[Sumaleth]

Thanks for the comments Spooge! I have some followup questions on a few of the comments before I do some fixups.

>Sumaleth cast and object shadows too similar, loosing the edges.

I'm not sure what you mean by this? You mean where the cast shadows join onto object shadows on the cone and sphere?

>Core shadow too far to left on sphere.

By this you mean that the cast shadow from the cube is too low and too much to the left on the sphere?

>Might want to reverse the figure-ground values on cube, surface darker, object lighter.

The cube is already lighter than the ground, if I switch them it will make the cube look darker?

--

I actually wasn't sure what values to use on the dark side of the objects and the ground - with a single light source the shadow areas would essentially be black and the shadows, where they butt up against the objects, would be the same value as the dark side of the surfaces? Or are we assuming that the evironment is enclosed in white walls which give more reflective light to vertical surfaces?

I think I'll have a go at the two-sphere task too, Fred's one was a big ask .

Anyway, thanks again for the comments.

Sumaleth

[Fred Flick Stone]

Yo, been busy, gotta get going toteach class, but to answer Frost's question;
Yes, Craig is the man when it comes to the form, he has been doing it for a long time. He was also technically trained at art center, which focus on the tidbit details of all this, but myself, Moze, Francis, Waco, especially Wacomonkey, who was also industrial design trained, and a co worker, can pick up the ball from here. I was trained in this sort of thing going through what classes I tried, I just focused on other aspects of art. I don't think you should fear losing any quality instruction. Spooge will be back in 6 weeks, that isn't a long period of time, escpecially since we are going so slow with this anyway. Spooge, if there are any drastic wrongs at that point will make it right, but I don't think you should worry one bit. I have faith in all this, I just hope I find a home soon, that's all I am really concerned with.

I will be posting some more information that will help out with these early exercises over the course of the next week. Just trickle in info to help support this topic, i.e. porjecting shadows, bounce light, reflective lighting, etc. this is i=such a finite topic and there are so many things to teach before we even get involved with going into great detail with the Indian portrait.
Keep the art coming in, and have fun with this...

[dines]

here is my contribution, maybe too late ?
I think there is something wrong with the shadow, maybe someone can help ?

Durrenberger David (dines)

[Nex]

Not that I would know this stuff but since I made the same error and spooge told me:.. The shadows should not be solid black and this sharp because we use natural lighting. (as stated in one of the first posts I guess)

[Alan]

I see errors in this, but I am too ignorant to know what to do. However, I thought I'd post it and learn from my mistakes.



-Alan

[eetu]

ok here's my turd (sorry for the un-finish)

i was thinking more about bouncelight than exact shapes (no shit eh )

i also did another one, but i cheated and rendered some cubes and used them as a reference.


to whomever that's critique'ing (sp?) you should probably look at my 2nd try on the next page ...

[This message has been edited by eetu (edited June 01, 2000).]

[ilmi]

now its my turn to submit my cubes.. I know.. its horrible.. but Trance-R's flash pic really help me understand how shadow is made.. thanx..

I really need u guys to point out what is wrong with my cubes and their shadows ... thanx again



------------------
Hyper-D Studio

[Sumaleth]

Here's version two, hopefully with the issues that Spooge highlighted fixed;



(Here's version one; [url=http://users.bigpond.net.au/sumaleth/sumalethshapes1.jpg)]http://users.bigpond.net.au/sumaleth/sumalethshapes1.jpg)[/url]

Personal critiq?

Hmm. I fixed everything that bothered me in the first version and have hopefully made the suggested changes from Spooge's post.

The positioning of the light is really the only problem that I can pick up. With the shapes mostly backlit we mainly see the shadowed sides and it's really hard to give the impression of "white shapes" when all you see is shadow.

Actually, I think the cast shadow from the cube on the sphere gets blury too quickly. Whoops!

Any thoughts Fred?

Sumaleth

[Sumaleth]

A question for Francis;

I've been wondering what the rules are for placing two cubes in a scene which are rotated at different angles?

