3ds Max - Tutorials - 20070606 - Onnovanbraam

Onno van Braam - www.onnovanbraam.
com - Tutorial Collection 2007
3Ds Max Tutorial

Collection 2007
Onno van Braam

www.onnovanbraam.com
1
Onno van Braam - www.onnovanbraam.com - Tutorial Collection 2007
Copyright © 2007 Onno van Braam, EVOlution WebDevelopment
All rights reserved. No part of this book may be reproduced or transmitted in any form of by
any means, electronic or mechanical, including photocopying, recording or by any information
storage retrieval system, without permission from the Author.
Although every precaution has been taken in the preparation of this book, the publisher and
author assume no responsibility for errors or ommissions. Neither is any liability assumed for
damages resulting from the use of the information contained herein.
Trademarks
All terms mentioned in this book that are known to be trademarks or service marks have been
appropriately capitalized. Onno van Braam cannot attest to the accuracy of this information.
Use of a term in this book should not be regarded as affecting the validity of any trademark or
service mark.
Warning and Disclaimer
Every effort has been made to make this book as complete and as accurate as possible, but
no warranty of fitness is implied. The information provided is on an “as is” basis. The author
and the publisher shall have neither liability nor responsibility to any person or entity with
respect to any loss or damages arising from the information contained in this book.
2
Table of Contents
MODELING.................................................................................................................................4
Polygon Modeling 1: Introduction...........................................................................................5
Polygon Modeling 2: Vertex Sub-Object...............................................................................11
Polygon Modeling 3: Edge Sub-Object.................................................................................20
Polygon Modeling 4: Polygon Sub-Object............................................................................23
Polygon Modeling 5: Techniques I....................................................................................... 32
Polygon Modeling 6: Techniques II...................................................................................... 40
Polygon Modeling 7: Car Body.............................................................................................51
Polygon Modeling 8: Car Rims.............................................................................................61
Polygon Modeling 9: Car Tyres............................................................................................74
Polygon Modeling A: The Process of Modeling a Car..........................................................84
Introduction to Splines..........................................................................................................92
Spline Car Modeling...........................................................................................................100
TEXTURING............................................................................................................................106
Map Channels.....................................................................................................................107
Multi/Sub-Object.................................................................................................................114
Composite Maps.................................................................................................................117
Car Mapping.......................................................................................................................134
OTHER....................................................................................................................................145
Blueprint Setup...................................................................................................................146
Depth of Field Using a ZDepth pass...................................................................................161
Making of... the Shelby Mustang GT500 Wheel.................................................................177
3
MODELING
4
Polygon Modeling 1: Introduction

Conventions
- In the tutorials sometimes next to an action there is a number, for instance like (1), this
refers to the next image in the tutorial and then the part of the image. So (1) refers to the most
left part of the upcoming image.
- EP: Editable Poly
- VSO: Vertex Sub-Object
- ESO: Edge Sub-Object
- PSO: Polygon Sub-Object
- MS / MS'ing: MeshSmooth / MeshSmoothing
Summary
In these tutorials I will be covering the basics of editable poly-modeling in 3D Studio Max.
From all the tools available when using editable polies to how to apply them and maybe more
importantly: when to use them. I'll assume you have a basic idea of how 3D Max works, and
where to find things, but don't worry if you don't since I'll try to use many, many (partial)
screenshots and icons to make clear what I do. At the moment I use Max 5 so everything you
see will apply to Max 5, but from what I have seen from the so called 'new features' PDF file
of Max 6, they didn't improve any of the modeling. And therefore everything explained should
apply to Max 6 without any problems.
State of Mind
I do not claim to know all, or that my techniques are the techniques and all other sucks. On
the contrairy, everybody should develop his own style of modeling but it's always very useful
to see other people do it. Since then you will see what he/she does differently and makes you
5
think about your own and his/her way. You'll balance them in your head, and try to see the
pro's and con's of each way and then decide: hm, this is smart, I'll try to use that, or think:
bah, my way is more efficient or a bit of both of course. But any way: you are thinking about
what you do and how you do it! This is essential, and is the basis of these tutorials: why do
you do what you do? Have you ever thought about it? Probably because you think that what
you are doing is best... and that may be true, but when you only know your own techniques,
you will never know.
It might be that this all sounds like bollocks, but I think that in the end knowing how to model
is all about a certain state of mind, an intuition, which is aquired by doing it a lot and looking at
other peoples methods and techniques. Then when you have aquired this intuition, then and
only then will you make progress in skill and art like never before. It will be this moment when
you have confidence enough to jump in and model what you want! I must warn though, this
moment doesn't come easily. It takes effort and dedication.
Alrighty, enough of the philosphical / psychological mumbling.
Editable Poly
Editable polies are a certain type of objects used in 3D Max. There's others like editable
meshes, NURBS, spline objects etc., but the one we will be using solely is the editable poly.
Why you ask? Because of many reasons:
- Do what a program does best: Max is perfect for polygon modeling if you ask me, as where
Rhino rules the NURBS modeling waves.
- Editable poly has a vast amount of built in tools, opposed to editable meshes.
- EP's (I'll use EP from now on, in stead of Editable Polies) are very intuitive, even though
when using MeshSmooth (which will be covered extensively) you are not WYSIWYG'ing
(what you see is what you get), it still is a very simple, easy going way of modeling. Vertices,
edges and polygons are simple and easy to comprehend mathematical objects.
- I use it for almost everything that I model, so it's logical that a tutorial I write is about the
6
technique I use.
EP's are built up from vertices (zero dimensional objects: points), edges (1 dimensional
objects: lines), and polygons (2 dimensional objects: surfaces). These are the three building
stones with which almost everything in the 3D modeling world is built. You may create them
differently (using NURBS, splines, meshes etc.), but when rendering you just need polygons
to actually see something.
A Simple Start
Okay, we'll be starting here: start 3D Max, create any object of any size, it's just for explaining
purposes, so for instance a box, then select it, right click on it, find the 'Convert to:' option and
select 'Editable Poly'.
An other way of doing this is creating the object, open the modifier tab, right click on the
object name (the one in grey) and select 'Convert to: Editable Poly'. In short: you need an EP.
7
When modeling with EP's, there are so-called sub-objects. I call it: 'you can go into vertex
sub-object mode', by which I mean that when you go into this mode, you can edit the
positions of the vertices of your EP. You go into sub-object mode by left clicking 'Editable
Poly' in the modifier tab (Ctrl + B does the same), it then becomes yellow. By clicking the plus
sign on the left you can expand it and see which sub-objects are available for your object, if
everything is okay you have 5 options: Vertex, Edge, Border, Polygon and Element. I
personally never, ever use Border and Element, so they won't be covered in detail. With the
shortcut 'Insert' you can change the sub-object to the next one (try it!). When you left click
'Editable Poly' again (or press Ctrl + B) you go out of the sub-object mode.
You will notice soon enough that each sub-object is very different, yet very similar in what you
can do. You can do almost the same in each sub-object, but the effect is not always the
same! There is a sort of symmetry or cohesion between the sub-objects which is very
beautiful. For instance you can 'Extrude', 'Grow / Shrink' in each Sub-Object, they do about
the same thing in each SO. Yet certain aspects are exactly the same in each sub-object and
do the same thing (only applied to vertices, edges or polygons depending on in which sub-
8
object you are, for instance the Soft-Selection works straight forward in the edge and polygon
sub-object, once you understand how it works in the vertex sub-object). Therefore I won't be
explaining (or just skipping) certain options since they come back three times...
There's also options which I will skip completely, either because I have never used it, or
because I think it's totally useless.
In the next section I'll explain my view on the sub-objects. Do not think this is all true or the
general definition, it's just how I see it. I won't go into detail about what you can do with them,
yet. This will be extensively explained in Tutorial 2, 3 and 4.
Vertex Sub-Object
The dots, the points of your model: vertices. Obviously when modeling you'll want to change
the positions of your vertices since that is the bare essential part of modeling, but something
you may not have realised: VSO (vertex sub-object) is mostly used to shape your model.
Edge Sub-Object
The lines of your model: edges. I mainly use this SO for fine-tuning. It's the place to be when
your model has the right shape, but you want certain parts to become less smooth when
using MeshSmooth. Chamfering will be the magic operation.
Additionaly ESO is very handy for creating new surfaces.
Polygon Sub-Object
The faces, surfaces of your model: polygons. When you are in PSO it's usually because you
want to create something, hardly ever to change your model. Not very often would you want
9
to change the form or shape by moving / scaling / individual polygons. Yet, when you would
want to create additional blobs, extensions, whatever, to your model then the PSO is the
place to be.
When reading tutorials 2, 3 and 4 and see what you can do in each specific sub-object, it'll
become clear why I said certain SO's are used for specific modeling actions.
10
Polygon Modeling 2: Vertex Sub-Object

Summary
In this tutorial, I will try to explain all options available when modeling in the vertex sub-object.
It may look like just a long list, but I'll will try to explain when a certain modeling action in the
VSO is useful and give examples of when you could use it.
If you haven't created an EP (editable poly) yet, then do so now, by creating a new primitive
and converting it to an EP by selecting it, right click, 'Convert To:', 'Editable Poly'. Then go
into sub-object mode (Ctrl + B) and to the Vertex Sub-Object (Insert, or click the plus and
select 'Vertex'). Then, if you haven't done it yet, open the modifier tab. What you should look
like right now:
Under this you can see all the options you have in the VSO.
11
Selection
Under this roll-out you can switch to other sub-objects, as indicated by the five icons, right
now the vertex one is higlighted since we are in that mode. While we are in this mode, notice
how in your viewport all the dots of your model are shown more clearly, they may have
become little plus signs, or small blue dots for me as seen on this screenshot (switch between
wireframe and smooth + highlights of your viewport by pressing 'F3'):
So now you could select one, or more vertices by selecting them and then moving (shortcut:
W) them. Selecting multiple vertices can be done by pressing Ctrl, deseleting vertices can be
done by pressing Alt and then clicking the ones you don't want. This is always the case in
Max: Ctrl adds to your selection, Alt subtracts it from your selection. Moving can be done by
using the XYZ euler , which works very intuitively.
12
This is all basic knowledge and you really should be aquainted with it before continuing with
this tutorial.
Also notice the little line: 'Vertex 6 Selected'. This may seem insignificant, but it's not. It can
be extremely helpfull, since sometimes you have, without knowing it, or by accident created
additional vertices which are at the exact same position as other vertices, so you don't see
them, but they do affect your model, for instance when using MeshSmooth. With this line you
can see that, since you could drag select certain parts of your model and expect a certain
number of vertices and then when this doesn't match the number shown, then you know
something is going on! I'll show an example under the Edit Vertices Roll-Out.
So first up:
Shrink: This is an option to make your current selection shrink (unsurprisingly). Keep in mind
when using shrink, that it needs a 'hole' in your selection to be able to shrink. For instance
when I select all the vertices of my model and then press 'Shrink' nothing happens. It needs a
start, so select all vertices then deselect (Alt) one and then press shrink to see what happens:
all the adjacent (connected by a single edge) vertices to the one you just deselected, are also
deselected.
Grow: The inverse of shrink and also works in the same manner: if you haven't selected
anything then there is nothing to grow from and it won't do anything. But select one vertex,
press 'Grow' and see that all adjacent vertices are selected. This I use more often than shrink,
usually when I have a large model but when the model is made up from separate parts: grow
doesn't continue growing from one loose part of a model to another. When there is a physical
gap, so no connection between two parts in the form of an edge or multiple edges, then grow
will stop there (but of course will try to grow until one whole part of the model is selected).
Ring and Loop: This is 3D Max logic at it's best: they're there, but you can't use them. They'll
13
be available in the Edge sub-object (and be very very useful).
Soft Selection
Before you can use Soft Selection, you click the radio button 'Use Soft Selection', only then
will all the parameters become available. Up the value of Falloff until you see the color of your
vertices change: the color indicates how much influence it recieves from your originally
selected vertex/vertices. Soft selection is exactly the right description of what it does: you
select certain vertices, but softly, so that when you move those vertices is has a certain
feedback on vertices which you didn't select. They kind of drag along with your originally
selected vertices. Try it out and play with the Falloff, Pinch and Bubble parameter to get a feel
of what it does. Shade face toggle we will come back on later, in the tutorial about the
Polygon Sub-Object.
I personally hardly ever use it, mainly because it has a vagueness and non-controllableness
that I don't like. Especially when modeling technical and detailed stuff it's in my opinion
useless. On the other hand, soft selection can be extremely handy for organic modeling!
Deactivate the Soft Selection again.
Edit Vertices
14
Remove: Not really sure why on Gods green earth they made a button to do this, but ah well.
Select vertices, press this button and they are removed. I have seriously in all my years of
modeling, never ever used it. Just select what you want to delete and then press... Delete!
Amazing, but true.
Break: When you perform this on selected vertices it breaks them loose from the current
model, in a copy kind of way. :) It creates a number of vertices which is equivalent to the
number of edges comming together in your original vertex. I'll show what I mean in the next
three images (with the text shown in the selection roll-out):
Obviously something has happened, since we went from one vertex to three. When you move
each one of the three vertices away from it's original position you can see what has
happened:
15
They are all independent.
Extrude: This one works a little differently from the buttons we have so far, since we can
either do some extruding manually (press the button) or do it by the numbers (press the little
window next to it). Try extrude on a single vertex (either way, manually or by the numbers)
and see what happens. You get a spike from the original vertex. I am not quite sure what to
make of it and when to use it. Maybe, now that I have fooled around with it for a bit, I can see
a purpose for it. Extrude creates something, and as mentioned in the introduction, I deem this
sub-object for shaping mostly.
Weld: Ah! Weld, now we're talking. This is something I use very, very often. It melts two or
more vertices together. So if you would have performed a break operation for instance and
then selected the three newly created vertices and then pressed weld then the Break
operation is undone in a non-undo-pressing kind of way. Here also there is the manual and
technical version. In the by-the-numbers one you can set a threshold which is effectively a
sphere around your selected vertices: all vertices within the sphere of influence of an other
vertex will melt together. And they will melt at the average of their positions. The sphere
cannot be drawn manually, so by pressing the button you apply welding with the threshold as
set in the by-the-number window.
Target Weld: The brother of weld, but used more often by me. With this you can drag one
16
vertex, in your viewport, to another one and it will melt the first onto the other, without any
averaging of positions. Click it, then select a vertex of your model (click and drag, notice the
dashed line!) and drag it to another vertex: congratulations, you have target welded! \o/ It is
extremely handy for cleaning up your mesh or correcting certain modeling operations (double
chamfers for instance, but I'll come to that in Tutorial 5 about Modeling Techniques).
Chamfer: Chamfer is in this sub-object almost the same as extrude (try for instance an
extrude with height = 0). But anyways: chamfering. Chamfering is one of the essentials in
polygon modeling, yet not in vertex sub-object. But what it does is create an additional face.
Here I selected one corner vertex and was in the middle of chamfering it:
As you can see, it smoothens it out a bit and creates a new, triangular shaped, face.
Connect: This connects two (and only two, not more, not less) selected vertices, by making
an edge that connects them. But only if the two vertices are part of the same face. You still
with me? No? Good, since it's totally useless.
17
Remove Isolated Vertices: It removes vertices that have decided to be alone. :) You would
have to seriously be doing something weird to create a single or more vertices that aren't
attached to anything. Something like this for instance (it would remove the one indicated by
the arrow):
Remove Unused Map Verts: I really wouldn't know. But, quoted from the manual:
Certain modeling operations can leave unused (isolated) map vertices that show up in the
Unwrap UVW editor, but cannot be used for mapping. You can use this button to
automatically delete these map vertices.
So there ya have it. I have never even gave it look, up until today. PS: You see how Dicreet
(the makers of Max) are toying with us here? Tsk, using 'verts' instead of 'vertices' because
else it wouldn't fit in the tab! PAH! :)
Edit Geometry
Create: Creates a vertex at the position where you click. This is for hardcore modeling only.
Collapse: Mighty handy tool, which resembles welding: select the verts you want to 'collapse'
and press 'Collapse'. Then all selected vertices will collapse into one single vertex at the
position, which is the average of the original bunch.
Detach: Don't ever use Detach in the VSO. Ever. It's just useless. It cannot be what you want
18
to do. Believe me.
Hide Selected / Hide All / Hide Unselected: I advise not to use this in any sub-object, but
only on complete models / meshes.
The rest of this section will be covered in the Polygon Sub-Object tutorial, since it's the same
for all sub-objects and because there is some small differences concerning selection of a part
of your model and then applying one of the operations available here and that difference is
best explained in the PSO.
The last three roll-outs (Vertex Properties, Subdivision Surface and Subdivision
Displacement) I don't use. Ever. They may be handy, but I have lived without even giving
them a single look.
19
Polygon Modeling 3: Edge Sub-Object

