective also make the shadows automatically correct?

Also, I wasn't sure about whether or not the shaded sides of the cubes should be darker or lighter than the cast shadow cast by the foremost cube on the rear one.

---

Judging by your shadow placement, it looks as if your light source is very close to your cubes (shadows point up), and very close to the ground plane (shadows very long). It is correct, but probably looks distorted to you because of the extreme angles.

As far as shading/shadow values, I'm going to defer to someone better qualified. I'm still learning that stuff myself.

By the way, sorry to everyone about the multiple posts. I'm just responding to posts as I read them.



------------------
TeamGT Studios

[eetu]

francis said:

>?? I'm not sure why you are having to do the thing with the "x." It IS another way to pick two new vanishing points, but unless I'm misunderstanding something, you can just pick two new points without having to go through that process.


i think the point was to help pick the vanishing points so that they appear to be 90 degrees apart, in the perspective dictated by the first two vp's.

(a good idea i think, thank you gimbal

btw thanks for all the detailed advice francis, damn helpful stuff.

eetu.

[AliasMoze]

Francis,
Maybe you should do a brief demonstration of using a good "cone of vision" and how to do it without getting distortion. I would give it a shot, but you are obviously a better draftsman than me. If you have time...

[Gimbal]

quote:

---

Originally posted by eetu:
francis said:

>?? I'm not sure why you are having to do the thing with the "x." It IS another way to pick two new vanishing points, but unless I'm misunderstanding something, you can just pick two new points without having to go through that process.


i think the point was to help pick the vanishing points so that they appear to be 90 degrees apart, in the perspective dictated by the first two vp's.

(a good idea i think, thank you gimbal

btw thanks for all the detailed advice francis, damn helpful stuff.

eetu.

---

I tried jotting down what I was trying to do in the short time I'm home for lunch, sorry for the crude drawing but it might help clear things up. Basically I was trying to do just as Eetu said: finding VP's for a rotated object that come from a 90 degree angle.



I thought of a different method last night before I went to sleep but I forgetten it already. Sorry I don't have more time to explain where I was going with this...or to make a better diagram.

[eetu]

okay, a nagging embarrasment forced me to try this again:



i thought i'd have the cube project the shadow on the cone, but after i had drawn all the projection lines i saw that it'll miss but i didn't have the patience to do it again =)

eetu.

[Sumaleth]

With Spooge gone and Fred held down at work we're unfortunately missing out on the feedback necessary to get the most out of this tutorial. With the second part coming tomorrow (or possibly not?) it would be a shame for everyone to pass over the first task without having really nailed the elements it intended to teach. So since I could sort of see where Spooge was directing the images, I thought I'd quickly run through all the postings since Spooge's last comments and try to give some Spooge-style critiques.

Quick disclaimer; my comments are only going to be based on what Fred, Spooge and Francis have said previously. I won't get anywhere near as accurate as Spooge would have, but hopefully it'll be enough to give everyone something further to think about (and work on) until the second phase is posted.

All the following comments are in the quick, to the point form that Spooge used. If you're easily offended, look away now .

--

Nex: The cubes aren't cubes. The tute was for "white matte cubes on white matte surface" and your scene looks very grey. Left hand cube top would get the most direct light yet the cube top on the right is the whitest. Too much contrast across the top of the middle cube, it appears bent. No bounce light from any surfaces. Ground lighting inconsistant with rest of scene. Shadow should not be black (in a fully white scene there is lots of bounce light). The cast shadow on the third cube (which is basically perpendicular to the lightsource and would likely not really be seen) is too blury for the distance it is from the second cube. The shadows are out of scene when they were part of the task.

Dizzogg: The cubes aren't cubes. Shadows should sharpen closer to the objects. No core shadow (shadow should be darker closer to it's cast object). Very little bounce light incorperated. Both left-side cube faces darken around the leading edges towards the light source when it should be the other way around.

AliasMoze: Cube on the left not really a cube (the one on the right isn't either but it's close). No bounce light incorperated. Cone does not appear "round" - check back to the original reference for shading the cone. Ground is a little grey and doesn't appear to fit the lighting of the scene accurately. Dark side on right cube would be darker.

Nex (2): The cubes are not in the same "perspective space" and both appear to have slanting horizon lines. Cast shadows should be perfectly sharp where they meet the object. Environment is grey. Cast shadows don't appear to be cast consistantly from the same light source. Shading is mottled. Edges of the cubes (particularly on the left) go light or dark along the edges. Bounce light is along the right path but needs to be considered a lot more - look at Spooge's cubes image and look at the different parts of the shadow to see how the light works.