Different angles means that they will each have their own set of vanishing points - is it simply a matter of making sure that the distance between the left and right VP's for both cubes is the same?

Thanks,
Sumaleth

[eetu]

i think you have to think of the horizon(/the line your vp's are on) as a circle around you rather than a line.
obviously it won't work if your both vp's are a meter off to the same side, you should probably think more in angles than distance.
but no, i don't have any definitive answer, i just pondered it myself an hour ago

so let's wait for someone who actually knows what he's talking about..

eetu.

[janne]

it seems that this is 100st reply to this thread. wow.
well, here is my contribution:



as many of you have noted, this is really hard stuff to do right and i'm also far from it. hopefully we are going to get kicked to right direction soon.

janne

[Francis]

Sumaleth -

That's basically correct, although the v.p.'s don't necessarily have to be equal distances apart. If the cubes are equally far away from the observer, and they are the same size, that is in fact what you'd do.

I think I'll just draw a quick pic.

The first image shows 2 cubes that are the same size, and are about the same distance away from you, so their two sets of v.p.'s have an equal distance between them. Note that this means that the 2 cubes are not aligned with each other anymore, but you probably knew that.

In the second image, each cube has its own set of v.p.'s, and the distances are not the same. That means (although it might be hard to tell from the sketch) that one cube is bigger than the other, and again they are not aligned with each other.





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

[eetu]

yeah, i know, after posting a cube pic like that my opinions have a lot less weight - but anyway:

i have a problem with the 'reflected light on dark side of a ball - rule' given earlier on the thread. if we only have a matte ball in a white surrounding and one lightsource, all light on the shadowed half of the sphere will be reflected light from the environment.
the points on the ball that are facing the cast shadow are the only points in this universe that don't have white floor in front of them to reflect light - and thus one shouldn't have a 'secondary highlight' there. thinking in a radiosity frame-of-mind you should imagine yourself in the place of the point being 'shaded' and visualize how much lit stuff is in your 'field-of-view' if you're looking 'out' of the 3d-surface you're drawing. (and how close and often in what angle if not everything is matte..)

ok let me illustrate. this is shaded more like instructed earlier, with reflected light
on the direct opposite side of the light.


and this one is shaded more like i'm describing

as you can see the reflected light is coming from the floor (and the front), but still i think it looks more natural. (my crappy drawingskills might lose some of my point here though )

here's a psd if you want, flip the layers to get a clearer idea of the difference. (400k)

i can see this in a few of the images posted here, and the most eyecatching it's in the case of A29's image as there is a white reflection coming out of utter blackness. ..and even spooge missed it in his critique

and now for the disclaimers

1. in realworld situations, especially in specifically lit objects, a reflection/fill light like that is often there (and sought after), it's just that we seem to get to the basics here and i wanted to bring this up.

2. realworld materials behave in various ways, the ball might be of a material that doesn't reflect head-on but becomes more reflective as the incident angle decreases, thus lighting up the edges. the floor might also be of material the has a lot of backscatter and would thus bounce extraordinarily much light back into the direction of the light. etcetera.

3. in this business.. if it looks right, it _is_ right - thinking of technicalities like this might just be confusing. i know the people teaching here are on a whole different level than me, so i'm a bit wary of trying to correct anyone.

eetu.

[Francis]

One more thing, and then I'll shut up -

The sketch below shows one way you might construct a perspective to ensure taht you're drawing a cube. The problem with the "pick two points on the horizon line" method is that you end up having to eyeball where the back edge of the face is. If you're anal enough, that probably bugs you a little.

The sketch below is actually 2 drawings - a top view and a perspective view. The top view allows you to determine exactly where your vanishing points should be, as well as how tall the front edge is and also how far back the back edge should be.

The "picture plane" is an imaginary sheet of paper that you are drawing on. It doesn't have to be right up on the corner of the cube - it can actually be anywhere, but putting it there is the simplest case. The station point is where the observer is standing. Because this is a top view there is no perspective - the edges of the cube are parallel and don't appear to converge to a point in the distance like in a perspective drawing. Because of that, we can draw two (green) lines that are parallel to the sides of the cube from the station point to the picture plane. The points where the green view lines cross the picture plane tell you where the vanishing points on your perspective drawing ought to be.