Summary
In this tutorial, I will try to explain all options available when modeling in the edge sub-object.
Under this you can see all the options you have in the ESO.
Selection
Shrink: Shrinks your selection, just as in the VSO but now applying to edges.
Grow: Grows your selection, just as in the VSO but now applying to edges.
Ring and Loop: Ring and Loop are very similar. They make selections easy. Ring does it
perpendicularly and Loop parallel to the edges' direction. To see what Ring does I selected
one edge, then pressed Ring:
20
On the other hand if I would select a different edge and pressed Loop I would get this
selection:
This obviously also works with more complicated meshes and shapes and is therefore a very
usefull tool when making complex selections in a fast easy way.
Edit Edges
Remove: Removes an edge you select. Works a little subteler than hard deleting it, in that it
rebuilds the polygon from which you deleted it.
Split: A sort of detaching, utterly useless.
Insert Vertex: Don't use that here, use 'Cut' in the PSO.
21
Weld: Welding Edges? Yeah right. Do not use this. Looks like nonsense to me.
Extrude: Creates a sort of spike, so very usefull if you want to make a spike, or when
selecting a whole Loop it could be used the create an outward seam for instance.
Target Weld: I think they forgot to turn this button off. It doesn't do anything for me.
Chamfer: THE tool for giving your model the right shape when you're modeling with the help
of MeshSmooth. This tool will be extensively used in all further tutorials, 5-9. It splits an edge
into two and thereby creates a new polygon. Can be set either manually or 'by the numbers'.
Last one is preferable if you want a consistent model.
Connect: Connects edges, by creating new ones from the middle of one edge to the other.
Also works on a larger selection. For an example between two edges see the image below:
22
Polygon Modeling 4: Polygon Sub-Object

Summary
In this tutorial, I will try to explain all options available when modeling in the polygon sub-
object.
Under this you can see all the options you have in the PSO.
Selection
Shrink / Grow: Works exactly the same as in the vertex sub-object, only now it grows /
shrinks by using polygons.
Edit Polygons
23
Insert Vertex: Use this if you really want a vertex at a certain position on a polygon. Just click
'Insert Vertex' and click where you want it on your mesh. Do notice the 3 or 4 edges it creates
along with it! This is a good way to make your mesh ugly. Try not to use this. Better use 'Slice'
and 'Cut'.
Extrude: Extrudes a polygon, either by it's own normal (perpendicular to the surface) or as a
group. Click the do-it-yourself button after selecting the polygons you want to extrude and
click-drag. Or do it 'by the numbers'. Then you have three extrusion types: Group (all in one
averaged direction), Local Normal (as a group, but each polygon along it's own normal), By
Polygon (each polygon along it's own normal and individually, so they won't form a whole).
See the next image, where I first have selected all the polygons in the middle of a cilinder,
and then applied one of the three extrusion types:
It's not like one type is the dominant one, they all have there purpose and I use them very
often.
Outline: Scales the polygons you have selected (either by a certain amount, or manually).
Could be useful, since it's not 100% a scale, since it scales each polygon, or group of
polygons, so that they still have the same normals.
Bevel: Bevel is a combination of Extrude and Outline. So when you click drag you extrude,
then let go and then when you move your cursor it outlines. This is actually a quite nice
feature and I use it quite often.
24
Inset: Inset is about the same at Outline, with the difference that it creates new polygons,
which is handy! So you'll be able to make polygons sit sort of inside others. This sort of
operations can also be done with slicing, but for quick-and-dirty work, Insetting is the way to
go.
Hinge From Edge: It does what it says: it hinges a polygon around an edge, so you'll create
an attached prism... sort of. You can set the number of segments and the angle it needs to
rotate when you do it 'by the numbers'. There you can also select the edge it needs to rotate
over. Quite a nice feature which came in in Max 5 (people asked for it). I hardly even look at
it. See the image below to get an impression of what it does:
Extrude Along Spline: This is a very, very fancy feature, which is handy when you try to
model organic stuff, but also for technical modeling it could come in handy. Basically you
draw a line, then click the button 'Extrude Along Spline', then click the line and it extrudes
your initially selected polygons along that spline. Obviously drawing a spline with a certain
shape is much easier then extruding all the time and trying to make it fit the line. Options such
as Segments (the number of segments used to try and make it look like the line), Taper (to
make it smaller near the end), Taper Curve (to make the smaller-going go linear or curvi-
linear) and Twist (to give a sort of twirl / twisted look) all are very nice. It's actually more
usefull than I thought.
25
Edit Geometry
Repeat Last: Don't understand why they bothered making a button for this, isn't there just a
global undo button? Because doing the exact same thing again but on a different polygon
doesn't make too much sense to me.
Create: With this you can create polygons, but not by a single click. You have to click a
number (3 or 4) of vertices in the right order and then it will create a polygon for you. So for
example, I clicked create and saw the first part of the image below (you will always see the
vertices of your model when you click 'Create' in the polygon sub-object.
So here I first click the right-top one, then then the left-top one, then the left-bottom one, then
26
the right-bottom one and then to complete the procedure, the right-top one again. Then it
creates a polygon and selects it for you.
Collapse: Crashes your selected polygons into one point (vertex). Same as in the vertex sub-
object.
Attach: Attaches other meshes to your current one. It's a bit of hardcore grouping.
Detach: The inverse of attach: make a selection of polygons, click detach and select the
option you want. A complete Detach, so it will become a separate object, 'to Element', then
you detach it, but it's still part of the same mesh, or 'As a Clone', then it copies the selection to
a new mesh and keeps the old one.
Slice Plane: This is one of the most important functions in the Polygon Sub-Object. Select the
polygons you want to slice (2), click 'Slice Plane', you'll see a plane gizmo appearing (3),
rotate and move it until it slices your selection in the way you want it and click 'Slice', lo and
behold for the result (4).
Slice: Performs the slicing as described above.
Reset Plane: Resets the Slice Plane gizmo to it's original position, extremely useful!
Quickslice: Quickslice let's you slice selected polygons, but dirty. It let's you 'draw' the plane
that slices, by drawing only one line, so this is usefull if you would have wanted to slice while
27
working from a orthogonal view. Do not use this for fine slicing, it's a very rough and very
quick method. Works nice, but don't use it too quickly or actively.
Cut: With Cut you can draw edges onto your polygons (they don't have to be selected), you
just click from point to point where you want an edge to be and it will create the necessary
edges and polygons to do as you tell him. This is very handy for drawing shapes onto your
model which you could extrude then for example. This tool is best used from already existing
edges to other already existing edges, or you'll get the same problem as with insert vertex: it
does what it's told to do, but along the way completely fucks up your mesh. Cut is a vital tool
when modeling, try to understand how it works, and what it's strong and weak points are.
Tessellate: Tessellate is kind of weird at the beginning and I have only found one purpose for
it so far (which will be covered in Techniques 1). When you select a polygon and then
'Tessellate' it slices it two ways, exactly in half, so you'll get four pieces where you started with
one. Would you press it again, then it would slice each and every one of the four newly
created ones halfway, twice etc. So it's kind of a quickslice if you want to slice halfway twice.
Polygon Properties
28
Material ID: Used for texturing purposes (or selecting purposes) mainly. I am not going to
cover texturing so I won't go into detail, but in short you can give polygons a Material ID which
you can then match with a specfic material from a Multi-Sub Object material or use it when
unwrapping.
Select by ID: Select the polygons that have a certain ID.
Smoothing Groups: Mainly used to create creases or pseudo-discontinuities on low-polygon

models. By giving adjacent polygons a different smoothing group they get a hard angle
between them. See for example here (left image with the center polygons with the same
smoothing group, right one with different smoothing groups):
29
Subdivision Surface
I never looked at this roll-out, until today and came to the conclusion that it's mighty handy. It
basically is a built-in MeshSmooth in your editable poly, so you don't have to add the modifier.
The advantage is that you don't have to add it (which makes your files bigger and more
instable), the disadvantage is that you don't have a quick 'Show End Result' on/off toggle
which you do have would you use the loose modifier. All the options are (almost) the same as
the ones in the MeshSmooth modifier.
Use NURMS Subdivision: Click it to make things happen, basically an on/off switch.
Smooth Result: This acts as an additional non-mesh changing smoother. Basically a built-in
additional smooth modifier.
Display: Iterations: Number of iterations (the higher, the denser and smoother your mesh
becomes, be careful not to set it too high (>4) or else Max has a tendency to crash. This
value can be different from the value used when rendering (see option below). Usually you
30
set this lower, so as to make the viewport still workable but have good smoothing and a nice
dense mesh when rendering.
Display: Smoothness: This is one of those weird options... Tweak it and you'll get a lower
polycount, but you pay the price in the form of an ugly mesh, I don't really see why you would
want use this option.
Render: Iterations: Same as the Display: Iterations, only at the moment you render.
Render: Smoothness: See Display: Smoothness, only at the moment you render.
31
Polygon Modeling 5: Techniques I