Joachim: (got his comments on IRC last night)

nori: Cubes are different sizes. Cubes appear grey. No shading incorperated. No suggestion of bounce lighting. etc

Trance-R: Cube on left not a cube. Cubes different sizes. Shadow lighting inconsistant - backs of cubes shaded far too light and the cast shadows too dark. No cast shadow from the right cube drawn across the left cube. Ho bounce lighting. Shadows should get sharper closer to the object. Boxes have white lines along some edges.

dines: Perspective incorrect on all shapes (well, except the sphere). Shadows not correctly modeled. Shadows too black. Bounce lighting minimal and incorrect - too bright under the sphere, side of cube gets brighter where it should get darker. The shading is generally wrong on each shape, check back to the original reference. Sphere is too whited out on the front thus loosing it's shape. Shading is quite mottled. Ground is grey.

Alan: Ground is grey. Cast shadows seem to be coming from an inconsistant light position. Cubes not the same size. Left side of right cube far too bright. Top of right cube is shaded with too much contrast - looks curved. No bounce light for the shadows. More bounce lighting can be incorperated on the backs of the cubes.

eetu: Cubes and ground are grey, not white. Cubes aren't cubes. Shading hard to work out. (second one) Tops of cubes too grey. No shading across surfaces. Bounce lighting only half incorperated.

ilmi: No lines, the shape must come from the shading (or as spooge put it "Ahhh linedrawer!!!! get it away!!!"). Inconsistant perspective on the cast shadows (particularly the far two). Shading is difficult to work out.

Sumaleth: Um, pass.

janne: Cube on right not a cube. Cubes appear more grey than white. Dark side of both cubes too dark - loosing the form because they almost merge with the cast shadow. Shadows should loose focus further out from the shapes. Bounce light only half included (from the floor up onto the sides, around the cast shadows, far top corner of back of right side cube etc).

sfr: Whole scene appears grey more than white. Bounce lighting is more random than relative to the way light would really bounce around the objects and environment. Top of right cube too shaded - appears bent.

anticz: Scene appears grey. Bounce light is more design-based than what the scene would really do. Sphere shadow appears too wide. Everything is very blurry.

AliasMoze (2): Left hand cube is not a cube. Cubes appear a little grey. Most bounce lighting not included - ground up onto cubes, one cube onto the other, cast shadows should take away a little light close to the shadowed sides of the cubes, higher up the cubes on the shadows sides should be more light etc.

BooMSticK: Objects and environment appear grey rather than white. No perspective on cubes. Shading on surfaces more random than based on the requirements of the scene. Cast shadows should loose focus as they move out from the objects.

geelimp: Cubes are not cubes. Objects and environment appear very grey. Shadowed side of cobes are getting lost in the cast shadows - loosing form. All shading ogf surfaces is bright in the central region and darker at the edges, they should be shaded cleanly across the surface according to the position of the light source but not so much to suggest bumps or curves. Both close cube edges have white lines running down them. Bounce light from the environment not included or not incorperated correctly.

Affected: Surface gets very grey while still in the scene area. Shadows should loose focus further out from the cubes. No bounce lighting from the environment or the cubes included. Cubes not the same size/shape. Shadow does not have the correct perspective (or maybe the light source is mega close to the cubes?).

Joachim (2): (needs to have a lot less jpeging to really see it We seem to have lost the top of the cone shadow (looks like it should extend past the cubes shadow although hard to tell for sure without drawing lines). The shadow line around the cone needs to be further around the shape to be consistant with the position of the light and also needs to be darker at that point (as well as slightly wider). Front of cone is too washed out, we've lost the shading. Slightly more bounce light needed on the sides as you move up the shape.

Hurri-cane: Inconsistant perspective. Shadow defocusing is odd. Inappropriate bounce light under sphere. Back of left cube is incorrectly lighter down in the center. No bounce light used. No shading on surfaces. Sphere is not rendered as per the original reference.

micke: Different perspective "space" for each cube. Appears a bit too grey. No bounce light incorperated. Not much shading on the cube sides. (the light is probably positioned a little bit too "directly above")

iska: Completely lost the form of the cube top on the left. Cast shadow should generally be darker (in all places) than the shadowed side of the cubes. Bounce lighting is on the way although needs to be more accurately based on what the scene would be doing. Can't see most of the shadow. The shadow line on the right doesn't seem correctly positioned.

Mozeman: Ground appears grey. Front of cone and sphere far too washed out - completely lost the shading and therefore the form. Cast shadow from cube on base of cone goes up far too high and should not be lighter at the base than it is at the top. Shading on near side of cube should be lighter top/left (nearer the light), with bounce light generally up high and a strip of slight darkening along where the cast shadow meets it's base. Cast shadows should grow progressively more defocused as you move away from the object. Cast shadow of sphere seems far too long. Shading on sphere does not appear to go around it's circumference and doesn't look to be at right angles to the position of the light source. Shadow line around the cone needs to come around further (to us), needs to be wider, and needs to be darker (the value needs to match whatever is seen on the sphere in the same position, with bounce light considered). Cast shadow of the cone seems to run away at the wrong angle (at least compated to the cast shadow of the cube).