Now on the perspective drawing, you can just plop down the horizon line, and then draw the front elevation of the cube as a reference. (Note that the elevation is placed below the horizon line.) Since you decided to place the picture plane right on the edge of the cube in the top view, you can directly trace the dimensions from your front elevation (they are the same size, being a cube and all..).

The light gray lines on the top view drawing originate from the observer and connect to the corners of the cube. The points at which these lines cross the picture plane give you the locations for the edges on your perspective drawing.



Just as a side note, I went through hundreds of hours of this kind of constructed perspective exercise in architecture school, and then used it on almost a daily basis when I did architectural rendering at the firms I worked at in Texas. A lot of this information can be found in just about any worthwhile drawing text - it's pretty much a standard procedure. The stuff I've been sketching out here are the simplest case scenarios (one cube, one cast shadow, etc.) - since this exercise is about drawing cubes as perfectly as possible, I tried to distill the information down into only what was required for the exercise. The pain in the ass part of this particular method is that it requires a giant area. For example, the greater distance between the station point and the subject, the less distortion ends up in your final drawing. On a number of occasions I ended up putting down pieces of drafting tape on adjacent tables and using a 48" ruler in order to be able to draw large scenes without too much edge distortion.

Anyway, I'm digressing, and I don't want to horn in on Fred and Spooge's project too much more.

Thanks for indulging me.

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

crap.

[This message has been edited by Francis (edited May 28, 2000).]

[dines]

Sumaleth,
did you used a gradient for your shading ?
It's so smooth! what is your secret ?

dines. http://www.nexen.net/koubis

[Trance-R]

Hahaha. Francis, you took Drafting/Tehnical Drawing lessons before haven't you?! I saw so many of those diagrams on my Drafting textbook. You know... I can't shade the planes correctly so I did one in Lightwave. I rendered a set of objects for you guys to study. As I trust the 3D program's shading techniques. The Ambient light in this scene is 0%. I used Soft Filters and Enhanced Low Antialiasing. Along with an Area Light with quality set to four. Btw, the cone isn't smooth, because I didn't do any tweaking. Just used the default settings.


Eetu:Oh yeah, that's what I was gonna ask. Why is it that when you use Radiosity, you get these white circles on the objects? What did I do wrong?
Sumaleth:No dude... area light isn't single ray trace. Spotlight is. That's why you see the soft shadow. Else the shadow would be very distinct and sharp.

[This message has been edited by Trance-R (edited May 28, 2000).]

[eetu]

i think you definitely should've rendered them with radiosity, as global illumination is essential here.
sorry these are a bit f*cked up in post and i don't have the clean ones here.


also one area light.

(sorry for the 3d intrusion)

eetu.

[sfr]

Well, here are my first cubes. I see lots of flaws, and there are probably even more that I don't understand myself... Painted these completely freehand in Painter - I guess I really should have made a perspective drawing first. Next time I'll try to follow Francis's excellent advice, I just have a deep-rooted dislike for technical drawing

Saffron / Sunflower

[leno]

i dont know if the image is showing up, but the image is at http://www9.50megs.com/lenoleum/leno-box.jpg

its a bit late but oh well. its actually the first thing ive ever posted on this board. i would love to have someonee critique it. (specially from francis, fred and spooge)


.lenoleum.

[This message has been edited by leno (edited May 28, 2000).]

[This message has been edited by leno (edited May 28, 2000).]

[This message has been edited by leno (edited May 28, 2000).]

[AliasMoze]

Francis,
Good post. I might add that Photoshop is a great tool for drafting. 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).

[This message has been edited by AliasMoze (edited May 29, 2000).]

[anticz]

Thought I'd give this a try. BTW, thanks Ron, Spooge, Fracis, and everyone else that thats put in the time and effort to help everyone here.

[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"?