Summary
Okay, now that we have covered most of the available tools in each sub-object, it is time to
apply them. Yet we will not apply them to a specific modeling problem just yet, such as
modeling a wheel rim. Why not? Because it's time to understand the interaction between your
Editable Poly and MeshSmooth. This interaction can be be delicate at times and at other
times it works completely logical.
The Essentials and 2 Examples
When you're modeling and you apply meshsmooth to a model you will have noticed that not
your entire mesh gets evenly smoothed (and not because of vertex weights, we will not use
them). This is always because of your mesh density: your mesh density, as I call it, is the all
important thing when polymodeling in combination with MeshSmooth. So what is the mesh
density? Well your model could look exactly the same as mine before MeshSmooting, but
afterwards it could look totally different.
Example 1: Suppose we have a cube (100x100x100 in size), on the left it has 1 segment in
each dimension, in the middle three and on the right it has ten. As anyone can see, and which
is pretty obvious they all have exactly the same shape... they're cubes, so duh! But then I
apply a meshsmooth (MS from now on) with one iteration, the result is shown below each
cube.
32
THIS IS IT, this is what poly modeling is all about. Try to understand what is happening here.
The result is very very clear: the cube with 1x1x1 segments is extremely smoothed (it has
almost become a sphere), the second cube is way less smoothed but then again, the cube
with 10x10x10 segments is still very similar to it's original shape.
The denser your mesh before MS'ing, the less smoothed it becomes.
That's the rule and that is what you will think of all the time when you're poly-modeling. Every
action you perform, you think: hm, what will happen when I MS? Or even better, know
beforehand, and therefor you will think: I do this and this, because when I MS, this will not get
smoothed since I made the mesh dense.
Let's redo the example above, only in a slightly different way.
Example 2: We make the second cube again (100x100x100, with 3x3x3 segments). Only this
time we will want to create the same end result, so after MS'ing as on the cube with 10x10x10
segments. As you will understand it is a lot more efficient if we could do that (MS'ing
33
quadruples your polycount on each iteration). So create it, convert it to an EP, then go to the
VSO and move the middle two rows of vertices closer to the edge (I selected and scaled them
275% from different viewports, it doesn't matter how you do it, as long as you get it done).
So I move all the vertices towards the edges (literally) of the cube, since what I want is a cube
with sharp edges, I don't want a blobby cube, I want a well defined cube only then with nice
smoothed edges. I did it in three steps as you can see on the images, below the 4 is the end
result (with and without meshsmooth), a bit zoomed in. You can see clearly that the shape
hasn't changed at all, it's exactly the same as before. The only thing we did was move some
vertices around to densify our mesh at the places where we want them, which is exactly the
whole idea behind polymodeling. On the final right bottom image the result is shown: a very
cube like shape, only with nice semi-smooth edges. It's a step in the right direction. Moving
vertices around to densify your mesh in desired places is a valid way of getting what you want
out of MS. Yet it means that the vertices that you move away from where they were didn't
define a shape you wanted... otherwise you would loose the shape of your model, but gained
definition at other places. There must be other ways to get where we just got.
34
Chamfering
There is indeed: it's called chamfering. As explained in the previous tutorials, and especially
the one about the ESO, I mentioned chamfering and how important it would become later on.
Well here goes.
Example 3: Create a new cube, 100x100x100 in size with 1x1x1 segments, convert it to an
EP, go to the edge sub-object. Okay we are going to get the same result as before, but this
time we create the smooth edges by chamfering, so you get smooth edges and a denser
mesh at the borders which will in return result in what you want when MS'ing. Select all
edges, click 'Chamfer' (the little window next to it), set Chamfer Amount to 1.0 (default) and
click Ok (not Apply). Congratulations, you have chamfered and therefore your are! ;) Apply a
MS with 2 iterations. We use more iterations because our mesh is so empty before MS'ing
that it gets very little definition would we use 1 iteration (try and see what I mean). Okay, so
our goal is accomplished, but there is still some things to explain.
Do the same as before only set the Chamfering Amount to 5. Both steps and results can be
seen in the next image, and work both as we expected: set the Chamfer Amount smaller and
the end result are very sharp borders / edges, set the Chamfer Amount larger and the end
result is smoother. So the denser your mesh, the sharper the end result.
35
This about wraps it up, concerning the chamfering and densifying your mesh. Try to
understand what I am getting at, since it's the core of poly/MeshSmooth modeling. Hereafter I
will explain some of the techniques I have learned over the years.
Making holes
This won't be an explanation of theory, but more an application, or serie of events to create
nice round holes in an object. The end result isn't very tidy (in that you are left with some
garbage vertices, but it gives a good idea of how to do something).
So create a new box, 100x100x100 with 1x1x1 segments, convert to EP, go to the PSO.
Select a side polygon (1), click 'Tessellate' twice (2) (this as mentioned in Tutorial 4 slices
your polies). Then select the middle 4 polygons (3) and click collapse (4). Now you have a
sort of star like polygon configuration where your original polygon used to be.
36
Go to the VSO, select the middle vertex of the 'star' (1) and chamfer it (to the size you want)
so that you create a new centered circular shaped polygon (2). Go to the PSO, select the new
center polygon (3) and extrude it inwards (4).
That's it, but we want it to look good when MS'ing! So go to the ESO, select on the edges (1)
of your hole and another one on the inside (2), and then press 'Loop' (3). Hopefully it selects
the two 'rings' of edges, in my case it didn't so I had to select them by hand. Then I chamfered
the edges by an amount of 1.0 (4).
Then to make the whole cube not become very blobby and smooth when MS'ing, I also
chamfer the other edges (1) by an amount of 3.0 (2), since I want them to be defined but not
37
as much as the hole edges. Then I apply a MS with 2 iterations (3).
If you want the hole go from one side to the other, then do the exact same procedure on the
other side of the cube, but delete the two center polygons and connect the two newly created
cubes by welding the vertices.
Another way of doing it: select the whole side and hole (1), for instance by using grow after
selecting the center polygon, then shift-rotate it 180 degrees (hold shift and then rotate your
selection which will in effect copy it, press a beforehand to rotate with steps of 5 degrees),
clone to element and move it a bit away (2), deselect it. Then rotate your view and delete the
side polygon after selecting it (3 + 4)
Then turn on snap (s) and set it (by right clicking it) that it snaps to vertices. Select one of the
polygons of the copied hole (1), start clicking grow like a madman until the whole hole is
selected (2), then move it (by using snap) towards the place where you just deleted the
polygon (3).
38
Deselect all, go to the VSO, select all, and then press 'Weld' (the vertex count should be 4
less after the welding), apply the MS again and lo and behold the result. It probably doesn't
work very well, since the original side is still different from the original one, this is because the
chamfering was done after the collapsing.So the best way would be to just copy the entire
side, including edges (1), delete the whole other side and do the same as explained before
(2). More vertices should weld and the result should be better (3).
The last part went a bit quickly and sloppy, but it's just to explain certain handy procedures.
This about wraps it up for this tutorial, more in Tutorial 6: Techniques II.
39
Polygon Modeling 6: Techniques II

Summary
More techniques and procedures will be explained in this tutorial. And also problems you
could encounter when applying the explained techniques.
Creases
Everybody wants to know about creases, especially when car modeling they're essential for a
nice result. Very often you see those blobby creases which make a car look very undefined
and ugly. In the next section I will try and explain way(s) of making them.
The steps will be:
- Edge: Chamfer
- Poly: Extrude
- Edge: Chamfer
Once again create a box, 100x100x100 in size and 1x1x1 segments, convert to EP. Go to the
PSO and select all polygons (1), from a side viewport slice your box in half (2) using Slice
Plane (3), go to the ESO, select the newly created edges (4) (or they should already be in
fact) and chamfer them with an amount of 3.0 (5).
40
You can see the crease approaching! ;) Back to the PSO, select the newly created polygons
(1), extrude inwards (2) (By Group, height -5.0). Then when you go back to the ESO, you
should notice that the edges are still selected (3) (exactly the ones we want), chamfer them
by an amount of 1.0 (4). Then if you want the cube to remain cubish (this is not necessary,
but we'll do it anyway) select (5) and chamfer (6) the other edges with an amount of 2.0.
The important and essential steps are of course: chamfer edge, extrude inwards, chamfer
41
edge, as described beforehand. If you want you could also chamfer the edges on the inside of
the extrusion part. Anyways, apply a MS with iterations 2 and your result should match the
next image:
That is it, there is nothing more to it: chamfer, extrude, chamfer. Remember that, apply it and
it always works. At least is has for me in the last years.
Problems
Off course it's not always as straightforward and easy as with this cube. In fact it hardly ever
is... And then it may well be that you (and me too) will run into problems. These problems
could be (an explanation of how to deal with them follows below):
1. You could have created a crease one way and then want another crease and that crosses
the original one. This problem is best to be avoided.
2. Creating crossing creases (the 1st one is almost the same, but this one is still different).
3. Creases / chamfered edges on bend surfaces: one of the mother of all problems when
polymodeling.
42
Problem 1: Crossing creases
Suppose you have made a couple of creases on your model (like I did on the image below)
and want to make an additional crease along the red line. It looks trivial, but unfortunately it's
not. The second part of the creasing process, the poly extrude will not work as expected.
I will have a go at it, so I select the edges (1) (by using 'Loop'), chamfer (2), since that
shouldn't be any problem. And then select (3) the polies that need to be extruded inwards.
Here lies the problem... we can't just go and extrude! Since then the part where there already
is a crease will be extruded inwards for the second time. There's a way to extrude anyways
now, but it's very, very cumbersome and hard work. It would encompass extruding parts, then
deletion of existing polygons, creating new ones and hoping it all matches up. As follows:
deselect the already creased part (2), extrude inwards (3), select polies that are in the way
43
(4), delete them (5), create new polies to fill up the gaps (6 + 7 + 8). It works, but then you
also have to make the chamfering of the edges match. And believe me all this is very very
tedious work.
So there's a solution, but is there others? Yes, yes there is. It's a trivial one: create all your
creases that cross at the same time! This will save you so incredibly much time. As you can
see in my Porsche WIP progress from image 2 to 3 I created all creases in one big sweep. It
was one 'chamfer' command, one extrude and one chamfer again, because they were almost
all linked together somehow it had to be done all at once.
Problem 2: Crossing Creases II
It's very easy to think that this is not a problem. But I will show you it is! ;) You don't have to
do this yourself, looking at my images and reading what I do will be enough to see what's
going on.
I create a box, 100x100x100 with 2x2x2 segments (1), and create the crossing creases
(select (2), chamfer (3), extrude (4), chamfer (5) and MS with 2 iterations (6)) exactly the way
I did in the previous example, only now with some more edges to which I apply the chamfer,
extrude, chamfer trick.
44
As you can see, where the creases cross each other a big tearing / hole appears. Not what
you could possibly want. The question of course is, before we can solve it: why is this
happening? It is happening because densifying you mesh when chamfering has a direction
and MS has too. MS tends to smooth negatively perpendicular to the average of normals of
the surfaces. So if you would take the normals of two faces, averaged them, and then inverse
it (make it negative), then you get the direction MS would work in. Sounds a bit complicated,
but it's a feeling you will get along the way of learning how to model. In the next image I tried
to explain by drawing some images, how the direction would be.
45
Now that we know the cause of the problem, we can go and do something about it. Naturally
our mesh densifying by chamfering comes to the rescue. Select the four standing edges, see
image below, and chamfer them, so that they create a sort of barrier for the MS.
Hooray! That'll teach 'em! We win, we will always win. :D Modelers: 1, MeshSmooth: 0.
46
Anyways, when MS doesn't do what you want, like it did just now, then densify your mesh in
the places you think the cause of the problem lies and see whether it gives the desired result.
If not then try other places to densify, densifying is always the solution. Always.
Problem 3: Creases / chamfered edges on bend surfaces
This is a problem that is hard to avoid and also quite hard to defeat. In short it's a problem
that you want to densify your mesh somewhere, and it does what you want, but as a side
effect it has effect on your model when you MS which you don't want.
I will use a tube in this example, which is curved and not flat as the cube was, converted to an
EP and some extra segments. I'll first chamfer the edges of the side by selecting two edges
(2) and then Loop (3), chamfer (4) and MS with 2 iterations (5). All expected behaviour: one
side with nice edges, other side very blobby.
All is well, yet now the problem: what if I want a rectangular inset in the side? I'll do it and the
problems will occur automagically.
So I selected the polygons I wanted to extrude inwards (1), extruded them inwards (2),
47
selected the edges (3) and also the edges on the inside (4)(including the standing ones in the
corner!), chamfered them (5) and then applies MS with 2 iterations.
Looks good, but in fact it's fairly hopeless. From this angle it's hard to see the problem, but let
me rotate the camera for you. See this side view:
It's not circular anymore see the right part of the previous image, there's a bulge right now. It
seems minor, but believe me this is something that you see when you use reflective materials
such as a car paint or chrome. Also realize that this problem will always occur when you have
a curved surface and you start extruding inwards and chamfer some edges. It's a problem
you'll be facing non-stop if you're working with cars. There's almost no car where there's no
things that go inside on the car body and those are the places where it goes wrong. With
NURBS this would be a piece of cake, but we're modeling indirectly, so it may look like it's
going okay but the only thing that counts is after you have applied meshsmooth.
48
And now for the solution: the solution isn't one way. There's no default solution. The only
general plan of work is to make your mesh denser around the problem area and form it the
way it would look after you had applies MS. This way you 'fool' MS: it's dense so MS doesn't
have much influence, and it has the right shape so it will remain good. Let's have a go at it.
I sliced a couple of times and moved the vertices around (with a Circle as guidance to see
whether it matched up with it) while using the 'Show end Result' button in the modifier roll-out
to instantly see the result of moving the vertices and slicing. It takes a bit of feeling (which is
hard to explain in this tutorial) but just slice a bit to dense it up, and then reshape it so that is
matches the shape you want again. So once more the densifying is the solution.
As said, there's very little that you cannot solve by making your mesh denser. And that is no
surprise, since what you de facto do is play to be MS, you are actually smoothing things out.
So try to make your mesh as undense as possible and let MS do the work, yet if that doesn't
49
give the wanted result, then you densify it at place where it fucks up. If that gives wrong result
again, try to tweak it and fool MS.
50
Polygon Modeling 7: Car Body