Sedone: To quote Spooge again; "Ahhh linedrawer!!!! get it away!!!" Bounce light modeling on the backs of the cubes appears to be more of a phong highlight than bounce light. Left side of both cubes need to be darker and have less of a shading gradient (they appear almost curved). Cube on right not quite a cube. Something strange happening at the bottom left corner of the right cube.

Gimbal: Shadow too dark. No hint at form in the surface. Both sides of cube too light. Cube needs to be slightly taller to be a cube. No bounce light incorperated.

Affected (2): Cube on left not a cube. Environment a little grey. Most of the cast shadow out of frame. Cubes appear to be in slightly different perspective "space" to each other. Need more bounce light incorperated into the scene (eg. left side of left cube has the same shading from top to bottom).

Shorty: Cubes apear slightly grey. The back portion of the cast shadow from the left cube does not appear correct (it seems to ignore the entire far side of the left cube). Bounce lighting inconsistant with the scene. Form of top of left cube completely lost. Top of right cube is too shaded giving the appearance of a bend.

eetu (2): Ground has no form - nothing to give it a point of reference in space. Cone is not shaded as per the description in the original posting. Shadows on both objects do not reference the same light position. Left side of cube is a little dark and needs the suggestion of bounce light from the ground surface. There would be the hint of bounce light on the right side/top of the back surface of the cube coming fromt he cone.

synj: Perfect except for the lack of green.

Zaphod: Ground and cube appear grey. Left side is too grey. Task did not call for a gridwork on the ground plane. Ground shading is inconsistant with the scene (should be whitest around the cube). Top surface seems to have an artifect - it's not smooth shading. Back of cube is too dark in relation to the shadows value. More bounce light on both side/back cube faces and the shadow edges.

--

Phew, thats the lot of them.

There are a lot of things that people are consistantly leaving out of their image;

. Spooge and Fred were after the -appearance- of "white cubes in a white environment". This doesn't mean you can paint everything white, it means that, based on where the light source is and the shading of the objects, they automatically -look- like they are white, even when their values are mostly shades of grey.
. If you were doing the two-cubes version, they were to be the same size.
. All surfaces are matte - no reflections or phong/blinn highlighting.
. Shadows loose focus further out from the object that cast them.
. The aim was to try and simulate real-world lighting and in the real world light bounces around from surface to surface. Have a really good study of Spooge's picture (and maybe mine) and work out exactly why each dollop of light/dark was put where it was.

Anyway, I apologize if any of the above comments are misleading, but generally speaking I was really just requoting comments made by Spooge about earlier images. I really recommend going through -all- of Spooge's comments for each image that he looked at (and all of the above, not just your own image), to learn from other's mistakes. So many recent postings make mistakes that Spooge had highlighted in earlier postings that I don't think enough people have been following closely enough - this is a group learning process, don't just blindly work on your own.

Hopefully Fred will be along soon to pick up on anything I said wrong and to take the critiqing to the next level. Feel free to ignore the comments if you feel they are wrong or ask for clarrification if you don't understand something.

Cheers,
Sumaleth.

[Zaphod]

Thanks for the critique!
Just want to add that those lines aren't groundplane grid lines. Just my construction lines (some of them at least). I just thought it looked a little bit cooler that way. And it also looked more 3d.

I didn't add the background until in the end.
So that's why my shading might look weird.

Normally I would correct this instead of explaining. But after I had flattend all the layers resized the image, I accidently saved over the original .psd *grumble*
And I don't fell like doing it all over again right now :/

------------------
/Zaphod
http://www.designmodule.com.bi

[eetu]

sumaleth: wow you did a man's job, great!

i actually thought quite a while about the "Left side of cube is a little dark and needs the suggestion of bounce light from the ground surface"

if the situation was an infinite matte white plane, the amount of bounce light coming to a plane that doesn't 'see' any shadow should be equal.

i think

eetu.

[sfr]

Sumaleth (and Eetu and Francis and GeneralConfusion and everyone), huge _thanks_ for your efforts to explain and critique! This thread has to be the most useful on this forum ever.
(I'll have to take copies of all that immensely useful looking perspective stuff for a later day, when I've found the courage to actually try learning that stuff...)

And a welcome to iska and Shorty - Shorty, you're an Innerloop guy too I noticed, I should mention your company rules

Saffron / Sunflower

[AliasMoze]

Sumaleth,
Thanks! The only thing I disagree on is that my left cube on #2 is not a cube. Trust me, it's a perfect cube, and they are both exactly the same size.

Otherwise, thanks for taking the time! I'll keep working on more cubes.

[This message has been edited by AliasMoze (edited May 30, 2000).]

[sydneyshan]

By the way, if anyone wants the main points of this first tutorial and the life-drawing one in PDF format for referal away from a computer, go to http://shan.bounce.to/forum.html and pick them up.

I hope nobody minds me doing that- but I'm not always near a computer to refer to the forum for direction.

------------------
Shannon Murdoch
www.bounce.to/shan