I used to know all this (I was studying to be an architect at secondary school but ultimately went on to do engineering at uni *coughwasteoftimecough*). It's just been so long that I only have the very faintest idea of how things worked . Time to buy a book and brush up I think.

BTW, I love those sketches you've been posting in other threads. Great linework.

Dines:

I did use a gradient to create a smooth lightening down the length of the ground shadow (and then subsequently did a lot of painting over the top of it), but the rest was just done with a big (300) airbrush and lots of very light wiping. The original image is actually 3 times larger than the posted version so that helps a bit too.

Trance-R:

The point of the exersize was to try and recreate reality and a single light source raytraced render is only a very calculation-simple approximation of reality. Your images are lacking most of the features that Fred and Spooge have taken their time to explain :}.

eetu:

Those radiosity renders are REALLY interesting! (Lightwave6, Lightscape or something else?) I know it's kinda getting away from the intended purpose of this thread, but it would be interesting to see the results of giving the renderer this tutorial task as accurately as possible.

--

Cheers,
Sumaleth


[This message has been edited by Sumaleth (edited May 28, 2000).]

[AliasMoze]

Here's my next image. I think this one is a little better. I tried to throw in some subtle shifts in value to make it more interesting. Anyway, please give me a grade, Fred. Thanks.



------------------
AliasMoze
:) :) :) :)
"That activates my hilarity unit."

[BooMSticK]

Well here's my bad excuse for using your bandwidth. Should probably be more focused on perspective next time Still this was as much a test for me getting into paths in PS... Still much fun, and pretty damn hard too!

,Boomer

[geelimp]

I had to do one too I know it is not a cube, sorry about that.



lines: http://www.3dluvr.com/glimp/images/boxes_w.jpg

[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.

------------------
Affected

Democracy is a lie

http://affected.xs.mw

[Joachim]

Well, here's my second attempt...I hope/think that it's better than my first, atleast. . . . I'm really looking forward to some feedback now, hint hint


eetu, I fixed the bouncing light, as you said, thx again !

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




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

[Hurri-cane]

here !...i just wanted to get it done...no help with lines, logic....only eyes...if it looks kinda good, ill be happy here goes..



i know the errors...the shadow of the top box...perspective faults...ya ya ya ya...GIVE IT TO ME RAAAWWHHH !!


------------------
-H-U-R-R-I-C-A-N-E-
[email protected] http://www.geocities.com/hurri_cane_1999/
-------------------------

[This message has been edited by Hurri-cane (edited May 29, 2000).]

[eetu]

joachim: i'm not one you were hoping to get feedback from, but i'll do this anyway

(this is a continuation of my rant above, please feel free to skip this)

the cone: you are following the 'rules' given here a bit too closely. i mean the "the form shadow is darkest near the shadow equator" one. think about what kind of surfaces are near. the gap between the cube and the cone is the darkest place on that world. no reflective/bounce light coming from there. instead the white floor bounces light quit well.

the cubes: better, although the shadow-side faces should rather have a slight upward gradient, the areas closest to the cast shadow receive the least bounce light.



here's a psd with the layers you can flip.

anticz: you're getting bouncelight from the cast shadow as well :P

i feel really dubious here trying to correct people who are obviously a lot better graphicians, but i know i have a point
sorry fred et al if i'm a bad pupil

eetu.

[Gimbal]

Thanks for the link, Aliasmoze! Long live the internet is right. Here I am getting art lessons from a group of great artists while I talk to my best friend on an almost daily basis who lives 800 miles away over ICQ as I'm listening to .mp3's and downloading a new game demo. I'm happy...

I just got back from the Universal Studios theme parks late last night and haven't been able to get my butt in gear and contruct some more cubes yet. I did lots of research on light and shadow while I was there though. Everytime I was waiting in line for a ride I was observing the shadows on the poles that make up the zig-zagging maze they herd us through, the myriad of different faces and how they are constructed, the spiraling cylinders that make up the roller coasters, everything I saw was broken down into primitive shapes. It was good practice to get into that mindset and greatly helped pass the time in that hour wait for the Spider Man ride.

Guess I'll get something to eat and get to making more cubes.