Summary
The modeling of the body of a car can almost always be sliced up into different parts, literally
and figuratively speaking. Litereally because most of the time the front and rear bumpers, the
entire 'cockpit', mirrors etc. are all loose parts. Figuratively because when you model it, it can
be done in steps: first a rough simple (but rightly shaped!) basis and then the refinements,
such as chamfering edges and making the creases.
Blueprint Setup
Setting up blueprints is a 5 minute job, I have explained the way I do it earlier in the tutorial
about how to setup blueprints. Only this time I chose a different car, since almost the whole
world seems to model it: the Audi TT. Another reason to model this car, is because it's one of
the easiest car to model. After we have done that (1), it's time to jump in.
PS: When I would be modeling a car I would try to find as much reference photo's as
possible. As I have stated before: I don't use the blueprints that much when I model my car,
only for the rough / global shape of the car I use them, photo's are way better reference since
they show all details. So get Google out and try to get your hands on wallpapers and personal
photo's of the car you try to model. Sometimes it can be very hard to find any (for instance for
the 993 GT2 and the Alfa Romeo SZ I modeled), but it's essential if you want to end up with a
good representation of the original. I am not talking about a couple of photo's... usually I
would try to find around 50-100 wallpaper sized (800x600 or larger) photo's. In this tutorial I
won't go into modeling all details, since only the concept is important, not the modeling itself.
51
Finding a Place to Start
I have the feeling that many people have a (healthy!) fear or starting to model a car. 'Where
do I start?', is a logical question. And there is no definate answer. Anywhere would be the
best I can think of. It just doesn't matter, yet the way you approach the modeling process
does.
So I am just going to start at the front, and model the hood. What I will think of from the first
second I start is that the whole car is symmetric (apart from the tank cap), so I am not going
to model both sides, but only one half. So I'll start with a single plane, but the way I create it,
is by creating a box from the side view (1) and then deleting 5 out of 6 polygons (2). This way
side of the polygon is exactly in the middle and I can use the symmetry modifier without any
troubles (3).
52
So now that we have made the basis, the hard work is up! I start from the top view and shift-
drag some edges, slice once and move the vertices to match up with the blueprints (but only
from the top view up until now!) to create the entire hood (1). After doing that I realize that the
most front part of the hood goes downward and that I need more polygons there (what you
see is smaller than it really is, because of the orthogonal top view) and so I model the front
grill too (2). I always model this with the symmetry modifier visible, so I can see the result
immediately.
Now is as good as any time to not only make it match the top but also from the side (top and
53
side view are most important, I think, front and rear view I don't use very often). So we move
the vertices so they match. Both the most inner row and outer row of vertices are easy to
place (3), since the blueprints have a line there (the crease of the hood) in the top and side
view. The line of vertices in between should be placed by feeling...
PS: Notice how the blueprints I used do not match up. This is a common flaw of almost any
blueprint, and there is nothing you can do about it, just live with it and correct your modeling
so that it looks best according to your own feeling. This is also why I say that you shouldn't
trust on blueprints alone.
54
After making the two other rows match too, you shold have something like (1). Looking good if
you ask me. Realize that the lines on blueprints usually correspond to creases and / or sharp
edges. When you want to model them, you need edges there (to create the creases as
explained in a previous tutorial).
55
Going Sideways
Well we can't go any more further to the front, all is done there (well roughly), so we should
go towards the side. I selected the entire row of edges (2), shift-dragged them out (3) (once
again) and editted the vertices I just created to match up again from top and side (4). Looks
fairly good, but I made a small mistake: in the last part I didn't realize I was looking from
above again and so I didn't have enough vertices to match the shape of the body, this I will
set right by cutting an extra row and place the newly created vertices correctly (5).
Since I won't be modeling the entire car (why would I? Then the tutorial would cover how to
model an Audi TT ;), I will now start with some detailing (which I would normally only do until I
have finished the rough model of the entire body.
The edges on the hood are a good place to start. First we select the edges (1), chamfer them
(2), move one lower than the other (3) then chamfer the new edges again only now smaller so
56
they will become more sharp (4). Add a MS modifier with 2 iteration to see whether the result
is satisfactory.
I now will model the headlights and the crease of the hood. Since the headlights look to be
perfectly smooth with the body I will just make creases around it (and not delete them and
later on try to remodel them back in). And with one sweep I will also create the crease of the
hood. So I select the edges (1) I want to crease and apply the technique explained in an
earlier tutorial (2). With a 2 iterations MS modifier (3).
57
Would I continue from here, then I would be in deep trouble, making the creases continue
onto the wheel arches is undoable (as explained). So realize that this is just for tutorial
purposes that I start detailing already.
Last thing I will do is make the grill. I didn't take into account modeling it (I have no vertices
there), so I will slice where needed (1), then extrude inwards (2), delete the not need
polygons (3) and finally chamfer some edges (4).
58
So in the end it looks like this:
59
Modeling the rest of the body is just more of the same. Only you would first model the entire
car, then make the creases and hard edges and then do all the final tweaking.
60
Polygon Modeling 8: Car Rims

Summary
Modeling car rims is actually fairly easy and straightforward. This is because there always is
some sort of symmetry, always. Whether it be point-, line- or suface-symmetry, there's always
one or a combination. This will reduce modeling difficulty and effort tremendously, since Max
has, since version 5, the Symmetry modifier which is very handy.
The Rim and Symmetry Analysis
Well before jumping in I think we should first pick a rim which we'll try and recreate.
After browsing to one of my favorite rim makers, Toora, I picked one out which is as good an
example as any:
Do realize we will probably not be making all those nuts and bolts, maybe if I have some time
at the end.
Anyways: as you can see, there is a five-fold symmetry here. 5 Center nuts and 5 arms. So
61
what we are going to do is modeling one arm, one nut hole, one bolt / nut and 1/5th of the
outer rim and that's it. Nothing more. No way that I am going model the whole thing. We are
going to let the Symmetry modifier do all the work for us.
Okay I've drawn some of the symmetry options that are available in the next image:
Immediately after I drew this, I realized: there's not only a 5 fold symmetry (the red and green
part), but even a 10-fold one (the blue part)! So all we have to do is model a part of 36
degrees (360 for a total circle, divided by 10) and symmetry it so that it makes up the entire
rim. I have been very lucky choosing this one as you can see... Would this rim have 5 arms
and 6 nut holes then it would be way, way, way more work, it's then when you start cursing
Max and wished you used Rhino. But that is not the case and it hardly is, as said, there is
almost symmetry on every rim.
Start: Outline and Basic Shape
I talk a lot about basic shapes with which I mean that I always model something that roughly
looks the way I want it to look and then start detailing and refining. I see many modelers work
in a different way and that is cool, but I feel that my method is especially good when you want
62
to learn and also works well for finding errors or misformed parts of your model more easily.
So first a rough outline which we will use as guidance. In the front view I create Circle with a
radius of 100 (1)(remember that in 3D modeling everything is about relative sizes, not
absolute sizes, so 100 is absolutely arbitrary, but handy to work with) and a cilinder with
radius 100, both at position 0, 0, 0 (2). In the Cilinder set the height to something like -30,
height segments 1, sides 20.
Now for a bit of math. I said that we will only model one part of 36 degrees. So click the 'Slice
On' radio button of the cilinder. In math the 0 degrees line is at the middle right (1). So for
example slicing from 0 to 90 gives result (2). We want a total of 36 degrees and I want it at
the center top. So we slice from 90 + 36/2 = 108 to 90-36/2 = 72 degrees and hooray, it works
as expected (3). Hm, but now that I think of it: having it at the center top isn't handy, because
what we will be modeling is only half an arm of the rim which I want at the center top at the
end. So we shift the slicing 18 degrees. Slice from 90 to 54 (4) This may look like nonsense,
but it'll become clear very soon.
63
Let's have a test run at how the symmetry modifier will help us out. Add one (1), and if
everything is correct it should be at the right place and angle already. Add another one (it will
be at the same position and angle as the first one) and rotate it (in the sub-object of the
modifier, yes modifiers also have an SO, usually for editting their Gizmo's) 36 degrees. This is
best done by clicking the 'Absolute Mode Transform Type In' which then becomes 'Offset
Mode Transform Type In' button at the bottom center of your screen. This button changes the
amount you give the rotate / move / scale from absolute to relative. When you set it to Offset,
it adds it (the amount you want) to the current value, which is what we want now. This way
you save yourself calculation time. :) So fill in 36 in the Z-direction (rotate!!, not move) (3). As
you can see we are not yet at 10 parts, so add another one, rotate this one 72 degrees (twice
the 36 of last time) (4). Now comes a part where you might think it'll go wrong: we are going to
64
add another Symmetry modifier. This will result in 16 parts in total, while we only wanted 10.
But the modifier has a Threshold, so it will symmetry it over the ones you created in the first
couple modifiers and weld them away. They won't be seen. So add, rotate 144 degrees (5). In
the last part of the next image I show how your modifier list should look right now (6) And also
the Transform Type-In thingie.
So far for the boring part, now it's time to take advantage of what we just created.
Copy the original circle and give it a radius of about 35 (1). This will work as a guidance when
we model the center part. Then copy it again, set the radius to 10, move it to make the
guidance for the nut holes (2). Now I drew a line (with Initial Type 'Smooth' and Drag Type
'Smooth') to help as guidance when modeling the arm (3).
65
Modeling
All outlines are set, so we can now start modeling. Back to the original cilinder, click the 'Show
End Result On/Off Toggle' so that you only see the single part (1). Set the Sides to 3 and Cap
Segments to 3 (2).
Now we want to have an EP (we want to model!!), but we have a cilinder, that's tricky since
we have all those modifiers and don't have the option 'Convert to Editable Poly' without
collapsing the modifiers along with it. But Max is to the rescue, it has a nice copy/paste
function for modifiers. So select all the symmetry modifiers by Ctrl clicking them and then right
click, Cut. You modifiers are now cut and placed in memory just as it would work with text.
Then convert your cilinder to an EDITABLE MESH (!!) and then to an EP (somehow Max
fucks up if you convert directly to an EP), right click it in the modifier panel (where you just
clicked Cut) and click Paste.
Before we go into the VSO, go to the PSO and delete the sides of your sliced cilinder (3 + 4),
they will give troubles when using the symmetry modifier, and the rear part too (5). So now we
have a total flat part of the original cilinder. Into the Vertex Sub-Object we go now, we want to
shape it (6)!
66
So modeling now, but first set the interpolation of all circles created to 'Adaptive', this way
they're perfectly smooth and better guidance.
In the VSO start moving the vertices around so that match all your guidance lines (1) and
delete what you don't need (2). Click the 'Toggle End Result' button to see whether we are
getting somewhere (3). I think we are! This is starting to look like something and all with very
little effort. You could work with the TER button On all the time, this is a very good way to see
the result of what you're doing immediately and all over the whole rim (especially with an MS
modifier of course).
67
So do a little editting of vertices and slicing to make it fit better (1 + 2) and we're done for the
moment. What we have now is an almost perfect 2D representation of what we want in the
end. From now on things will go faster: you have the knowledge know of how to setup the
symmetry and then all you have to do is model one bit perfectly and you're set. I changed the
shape of the outlines since I noticed the rim was getting too fat compared to the original,
editted some more vertices (2).
The next thing we will do is make the nut hole. Go to the PSO and slice the shape so that it
matches your guiding circle (1). Cut some additional lines to prevent errors later on (2) (keep
your mesh tidy!). Select the polygons that make up the hole and extrude inwards (3), delete
the polygon in the middle that you just created, you don't want that, once again because of
the symmetry modifier.
68
Now in the ESO select some edges that make up the side of the arm (1), to give it thickness,
by shift-dragging them (2), and the end result should look someting like this (3):
All to do now is chamfering edges to define the shape better and 3D the whole lot to make it
more look like the original. I am not going through all those steps it would be a repitition of
previous tutorials and also a lead-by-the-hand type of tutorial which I don't like. Tutorials
should be about concepts not step by step explanations. This is what I got after 5 minutes of
editting and chamfering etc.:
69
Either you call it a day here, or you make one final addition: the outer rim cilinder. This you
could make separate from the part we just made or make it a whole. We'll try the latter.
I didn't change much at the top end of the arm and so you shouldn't have too much trouble
folliwing the next steps. Create a tube with radius 100 and 103, same slice as before (90, 54)
or at least make it fit the rim (1) and then delete all polies except for the inner ones to be left
with a small strip (2) then delete the last polygon (3), attach the rim to this tube part and weld
some vertices to make it form a whole, then select some edges and shift-drag them (4 + 5).
70
Some more chamfering and shift-dragging where need and you should be all set. Then you
should (or could :P) have something like this (1) and I can think you can do the rest yourself
now! It may not resemble the original perfectly, but you only have to change one single part...
71
72
And here an example of a little more worked out rim, with a semi-nice texture.
73
Polygon Modeling 9: Car Tyres

