[ Showcase ] GoW - Age

[Oneironaut]

Thx for the compliments!
I think, making the Age playable is a long way.
For this i need help, or better, become member of a team.
But first, i have to read, read and once again read the forum ^^
This will take some time...

[Oneironaut]

Update...
UV Unwraps on the left side are also finished.

[Christian Walther]

Looks very cool! Question from the interested layman eager to learn from the pros: Is this model composed of few large closed meshes, or does it consist of many small, partly open and overlapping, components?

If I may stray from the topic a little and extend the question to all the experienced modelers around here (and let me know, Oneironaut, if you think I'm hijacking your thread with this):

I have the vague impression that amateur modelers (to whom I count myself, and also because I come from POV-Ray where that kind of modeling is the norm) tend to think in aggregations of solid primitives, while professionals more often create contiguous closed meshes.

Is there any truth to this? And if so, what is the reason, or what are the advantages of that style of modeling?

For example, let's say I want to model the solid shape at the left in the following picture.

open-closed-meshes.png (7.73 KiB) Viewed 3885 times

I can either do that using two disconnected, intersecting meshes, one of them open, as in the middle of the picture. (An even more complete newbie than I would perhaps even make both meshes closed, adding one completely redundant face.) Or I can do it as a single, closed mesh, as on the right side.

What are the advantages and disadvantages of either style?

Some guesses: I can see that the middle one has three faces (quads) less than the right one, and therefore probably renders faster, even though it creates some more fragments that will never be visible (that's my unfounded impression, at least - I haven't measured it). The right one somehow looks cleaner, on the other hand. I could also imagine that relying on depth buffering to create the correct intersection lines, as in the middle one, might sometimes suffer from the limited resolution of the depth buffer.

[Oneironaut]

Looks very cool! Question from the interested layman eager to learn from the pros: Is this model composed of few large closed meshes, or does it consist of many small, partly open and overlapping, components?

The Models in the Picture above are partly open but not overlapping.
Why? I think its a question of the workflow while creating. I "prefer" this type of modelling
because it fits my needs best. Im not a Pro, i would call me Amateur or Hobbymodeller.
The thing i know is that the Uru engine supports the way im modelling.

If I may stray from the topic a little and extend the question to all the experienced modelers around here (and let me know, Oneironaut, if you think I'm hijacking your thread with this):

No, I'm glad to help in any case.

I have the vague impression that amateur modelers (to whom I count myself, and also because I come from POV-Ray where that kind of modeling is the norm) tend to think in aggregations of solid primitives, while professionals more often create contiguous closed meshes.
Is there any truth to this? And if so, what is the reason, or what are the advantages of that style of modeling?

The advantages of the solid modelling are in my point of view the flexibility. If i take your picture and count the models from left to right 1-3,
the second has fewer faces but also fewer faces to edit. Another advantage is, that you can convert the third model in small(?less)time to the second.
If you want to do this vice versa, 2. to 3., you have to draw new faces.
There are many pro and contras. I will go more in detail at the end of the Post.

Some guesses: I can see that the middle one has three faces (quads) less than the right one, and therefore probably renders faster, even though it creates some more fragments that will never be visible (that's my unfounded impression, at least - I haven't measured it). The right one somehow looks cleaner, on the other hand. I could also imagine that relying on depth buffering to create the correct intersection lines, as in the middle one, might sometimes suffer from the limited resolution of the depth buffer.

You got the point. In fact, both ways are right!
My opinion is, once you have learned (or mastered) the programm youre working with you can choose
between both solutions. Its a progress of trial and error. Ive thrown a lots of model away because i made mistakes during creation.
At the next model im trying to prevent these mistakes. After some time the way of modelling "grew" due this progress.
Sadly, i cannot help you with blender, but i wish to help on any way i can. (If you can bear my grammar ^^)

[Oneironaut]

Here are the first two models from this Topic for Open source.
Credits are not needed.


EDIT:
(Thats strange, unescpected end of archive. I think it has to do with the name: *01.rar, *02.rar. Like a split archive)

I've taken the files offline, i look forward to upload them to the OpenSource libriary at GoM

[GPNMilano]

Looks very cool! Question from the interested layman eager to learn from the pros:):

I have the vague impression that amateur modelers (to whom I count myself, and also because I come from POV-Ray where that kind of modeling is the norm) tend to think in aggregations of solid primitives, while professionals more often create contiguous closed meshes.

Is there any truth to this? And if so, what is the reason, or what are the advantages of that style of modeling?

Christian, to me, the third one in your picture is the best one to use in terms of Uru. Reason being is that is that an object, with pyprp, can have multiple materials. This saves, in the long term, on the amount of separate objects in an age and thus on the amount of export time, as well as engine performance in Uru. Second, is vertex coloring, in Uru, is a large portion of how Cyan, and we in term, paint our models. Most cyan stuff used vertex colors, rather than material coloring for materials. The way it is in Blender, (not sure if this is the same in 3DMax or not) is that when you paint a vertex in blender, it literally paints just that vertex that color, so if you had a flat plane, and painted three vertex black, and another white, you'd get a look one corner painted white, that fades to grey as it reaches the black part of each of the other vertexes. So in effect you'd have a triangle on one side of the mesh thats white and grey, and the rest of the mesh would be black.

So, say you wanted to paint the crease of the model picture you posted, a darker color, to acheive the look that its shadowed where the two meshes meet. You'd paint that a darker color than the rest of the mesh. With two separate objects, that overlap one another, you won't get the same faded look as if they were one entire mesh.

[Oneironaut]

Ahh, this is an intresting Point you are mentioning, GPNMilano.
I havent used it yet. That goes on top of my list.
The Alpha Channel can also be controled with vertex paint, right?

[GPNMilano]

Ahh, this is an intresting Point you are mentioning, GPNMilano.
I havent used it yet. That goes on top of my list.
The Alpha Channel can also be controled with vertex paint, right?

Yes, the way PyPRP is set up to export an age it export its alpha channels based upon the vertex color layers. So any object with two vertex color layers (alpha and color) we'll automatically set the alphablend flag in Plasma, and use the alpha color layer for its alpha channel. So as an example, we'll take the easiest thing, a shadow for a book. You make a single plane, little bit larger than the book. have a simple black texture, very small say 64x64. and set that texture for your plane. Then you vertex paint the whole mesh white, and name it color. Then you make a second vertex color layer, and fades to white along the edges of the plane. Name that vertex layer alpha. Upon export, PyPRP will use that second layer as the alpha channel, and the first layer as normal vertex colors.

This is how you achieve the look that you see in Relto, where the islands seem to fade into the distance when you look down through the clouds.

Regular textures that have have transparency already however don't require an alpha channel. So a lamp glare doesn't need one, (though cyan uses them here and there depending on the type of lamp glare they are going for. They still need the alpha blend flag though, which you can automatically set by selecting the "usealpha" button in the blender texture screen.

[Lontahv]

The vertex calculated light (and vertex-paint) looks better if you have connected meshes.

Sorry for farther moving this off topic. I'm just going to add this one post.

EDIT: Removed long quote

[Oneironaut]

It seems that I'm nearly finished with modeling.
Now, the paper-warfare can begin

What is?
~ 11.000Polys
~ 16.500Tris
- 70x90Meter

What will be?
- Detail
- HQ Textures
- Five more Districts (step by step)

What is needed?
- A name
- Contributions
- Feedback and Organisation
(like a Moderator with an observing eye
or contacting the right people at GoMa etc.)

First I'll ask boblishman for the techical realization.
Follow this topic for infos