Summary
Modeling car tyres is a polygon eating exercise (you end up with quite a few in the end), but
definately worth it if you want to make a realistic looking car: using bump maps for threads is
a bit too obvious.
Reference and Interpretation
So what tyre will we make. Of course there's only one brand that comes to mind: Pirelli. Once
again, any tyre would do... this is all about the concept.
After having a quick look at the Pirelli site (through Google most you find is Pirelli Calendar
images... ah what a shame! :) I chose the Pirelli PZero Direzionale tyre:
More than this you don't need as reference. When you look at this tyre you can see the
simplicity with which we are dealing: it's a repeating pattern! And this one is symmetric
through the middle too so it's even more easy (on my Alfa Romeo SZ I have Pirelli PZero
Asymmetrico's... they're as their name suggests asymmetric).
74
So what part of this tyre are we going to model? Well only the part highlighted in the next
image, that will be all! This part we will then mirror (with the Symmetry modifier) and then
Array Rotate it (this will be explained in a short while). Having a quick look at the reference
image I can guess that there will be around 30 parts (in the Array Rotate) needed to complete
the whole tyre. 30 parts for 360 degrees is 12 degrees per part (remember!).
Starting
It's quite obvious that a tyre resembles a tube and therefor it would be logical to start
modeling from a sliced tube. Yet we won't. The way I do it, is make a tube as a basis (a sort of
background / underlying model) and over it model the thread (only one part) which will be
'glued' on.
So first the underlying tube. Nothing fancy, I just created a tube (1) (100 Radius 1, 90 Radius
2, 70 height, 1 Height Segment) and posistioned it at (0, 0, 0) which will be the basis. After
that I applied an MS modifier with 2 iterations (2). This is not good so far, it's way too smooth
and thin. So I chamfered some edges and sliced some stuff using Slice Plane (3) to make it
more defined and retain its shape after MS (4). All this is just rough modeling, no measuring
or thinking but all by feeling.
75
The Thread
So now that the basis is there, we will start modeling the thread. First some guidance
modeling is need to create virtual boundaries and then the actual modeling can be done.
The reference will be a sliced cilinder. So create a new cilinder with a radius of 110 (the rest
of the settings, other than the slicing, aren't important) and set the slice from 96 to 84 (12
degrees centered around the middle top, see the previous tutorial). So the part we will model
is roughly the size of this cilinder (1 + 2).
76
We will start out the thread by making a simple box, which will serve as a rough start, give it a
height of 35 (half the height of the cilinder) and convert it to an EP. Now move it to the logical
position from where we can start modeling (1 + 2).
Delete the innert part of the box (1) and add a Symmetry modifier (2). Now move and rotate
the Symmetry gizmo (so in the Symmetry sub-object) so that it mirrors the box, (probably
move -35 in the y-direction and rotate 90 degrees in the z-direction from the top view) (3).
77
Very good. Now we're almost done with laying the basis. Only two more steps to go: adjusting
the Pivot Point and the Array Rotating.
Go out of the Symmetry SO and to the side view (1) go to the Hierarchy tab (this is next to the
Modifier Tab, at the right top of your screen). Then you should see the Adjust Pivot options.
What we will do here is change the Pivot point of the box, so that it can rotate around a
different center, namely the center of the tube. Click 'Affect Pivot Only' (2) and move the pivot
to (0, 0, 0) (3). Click the 'Affect Pivot Only' button again to go out of the editting.
78
Only thing left is the Array Rotating. So either click the Array button, use your short-cut to
Array or go to 'Tools' > 'Array...'. Array can make copies of your selected object(s), either
normal copies, reference copies or instance copies. We will use Instance copies. Instances
are object that have a two way information stream between itself and the object from which it
was created. So suppose I make one object, instance it and then edit the new one, then the
old will change too and vice versa. Very useful.
So set the 1D Count to 30 (remember?), and the z rotation to 12 (360 degrees / 30 parts),
Type of Object: Instance and click 'OK'. Excellent, most of the hard work is done now. We
have 30 parts which are all identical and remain identical (they're Instanced!), so we once
more only have to edit one part and the rest will come along.
79
I think the rest of the tutorial isn't as important. I will go through it more quickly, since all you
have to do now is just edit your original box to make it look like one part of the Pirelli tyre.
Only screenshots of what I did to make my thread look the way it did will be shown.
80
81
Almost done with the modeling now, mainly smoothing it (by hand, no MS) and making all the
parts fit each other is to be done. After I have done that, I unhide all parts (I usually work at
one part alone, only the matching I do with three parts (the top center and the one left and
right of it)) and test render it:
More detail can be added if you want (I made it very sloppy now), chamfered edges do make
it look better but you will get a high poly count. It's one or the other. :)
The following tyre (of my latest model, the Alfa Romeo ES30 / SZ) is made with the same
technique, the rim with the technique of the previous tutorial. Only now with some custom
made textures added. I am quite proud of it and think it looks quite fancy.
82
I hope you liked these tutorials and learned something! Happy modeling!
83
Polygon Modeling A: The Process of Modeling a Car

Summary
In this tutorial I will try to describe and explain, as far as possible, the process of modeling a
car. Do mind: the process, so an overview of the steps I would undertake, not the details of
how to model a rear spoiler, or front bumper... Be prepared for a lot of text and very little
images!
Once more I would to emphasize: this is how I see the process of modeling a car, my
personal view on it. I do not claim it to be how it should be done, it's just a look in the kitchen
for all you people out there modeling cars too. Take from it what you like, laugh at what you
think is stupid, but learn from it or at least enjoy it. :)
Where to begin?
A good question, and obviously the only question when you want to start modeling a car, and
as always there isn't really a definite answer to where. But I can tell you how I would do it, or
what I would suggest you to do.
Throughout the first bit of the tutorial I will assume you want to model a car and do not have to
model a car. There is a massive difference between modeling something because you want
to (personal) and because you have to (work).
So you want to model a car: good, very good. A question like: can I model a car could arise at
first, but it's not really important. You can always try and you have to try if you ever want to
become a better modeler. So a car it is, yet what car? When modeling something for yourself
you could think this is not really important either, but I think it is. I tend to believe that when
you model a car and you have some sort of connection with it, that the end result is better and
84
more easily reached too. One is more motivated when one models something that brings out
a smile on your face, or gets you going because it makes such a nice noise, looks so nice etc.
The whole love for the car in real life (would you model a car that actually exists of course) is
a great drive for anyone to make it as good as one can. Suppose I would (for some reason... I
wouldn't know why, but just suppose) start to model an Opel Kadett from 1985. I can tell you
now that I won't finish it: how can I be motivated to model such a totally crap looking piece of
****? I couldn't, maybe you can, but even so, when I would model a Ferrari F355 (happens to
be the car of my dreams), I would give it EVERYTHING I got, put in the hours it takes to make
it as good as I can make it.
Modeling a car isn't a one day thing, unless you're some magician. I don't think many can
model a high-poly car in a day. I know I can't, and I have been told to be an extremely fast
modeler. The modeling alone can take anywhere between 10 and 60 hours or so, depending
on the level of detail and how fast you work. But you see the problem here: you seriously
have to be motivated to start working on a project like that... 60 hours is comparable to 1.5
full-time working weeks. That's quite a lot.
I am not telling you this to scare you off, but only to make you face reality or know what you're
getting into when you start modeling cars. The flipside of the whole story is of course the
immense satisfaction of the whole modeling process, the test renders, the final result. To me
at least: everytime I see some of the renders that I consider my best work, I get this feeling of
pride, of joy and that only gets stronger when you realize how much time is has cost me to
learn 3D Max, PhotoShop and to actually build the whole thing. Maybe we're all mad making
cars in 3D, but I for one couldn't care less. I love it.
Reference Images
After choosing what car to model, the thing I always do is start looking for reference images.
Not a couple, but every single last image you can find on the highest resolution possible. If
you know the car you're modeling exists somewhere in your area: try to get a digital camera
85
(if you have one: great, otherwise borrow one) and make as many pictures of the car as
possible (car dealerships, private owners (ask them first, some people are a bit freaky about
taking pictures of their car)) and especially of the details such as head lights, tail lights,
doorhandles, windscreenwipers, interior (if you're up for the modeling of that too), logo's /
decals etc. etc. The details are what makes a great car and it's always difficult to see them on
wallpaper images which tend to show the whole car.
The second part of the reference image quest: blueprints / 3-view drawings. If you can find
these for the car (or whatever you're modeling) you want to model, then that helps
immensely. There are a couple big sites on the web (Suurland, SMCars,
Blueprints.Onnovanbraam.com) and together they have around 20000 blueprints in total, so if
you're lucky enough they have the blueprint you're looking for. Newer cars (last year) and
classic cars (70's and older) tend to be more rare, but most 80's and 90's cars are
represented, especially the special ones, like italian exotics, M3's, Skylines, Supra's etc..
If you can't find blueprints: that sucks, I personally can't model very well without them. It can
be done absolutely, but it's obviously a couple of magnitudes more difficult than when you
model from reference. I personally always check if a blueprint exists for the car I want to
model... if it doesn't exists (which hasn't happened yet, but then again I haven't modeled that
many cars) then I wouldn't start.
86
The modeling process
The actual modeling process I usually see in a couple of stages:
1. Blueprint setup
2. Body
3. Rims and Tyres
4. Exterior details
5. Interior
----------------------------------
6 + 7. Texturing + Lighting
This is fairly straightforward, apart from things that are normal to me, but I don't think many
people do: I split the modeling and the texturing/lighting part completely. I model the entire car
(EVERYTHING) without applying any texture to it, apart from a grey clay material. That's right
no texturing at all until the whole car is done. No lighting until the whole car is done.
But... WHY? That's what you're probably thinking. And with a reason, because it's not very
normal I think. Then again, I am not normal, and I do have a reason to do it this way. Keeping
different aspects separated helps me in focussing better: I hate switching between different
aspects of a project. I have that with everything I do, whether it be modeling, or making a
website: when you focus on one particular aspect of a project you can go up into better, focus
completely on a single process, a single thread of the whole wire. Switching back and forth
slows anyone down significantly, plus you have to get you're mind switched to every time from
polygons to raytraced materials, blurred reflections, opacity, back to polygons, edges, back to
specular materials, blend maps etc. etc. It'll slow anyone down and will drain energy.
Another reason to do it the way I do, is because projects can come to a halt halfway more
easily because it seems like you achieved quite a deal when you render a body with full
textures and some nice lighting, but if you have no windows, no wheels, no rims, no exterior
details, what good does it do? I think one tends to get complascent with what one has
87
achieved then, and it's more easy to quit when you have a feeling you're done. Which is
obviously not the case: there is loads of work to be done. GET BACK IN THERE SOLDIER! I
know it it's easy to fold though, especially when you do something for yourself, in your spare
time. Motivating yourself into working that long and/or hard on something with no payment,
other than the satisfaction and pride when it's done, isn't easy.
1. Blueprint Setup
The setting up of the blueprints is a 10 minute job at most. You just get the blueprint you
found (if you have of course), crop the views in PhotoShop and put them on planes in 3D
Max. See also this tutorial I made a while ago.
2. Body
The body of the car: obviously the most important part of the whole car. Use the blueprints as
a guidance to modeling it, but never rely on them completely. It is almost never so that the
blueprints are very accurate let alone match up in all the views (top, side, front, rear), so use
them as a rough outline but double check what you model using the reference photo's you
found. Especially check the position of characteristic features of the body: compare their
relative sizes and positions compared to other parts of the body or other parts of the car.
3. Rims and Tyres
Ah yes, the wheels. Wheels make or brake a car. Believe me, I have seen a lot of cars in my
life (seen seen, as in observed them) and the rims on a car are very important. Yes they can
be too big (anyone putting 19" rims under a BMW E36 M3 for instance is nuts! ;) ), they can
be too small, they can be wrong and good in a dozen ways. But knowing this: spend time on
them when you model them, try to make them accurate. To me they are almost always one of
the most enjoyable part of the whole process, because they're like little gems of their own.
88
Also because they have a great variety of texture in all the separate parts (brake disc, brake
calliper, the rubber of the wheel itself, bolts, nuts etc.). Of course to each his own, I can't
judge taste, but what I do know is that well modeled cars, have well modeled rims and real
modeled wheels (including a modeled profile!). For the modeling of the rims see a technique
here and for the modeling of the tyres see a technique here.
4. Exterior details
The details are what makes a modeled car look good. Model as much as you want, and
realize that the more the better. Not more than there are on the real thing, but as much as you
want from those available: door handles, the grill mesh, locks, badges, decals, antennae, little
bolts, small hood clips. It can be tedious, but once more I will state it: the more, the merrier. If
you can't be bothered anymore but there are still things to be modeled: think about the
immense satisfaction of seeing all those details when you render out the entire car at
6000x4000... I know I have done on numerous occasions and I was very happy with all the
details I put in there. On large scales the large surfaces tend not to be as important; they are
of course, but only as background, or filler for the details. Your eyes, or the ones from
whomever looks at the image will trace the image flowing from detail to detail and zoom out to
see the whole picture. Details make you want to touch it, make the illusion of it not being real,
but looking as if it were real greater, much much greater. It's all in the details, in the subtleties.
Beauty lies within subtlety. Chinese wisdom? Maybe, it's what I believe in anyway.
5. Interior
Modeling the interior is difficult, very difficult. Many organic shapes usually, like the
dashboard, the seats and others. How detailed you want your interior to be depends on how
tinted your windows are (and yes, that's a non-subtle hint) or how little motivation you have
left when you are at this stage. Modeling it gives another big leap in realism, but I do have to
admit that I never went full-out on it so far, I haven't modeled any interior as well as I could or
89
as well as I should have. You can get very far with modeling the inside of the doors, the seats,
the dash and the steering wheel at a medium level of likeness to it's original. Maybe also
because I model the cars for their exterior mostly, and make renders of the exterior that I
haven't put as much effort in as I have on modeling car bodies. Only the SZ I made has
almost everything in the interior, steering wheel, handbrake, gear leaver, seats, rear seats,
floor, all gauges in the dashboard (with needles and background textures, yes, thank you very
much ;).
So depending on whether you are going to make renders of the interior, or are making a
convertible/cabrio, then you have to model it. And I can understand it is great fun, but usually
you already have done so much on the rest of the car that you either want to continue to the
texturing / rendering part, rather than modeling every single detail of the interior (and believe
me, there are more than you can think of, in any car).
6. + 7. Texturing + Lighting
I don't have much to say about these two aspects of the process. Some notes though: decide
beforehand which renderer you want to use, switching back and forth between scanline /
Brazil / Mental Ray / VRay is horrible half-way: you have to re-do most if not all of your
textures. Maybe as an exercise (if you're up for it) you could texture and render the whole car
in one renderer and then try to get the same or better result in another, but I wouldn't do that
for 'fun', so to say. Obviously you learn a lot doing something like that, you will learn the finer
details of the differences in materials and lighting settings... not unimportant at all, but I think I
know in which state of mind I am at this stage of the process: I want to get it done and as
good as possible, not mess about switching renderers and doing tedious material overhauls.
Again these two steps are vital to a good end result (even though I can enjoy a clay render of
a well modeled car as much as any). Don't trust the modern day renderers to do all the work
for you. (fake) GI isn't everything, try to get a nice lighting setup going, with or without a studio
environment with cards, or even better a real environment such as a parking garage...
90
Oh, how I dream on of my next project... :)
All in all: that's pretty much all I have to say on this topic. Try to work methodically, make lists
of what you want to do (list of objects for instance), to get an overview of yourself of what
needs to be done. You usually know for yourself how much work something is, but it's always
nice to be able to tick off that object which you were afraid of to model...
Happy modeling!
91
Introduction to Splines
Welcome to this introductory tutorial about splines! Splines are a powerful modeling tool,
which has it's pro's and con's, but for me until now, I have found mainly pro's. Modeling with
splines is very different from box/polymodeling, but in the end the results will be much alike.
The Basics
I'll start right away with showing you the absolute basics, that will come back all the time when
you start using splines. So fire up 3D Max, go to the Create Panel > Shapes > Splines > Line.
Click the Line button. You now get some options in the panel: Rendering, Interpolation,
Creation Method and Keyboard Entry. If everything is ok, then Creation Method should be
rolled out. Set Inital and Drag type both to corner.
Now go the front viewport and draw a hexagon (a six sided 'circle'), by clicking where you
want each point, and connect the last (the 6th with the first again) and when the pop-up with
'Close Spline' comes, click Yes. Your result should look something like so, it doesn't matter if
yours isn't exactly the same as long as it resembles it and has six points:
92
Alrighty, now go to the modify panel. You see it's called Line01 and that the object is of the
type 'Line', and has a small 'add sign' before it, this means that it has multiple sub-objects
(just as with editable polies, they have Vertex, Edge, Face, Poly and Element sub-object...).
Click the 'add sign' before Line and see that Splines have the following sub-objects: Vertex,
Segment and Spline.
Click Vertex to go into the vertex sub-object, you see a small button lights up in the modify
panel (under Selection), this button is the exact same thing as the one you clicked just now,
these let you go in to specific sub-objects. You also see that the vertices of your hexagon are
shown, with one square one. That's your initial vertex, the one where you started drawing the
hexagon. Move some vertices around and you'll see that it behaves normally, the lines that go
from vertex to vertex move along, so nothing spectacular.
Click the second sub-object (Segment) and you see that your vertices changed from small
crosses to different small crosses! ;) You can still recognize the initial vertex, is has again a
small square around it. The segment sub-object and one step 'larger' than vertex. It can be
compared with the edge sub-object in editable polies. When vertices have dimension 0,
edges/segments have dimension 1. Select a segment and move it around, you see that again
it behaves natural.
The last sub-object, is Spline, go in it (nothing changes in comparison to the Segment sub-
object (maybe your selection is gone), again the same diagonal crosses for vertices) and
select something, you see that the entire hexagon is selected. So this is the largest sub-
object, entire splines can be selected, nothing smaller. You'll see later on, why this can be
handy. For now go back to the vertex sub-object.
93
PS: Tip: To go in and out of the sub-object of an object (Spline or Editable Mesh/Poly) use
Ctrl + B, to cycle through the different sub-objects, use Insert.
Now pan down the modifier panel to the 'Geometry' roll-out. The options here are all essential
to spline modeling, stuff like Create Line, Attach, Weld and Refine are used ALL the time.
We'll start with Create Line.
Click Create Line (make sure you're in the vertex sub-object), and now you can draw new
lines, but that are part of the original spline, so you still have one
object. So for instace, draw a rectangle inside the hexagon. It may
be that it doesn't become entirely square when you click 'Close Spline > Yes' again, but we
won't worry about that until later. But what if you don't want to close the spline all the time?
What if you wanted to make a U-shaped line? Ah, yes, that's easy, just right-click when you
want to end drawing a (new) line. Try to make one, this time inside the rectangle.
94
Right click to go out of the 'Create Line' mode, go into Spline sub-object, and select the
hexagon, then the rectangle and then the U-shape, you see how it works? You have created
three new splines, but they are part of the same object (Line01). Now remove the last two
mentioned, by selecting the rectangle and the U-shape and hitting delete.
Back in the vertex sub-object, click Refine and move over a segment of your hexagon, you
see that the cursor changes from the original arrow to a cross with two lines above it, one with
an extra dot. Refine adds vertices to existing segments! So try to break the top and bottom
segment (in VERTEX sub-object!!) into two parts, by adding a vertex in each of them.
This is very handy for adding detail to an existing spline model and something you'll be useing
intensely once grabbed by the power of spline modeling. Right click to go out of the Refine-
mode.
I'll explain something new one, that you may not have thought was important, but it is! If you
already undertsand how it works, then you can skip this part: Snap. Snap let's the cursor snap
to something you want, for instance vertices or grid points. This may be handy when you want
to draw/make something that certain measures, or when you want something to line up with
an existing object. Press 's' to turn snap on (you should see a button turn on and off
somewhere in the UI (user interface) when you press it). As default it snaps
to grid points, but we don't want that, so right click the snap button to bring
up the Grid and Snap Settings, click the Vertex radio button to turn it ON and
click the Grid Points radio button, to turn that one OFF. Move your cursor over the vertices
95
and see how it 'snaps' to those points and a small cross turns up when you are around one.
Okay, now click 'Create Line' again, and draw a line between the two vertices you just
created, with snap turned on. If snapped isn't turned on, then you won't draw a line that goes
from one vertex to another (which is required later on), but a line that is in the vincinity of
those two. And in three dimensions it is absolutely imperative to use snap, or otherwise it's
impossible to draw a line from one vertex to another... Believe me.
Okay, now refine the newly created segment and draw a line from the most left vertex to the
middle of the new segment (where you just made a new vertex) and on to the most right
vertex. We are taking this up another notch! :)
As said, snap is important. The new vertices that you draw will come at the EXACT same spot
as the ones you are trying to snap them to. This does not mean that you can't distinguish
them anymore, you can still select them both, but they need to be at the same place to form a
whole.
Vertex Types
Up until now, we have only used Corner vertices (we set that in the complete beginning,
remember?). But this way only very square objects could be created. So naturally there's
more, there are three other types:
Smooth: Non-adjustable vertices that create smooth continuous curves. The curvature at a
smooth vertex is determined by the spacing of adjacent vertices.
96
Corner: Non-adjustable vertices that create sharp corners.
Bezier: Adjustable vertex with locked continuous tangent handles that create a smooth curve.
The curvature at the vertex is set by the direction and magnitude of the tangent handles.
Bezier Corner: Adjustable vertex with discontinuous tangent handles that create a sharp
corner. The curvature of the segment as it leaves the corner is set by the direction and
magnitude of the tangent handles.
One of the things that probably isn't clear when you read this are 'tangent handles'. These are
helpers that you can move around to set the angle of the corner around a vertex. I'll explain
some more, with the spline we have made until now.
Go into the vertex sub-object, select the upper left and upper right vertices, hover over one
with your mouse until it becomes a cross (the one for moving or selecting, it doesn't matter, as
long as you're on one of them) then right click to bring up the Quad Menu. In the upper left
corner you see the four different Vertex Types, click 'Smooth'. You can see how your spline
has changed, it's smooth around the two vertices you are selecting! Move them down a bit to
smooth out the line some more, so that it goes through the middle vertex without a
discontinuity of the derivative.
Now select the lower left and lower right vertices, right click one of them, set Vertex Type to
Bezier. Now try to move the bezier handles (the orange lines with squares at the end), by
moving the squares and see how your splines reacts to it, it tries to keep it smooth no matter
97
what you do: in the vertex the derivative is continuous. This is exactly what it said in the
description: 'continuous tangent'.
See if you can make something similar to the shape I made below, it's not hard, but if you can
make it match this, then you could make it match (almost) any shape you want.
Select the most right vertices (there are two, but you can only see one, because they're
snapped) and set the type to 'Bezier Corner'. This time the bezier handles are there again, but
the shape hasn't changed and it's not continuous either. Once more exactly as in the
description of this type of vertex: 'discontinuous tangents'. Move each of the three handles
around and see what it does with the shape of the spline. This type of vertex I use the most:
you can create smooth surfaces that end in a 'square' form, you are in control most with this
type of vertex, you could even make it continuous if you wanted it, but also square, this one
has it all. It takes effort to get exactly how you want it, but at the end this ones the most
powerfull. Yet still, sometimes you don't even need it, for instance when making very smooth
objects...
98
Now that you know the basics and the vertex types, try to fool around, make some trial spline
cages, see how it reacts when you make all the vertices of the Smooth or Bezier type, try to
understand why it behaves the way it does, because if you do, then and only will you be in
control to make whatever you can imagine!
I haven't mentioned the surface modifier yet, but I'll shortly go into what it does and how to
use it. The surface modifier is your tool to make something out of you spline cage. Once you
have the cage, then apply the modifier and voila, one surface! But there are rules that you
have to attend to, to make the surface modifier work the way it should: 'Make sure that the
Spline vertices form valid three-sided or four-sided polygons. Vertices on splines that cross
one another should be coincident'. Look in the help file that ships with 3D Max under 'surface
modifier', you'll find all the information you need there. Again, fool around with different
settings for the Surface Modifier, try different Patch Topology Steps, maybe add an additional
MeshSmooth... Trial and error is the way.
That's it, this is all I have to say for the time being, maybe you should try to apply the newly
acquired knowledge in the next tutorial: Spline Car Modeling.
99
Spline Car Modeling

PLEASE MAKE THE RHINO HEAD TUTORIAL THAT SHIPS WITH MAX BEFORE
STARTING!! That is the tutorial about making a head of a rhinoceros (under F1), and not the
tutorial about a head in Rhino3D! :)
Or else this tutorial will go way too fast.
Then check these links:
http://www.comet-cartoons.com/3ddocs/facialmodeling/modeling.html
http://www.cyberkreations.com/kreationsedge/tut/splines/cage.htm
Then you can proceed with this tutorial :)
First take the blueprints, crop them in Photoshop and place them on planes (for a tutorial see:
100
here)
Now start drawing splines along the edges of the car in the front view (or side, just the view
where you see the side of the car). I assume you know how to draw splines and how use the
sub-object qua vertices (bezier, bezier corner, smooth and corner). If not then look in the Help
file that ships with Max. Make sure that you always form triangles or preferably squares,
never make a 'face' with five vertices, because it will not work then (the surface modifier that
is).
Next is placing all the vertices of your spline at the right place in the last and third dimension.
Do that by using the top view: align all the vertices at the right spot. From here on it is pretty
straight forward: do this for all the different parts (front / rear bumper, roof, hood etc.) Make
use of the great tool called snap (see Help file for more info). When you did that, add a
surface modifier to your spline(s) and you have a perfect mesh, try to play around with the
segments number and see what happens. An extra meshsmooth always helps too for a final
101
touch.
The rest of the images is the progress of the scene I used (one of my favorite cars: The
Lancia Delta Integrale Evoluzione).
102
103
104
And here the final result:
105
TEXTURING
106
Map Channels
Step 1. - Reset 3D Studio Max
Step 2. - Go to the Front View
Step 3. - Go to the Create Panel and make a tube. Make Radius1 = 100, Radius2 = 80
and the Height = 60. This is the basic shape that we'll use for this tutorial. It's supposed to be
a tire: we'll make a tire (without the alloy) which has some profile AND some letters on the
side like this (don't pay any attention to the low ammount of segments, because that can be
altered any time. And the inner side is also textured, but that does not matter either since one
would usually make an alloy inner tire too...):
Step 4. - Before you even begin to make this tire, you should make the two textures needed
to apply to the tire. I would suggest something like the following two, since I already used
them and they work very nice for the wanted effect: double texture on this tire. Save these two
pictures as tire-side.JPG and tire-track.JPG.
107
Step 5. - Select the tube and convert it to an editable mesh. Go to the polygonal sub-object
level and select the faces/polygons on which you want the tire-track to be applied. When
selected go to the material editor and make a new material. Click the maps roll-out and click
the bump button, select bitmap from the pop-up and select tire-track.JPG. Set V-tiling to 25
(we need the track to be repeated a lot of times because it has to go all around the tire) and
108
the W-angle to 90 degrees (on the right). Click Show map in Viewport and then Go to
Parent . Set bump amount to -999. And to complete this material set the Diffuse and
ambient color to (45, 45, 45) and specular level to 45.
Step 6. - Go back to one of the viewports and press Alt, e, i which means Select Invert (or
make a shortcut like me: Ctrl + Shift + I). So now we have all the polygons selected which do
NOT need the tire-track on them, BUT need letters on them, because they are the side of the
tire. So now go back to the material editor and again make a new material. Again click the
maps roll-out, and now click the diffuse button, select bitmap and then tire-side.JPG. Again
click Go to Parent and set the specular level to 45. Again click Show map in Viewport. Now
you should have two materials that should look like this in the Material Editor:
Step 7. - Close the material editor. Go back to normal mode of the tube (= out of the
Polygonal Sub-Object level). Apply an UVW map. Click cilindrical. That should do the job
nicely, you would say, and I used to think so too. But what happens when we make a test
render? IT SUCKS (left image)! Ok, but what do we do then? A planar map! Yes! Again we
make a test render... IT SUCKS (right image)!!!! WHOAAAAAA! :) We need both but we can
only have one at a time to work... But here is where the map channels come in!
109
Step 8. - How do they work? Go to the material editor. Select the first material (the tire track
110
one) and click the bump map again so that you are in the bitmap screen. On the right in the
coordinates roll-out you see Map Channel. Set it to 1 (it should already be but make sure it
is). Go to Parent and now go to the other material (the one for the side of the tire). Click the
diffuse button and set the material channel of this bitmap to 2.
Step 9. - Now close the ME. Remove all UVW maps off the tube. Apply a new one: set
Mapping to cilindrical and make sure the Map Channel of the UVW map is 1. Now apply a
new UVW map, set Mapping to planar and set the Map Channel to 2. Make the Length and th
Width of the second UVW 225 (because the tire-side.JPG is a bit too wide and high). Make
another test render.... YES! It's good:
Wrap-up - Well that's about it. I explained how the idea behind Map Channels work by
making a tire with two materials that needed different UVW maps. An important thing to
remember is that the Map Channel of ALL sub-materials have to be set to the right number:
suppose you have a material that has a diffuse, bump and opacity map then all map channels
within each map have to be set to the right value. I hope that you understand that this is only
a simple example of how Map Channels work and that you can apply it in a more complicated
111
way with more UVW maps and more materials.
Applying this technique is also possible without eating up all your material slots (I know that
you can have more than 24 that are available... but still), by using Multi/Sub Object materials.
This is very simple: everything works the same only now, you would have to set the right
Material ID's of the selected polygons. It does make it a lot more easy to comprehend,
especially with a lot of different materials on one object.
Help - Here's a piece of text from the Help file that ships with Max. It's explains the definition
of a Map Channel:
A map channel associates a map with an object's mapping coordinates.
When you turn on Generate Mapping Coordinates for an object, they use map channel 1. You
can assign new map channels with new mapping coordinates by applying a UVW Map
modifier to the object. Map channel values can range from 1 to 99.
112
For NURBS surface sub-objects, you can assign a map channel without applying UVW Map.
The surface sub-object has a different set of mapping coordinates for each map channel you
use.
A map's map channel value identifies which of an object's mapping coordinates to use.
Different map channels allow maps for the same object to use different coordinates. For
example, you might use one channel for diffuse mapping and a different one for bump
mapping. Map channels also let different maps use different coordinates within a compound
material, a compositor map, or a multi/sub-object material.
Different map channels can have different U and V tiling values, different U and V offsets, and
so on. In the UVW Map modifier, you can also set different map channels to have different
mapping types (planar, cylindrical, spherical, and so on).
If you apply a map that uses a certain map channel to an object that has no mapping
coordinates for that channel, the map doesn't appear on the object. When you render, a
Missing Map Coordinates dialog appears to warn you of the problem. The dialog lists the map
channel and the object name.
113
Multi/Sub-Object
Step 1. - Reset 3D Studio Max
Step 2. - Go to the Front View
Step 3. - Go to the Create Panel and make a cilinder. It does not matter what
dimensions, whatever your like. Go to the Modify Panel and set the height segments to 3.
Now that you have done that you should have something that looks like this:
Step 4. - Now go to the Modify Panel again and apply an Edit Mesh modifier to the cilinder.
Choose Sub-Object => Face (or Selection => face). Now select the bottom one third of the
cilinder. Scroll to the bottom of the Edit Mesh modifier panel and set Material ID to 1. Now
deselect the bottom section and select the middle third and set Material ID to 2, and for the
top third to 3.
114
Step 5. - Deselect the Sub-Object mode and open the Material Editor (m). Instead of
Standard choose Multi/Sub Object. At the pop-up choose 'discard old material' because we
are going to make new ones.
Step 6. - Set number of materials to 3. Now that you have three little sphere's in front of you
with the 3 Material ID's next to them. You can make the three materials any way you like: I
made a Dent, a Brick and a Planet material.
Step 7. - Do not exit the Material Editor yet, because it still has to be applied to the cilinder.
So click on the cilinder without closing the ME and apply the material: the 'sphere-to-cube'
button. Now every Material ID has it's own material.
Step 8. - The final thing to do is apply an UVW-map to the cilinder: because you have applied
an Edit Mesh to the cilinder, Max doesn't understand in which way he has to apply the
material. An UVW-map simply solves that problem. Go to Modify and apply UVW-map.
Click cilindrical as Mapping and at alignment click Fit.
Step 9. - After this tidying up, you can render it by pressing Shift + Q and you should have an
image resembling to this:
115
This of course is a very simple tutorial, but still: this way of adding more than one material to a
single object can be very handy and useful. So no time wasted! :-p
116
Composite Maps
This tutorial will describe how to use composite maps in 3D Studio Max. Composite maps are
a different way of putting decals/logo's on your object, other than having to unwrap it and
make large bitmaps in PhotoShop.
First step is to create the object on which we'll put the composite map, and for simplicity I
made a single plane of 200 by 200 units (square).
With the plane selected, go to the Material Editor and in the diffuse slot add a new map: a
Composite Map.
117
When you have done that, you see something like the next image: you can set the number of
maps by clicking the 'Set Number' button, but for the time being we'll just use two maps,
because we want to put two logo's (the Pirelli and the Brembo logo) on the plane.
To do this we need two different maps for each logo: one for the color and one mask. The
mask image will 'blend' the base material color with the diffuse color of your color image. It'll
become clear in a sec, if you don't exactly understand what I am saying.
Blend/Mask and Diffuse map for the Pirelli Logo.
118
Blend and diffuse map for the Brembo Logo.
Combined diffuse and blend maps: I made the maps perfectly square so that they're easier to
work with later on. When we'll apply UVW Map modifiers to our object, I don't want to think
about what measurements/dimensions/relative x / y size it should have, so I make it square
and that way it's easy to set the x and y size of the UVW Map since they're the same, they
have a ratio of 1:1.
119
Map 1: Pirelli Logo.
So on with Map 1 of the Composite Material. Click the 'None' of Map 1, it'll bring up a pop-up
with different type of maps and select Mask.
The Mask Map has again two slots: one Map (we'll use the Pirelli color bitmap for this) and
one Mask map. The Mask map blends (as said) between the base color and the Map.
120
Click the Map slot, select Bitmap from the list, and select the Pirelli diffuse/color map
wherever you have stored it. Note that in the screenshot it's still the small logo, not the square
one, I changed them only afterwards and forgot to remake a screenshot...
Uncheck the U and V tiling radio buttons: we only want to put a single logo on our plane, not a
121
repeating pattern.
After you did all that, click the 'Go to Parent' button to go up one level in your material.
Repeat the previous process for the Mask map, only now select the mask bitmap, in stead of
the Color bitmap.
So the Composite Map now has a Mask Map in Map 1, and the Map map (pardon the jargon
122
here, but they decided to call the diffuse color map in the Mask map 'Map') has the color
bitmap in it, and the Mask has the blend/mask bitmap in it.
Next up is to repeat the previous steps for Brembo logo which will go into Map 2 of the
Composite map.
The only difference is to change the Map Channel to 2 for both the diffuse and the blend
bitmap. Change the name of the Mask maps to something like 'Brembo Logo' and 'Pirelli
Logo' so that you can find them easily.
That pretty much wraps up the work in the Material Editor! The nasty part is over, now comes
the fun part: playing with the logo's on your object.
Adding UVW Maps
Add two UVW Map modifiers (Planar ones), rename them to match the material names
('Brembo Logo', 'Pirelli Logo'), by right clicking them in the list in the Modify panel. Change the
123
Map Channel of the Brembo UVW Map to 2, so as to match Map Channel you set just now in
the Material Editor.
Go back once more to the Material Editor and set the Show diffuse Pirelli map in viewport, by
going into the Composite Map then the Pirelli Logo Mask and the Map (diffuse) slot and click
the 'Show Map in Viewport'...
... and if everything went okay, it should show up in your viewport as below.
124
And now for the magic: go to sub-object of Pirelli UVW Map, by clicking on it in the Modify
panel (so the Pirelli Logo UVW map) so that it becomes yellow to indicate you're in the sub-
object. Now you can move your UVW Map around and the beauty is: your logo moves along
with it!.
125
I moved the UVW Map a bit downwards just for examplary purposes.
126
Go to sub-object of Brembo UVW Map and move it too, only a bit upwards.
127
This is basically the whole key to Composite Maps: you can move your logo's/decals, all
around the object you are mapping, without doing complicated unwrapping, or making large
UVW Maps in PhotoShop which kills memory wise when rendering.
So time for a test render! Below you can see it, and there are some notes about what is what.
Also not the horizontal lines, which are due to me making sloppy Masks which had some
compression in them and therefore weren't perfectly white/black.
128
Hierarcy/Order of Maps in Composite Maps
Now it's time to dig a little deeper into how Composite Maps work. You could for instance ask
yourself what would happen if the two maps were put on top of each other by moving the
UVW Maps both the center of the plane. Well, why not do it, but first I changed the
diffuse/color Map of the Pirelli Mask Map to a black version so that you can see the difference
between the Brembo and Pirelli logo (they were both red).
Left is the original order of the maps in the Composite Map (Pirelli in Map 1, Brembo in Map
2) and on the right is them reversed (Brembo in Map 1, Pirelli in Map 2).
129
You can swap the materials by dragging one onto the other and selecting 'Swap' as the Copy
method and click OK.
So this means: the higher the Map number in the Composite Map list, the more on top it is. So
if you would want to make a complicated textured car with lolads of vynils for instance, you
would work 'back-to'front': everything what is in the background goes in Map 1, then what is
on top of that in Map 2 and what if above that in Map 3 etc. etc.
130
Scaling of UVW Maps
So far we have only worked with the default size of the UVW Map (i.e. being as large as our
object), but what if we want the logo to be smaller than what it is now? You could make it
smaller on the bitmap, but that's a lot of work (you have to do the Diffuse and the Mask
bitmap) and besides, it's a irreversible process, you're resaving your bitmaps. All in all, that's
not the way: you can just scale your UVW Map. Go into the Sub-Object of the Pirelli Logo
UVW Map and make it a bit smaller by scaling it, using the 'Select and Non-Uniform Scale'.
This is one of the beauties of Composite Maps: you can move and scale your logo's and
decals at will all over your object without having to worry about unwrapping, changes to your
mesh, the size of the logo etc.
Adding additional logo's
If you want to add more logo's, you have to add more Maps to the Composite Map, as
follows:
- Set number of Composite Maps, 1 higher than it is now.
- Copy old map (of the logo you want to duplicate or if you want to make a new to save some
131
time making the Mask etc.) to the new slot.
- Change map channels of diffuse and blend bitmap.
- Add a new UVW Map to the object and change the UVW Map map channel to the one you
just set in the Material Editor.
I think it's good practice to match your Map Channels of your bitmaps and UVW Maps to the
Map number of your Composite Map. That way you avoid consfusion and errors.
If one would use Composite maps compared to UVW Unwrapping, there are pros and cons:
pros:
- Fast placement of each decal/logo and you can instantly put it where you want it.
- No huge bitmaps, that take a lot of memory while rendering.
- No UVW Unwrapping, which can be a really big pain in the rear and which you have to redo
every time you change your model.
cons
- Your material can become quite extensive (having a lot of Maps) when you have loads of
132
decals, like on a race car.
- Your model would need a lot of UVW Maps when you have a lot of Maps, which in turn is
memory intensive.
133
Car Mapping
This tutorial is a combination of two tutorials I made earlier and applied to a real-life example.
You can find the two tutorials here: Map Channels and Multi/Sub Object.
In this tutorial I'll try to give an impression of how I apply textures to my cars. It's very simple
once you get it, really.
As subject I'll use my latest car, the Citroen C3 'Custom'. I could have picked any car, in fact I
might as well have chosen a box, as long as the object you use has a couple of polygons.
What is all comes down to is this:
Create your car, plane, whatever, select parts of your mesh and assign a unique material ID
to that part, repeat this step for all parts, then apply an UVW map for every ID, create a multi-
sub-object material and voila, you're done. Easy as that.
Here's a screenshot of the car we'll be working with:
134
Okay, now you should think by yourself how you are going to break it apart, so what pieces
you'll want. I want the front bumper to be one part, the side, the other side, the roof and the
rear bumper (I detached the doors, hood and rear door, because I wanted them to be able to
move separately, they're all just planarly mapped since I only put one texture on those parts).
So I select the front bumper, and set the material ID to 1, select one side, set that ID to 2, the
other side: ID = 3, the roof: ID = 4 and the rear bumper: ID = 5.
135
136
So far so good, and very easy, isn't it? Now we'll add all the UVW maps. I am only going to
use planar UVW maps, so they need to be exactly the size of your selection, so use 'Select by
ID', apply a new UVW modifier and click 'Fit', or fill in the size of the map manually, works just
as good. So for the front bumper your UVW map would be like this (see the Gizmo?):
And the sides (I use one UVW map for the sides...), roof and rear bumper respectively:
137
Now give each UVW map a unique Map Channel, you could make it match the material ID
(this would be very conveniant and will avoid problems) but it doesn't really matter as long as
they are all unique! You will use this map channel when you make your materials for each
part of the car.
Your modifer tree should look something like the following image. Mine has some additional
modifiers (a symmetry, because I only modeled half the car, a MeshSmooth for good looks,
and the Editable Mesh to set some of the ID's), so don't really pay attention to them. I may
138
have more or less UVW map modifiers, since I only have four parts I want to texture... This
doesn't matter at all, it's the way it works that matters here.
Okay, great, so now everything is set up, we can start making the material. I use one of the
MultiSub-Object type: here you can make a material for each ID and assign this overall
material to the object and if you had your ID's and Map Channels right it would look as you
wanted to. So go to an emtpy slot in the material editor, click 'Standard' and select Multi/Sub-
Object from the list. Discard old material. Set number to 4 (we have four different ID's!). Now
you can start making all your materials, just click the first one from the list, the one for the
front bumper and do what you like with it. But keep the following in mind:
Suppose my front bumper would have ID 1, and the UVW map that maps the front bumper
has a map channel of 1, then the map channel of textures in your material should ALSO have
a map channel of 1, since you want those textures to mapped correctly for the front bumper. If
you would set the map channel to match a different UVW map then it wouldn't fit the fron
bumper corectly. In the following screen shot, you can see how I did it: my UVW map for the
front bumper had a Map Channel of 4 (what you see here is the diffuse bitmap):
139
Keep the previous at the back of your head at all times! So the material you with ID one is
given to all the polygons that have ID 1 and is mapped by the UVW map that corresponds to
the map channel in your textures. Well from here on it's more of the same for all the different
parts and in the end you have something like this (I used the same material for the hood,
doors and rear door...):
I recommend using 'Show Map in Viewport' a lot, since it immediately shows whether you
have done it correctly or not.
Here is an example of how the side texture looks like (drawn with Illustrator on a screenshot
of the wireframe of the side):
140
In the next 4 images you can see the 4 stages of the process for the side view: first you see
the side view (at the time I was modeling this car I just took a screenshot in stead of using
Texporter to do the unwrapping), then with the UVW Map, then the actual texture map that I
made in Illustrator, and the final Shadow Mapped viewport with the 'Show Map in Viewport'
button ticked.
141
142
And the final result:
143
144
OTHER
145
Blueprint Setup
Well pretty essential in this tutorial is naturally having actual blueprints... I picked the one that
I just (re-)made an hour ago, the ones for a Ferrari Enzo.
Now crop all the different views in PhotoShop so that they fit exactly in the image (see below).
I have made a small error with the mirrors (the Enzo has two differently sized ones, but I used
one for the front and one for the rear, and naturally picked the two different ones...:), so the
front and rear are not exactly as wide. Doesn't really matter, just that you know.
146
Alright: now save these images (or the ones from your blueprints) with some easy names, like
bp-front.jpg, bp-side.jpg etc. This will come in handy when we have to pick them later on, bp-
01.jpg doesn't really say much...
147
But before we go on do this: in the customize menu select 'Units Setup':
And set it to 'Generic Units', meters don't really mean much when your 3D'ing, it's all about
relative sizes, not absolute, as seen below.
Now in explorer (or with something like ACDsee) try to find out what resolution all the different
parts have, if the blueprints are good then the top would have to be as wide as the side, and
148
the front and rear should be as high as the side and finally the front and rear should be as
wide as each other. In this case you can see (image below), that the front and rear are not as
wide, this is because of the mirror-issue. But the rest matches up very nicely, if I may say so.
Now create a plane with the size of the resolution that the side image has, so in this case
791x195, set the length and width segments to 1.
Fill in the following values in the absolute position at the bottom of the UI (User Interface). I
used 195/2, because the pivot point is at the middle of the plane, so the z position of the
plane is half as high as the height of the plane: 195 / 2.
149
Et voila: the plane is positioned perfectly:
Now clone this plane (Edit > Clone) and rotate it 90 degrees.
150
Change the size of the plane to match the front view:
Before we continue we first set some viewport options right, so that it won't cause problems
later on. For instance when you work with planes and don't put 2-sided on, it doesn't show in
certain viewports, very frustrating. So right click on 'Perspective' and click 'Configure...'.
151
Click the '2-sided' and 'Default lighting' radio buttons and set 'Apply to' to 'All viewports'.
152
Now your perspective view should look something like this (if not then it might be that you
have to press 'F3', this toggles between Wire-frame and Smooth and Highlights mode.
Create the remaining two planes yourself, with the right sizes (= matching their image view
sizes) and positions. At the end of that you have something like this (I don't put the front and
rear at the same x-position since they would interfere with each other):
153
Rename all your planes to something so that you can find them easily, stuff like BP: Front
always works easily, I tend to just name them after the view they represent, so Front, Rear,
Side, Top.
That was the setup of the planes, now we are going to apply the right materials to the right
planes. So open the Material Editor and in the first slot click the small square for the diffuse
material...
154
...from the pop-up click 'Bitmap'...
...and select the side view from the blueprint, in my case bp-setup-02-side (from this tutorial):
155
Select the plane called 'Side' and apply the material, also click the checkered square to 'Show
map in Viewport':
Repeat this step for all planes and all their corresponding bitmaps.
Now right-click perspective again and select 'Texture Correction', this makes sure they are
shown rightly in the viewport, sometimes they don't really do what they're supposed to, the
texture correction solves this.
156
What you should have is this (almost finished now! :):
157
The final step is to change the setting to optimize the quality of the texture in your viewports,
so go to Customize > Preferences... > Viewports > Configure Driver
The only things you should change are the 'Apperance Preferences'. I have them like this, this
may differ for other video cards (I have a NVidia GeForce 4 Ti4200 64Mb... [AT THE TIME
158
OF WRITING]).
That's it, we're done. You may want to rotate the planes if they don't match with your viewport
names (as I have here, the side is in the Front viewports), but that doesn't really change
much, they're all orthogonal...
159
160
Depth of Field Using a ZDepth pass

This tutorial will describe how to use the Lens Blur filter in PhotoShop on your renders, using
a ZDepth pass.
A ZDepth pass is a black and white image which gives the depth in the z-direction (viewing
vector, z component). I will describe in short how to make this pass and then what to do with
it.
First of all: you need an actual scene where you want to apply the Depth of Field on. In my
case I chose my latest project, which was an artists impression of the center console of the
Bugatti Veyron 16/4 (the car).
On the next two screens you can see my camera (you need a camera to do all this! always
use camera's to fix certain views on your model) from the top view and the view through the
camera.
161
Select the Camera, and in the Modify panel go to the parameters and then to the environment
ranges. Click the checkbox 'Show' and change the Near and Far ranges to what you want as
a 'start' and 'end' point of your focal area. Basically you want as much contrast as possible
throughout your whole image, and thus try to tweak the Near and Far ranges so that they ...
162
My settings for this scene were 3000 and 4500 roughly. Remember those values, we will be
using them later!
Go to the Material Editor, make a new Standard material, set the self-illumination to 100, click
the diffuse slot...
163
...select Falloff...
...set the Type to Distance Blend...
...and the Near and Far ranges so they match your ranges in your Camera.
Apply this material to all your visible objects in your scene, delete all your lights, set the
renderer to Scanline and press shift + q or F9 (quick render).
If everything is correct, you should have something like the following:
164
Save it to a new file (something like Dash 015 Z Cam04.jpg for instance).
Here's my original render onto which we will apply the DoF trick.
Open up PhotoShop, and load your original render, copy it to a new file (never use the
original render when you edit it), and press ctrl + j to duplicate it once more:
165
Add a new Layer Mask by going to Layer > Layer Mask > Reveal All
In the Layers Panel of PhotoShop (the one below your Navigator Panel usually) you now see
the new layer mask.
Click the mask in the Panel so that it gets those little brackets around it and then click the
Channels Tab.
166
Open your ZDepth pass image, copy it (select all, ctrl + c), close it, and paste it into the empty
channel (so not into the R, G, B ones!) but in the one we just added.
Go back to the Layer Tab and you can see your mask there...
...and you see it in action too; our image has a reddish hue to it! Oh no!
167
Not to worry though. Now for the magic: go to Filters > Blur > Lens Blur...
168
set the Depth Map to Layer Mask, if it isn't already. Make sure Preview is ticked too.
Play with the settings until the result looks how you want it.
The most important settings are:
Blur focal distance: this determines where your focal point it, so where it is sharp
169
Iris > Radius: the amount of bluriness
All the other ones are not as important, but are great for more tweaking and optimizing.
When you're done, click OK.
What the hell? It doesn't look at all like what we just saw in the preview! CURSES!
It's the Layer Mask that does the damage here; it makes all the white areas in the ZDepth
pass more opaque than the black area's. It acts as an opacity map!
170
So get rid of it, since we already used it for the Lens Blurring.
DELETE! GO AWAY!
Now we're talking; this is like what we saw in the Lens Blur preview window.
171
It looks pretty good, but there are some problem area's, such as the one I circled here. The
blurring isn't very smart and thus overlapping area's don't really smooth out. In this case the
CD player was properly modeled and thus was empty inside, the ZDepth was larger there
(scroll up to see it in the ZDepth pass).
The solution? Manual blurring, using the Blur Tool (press R in PhotoShop). Set the feather tip
to something appropriate and fix what needs to be fixed without messing up the nice effect
created by Lens Blur.
172
Voila!
And the original render (straight from 3D Max, low VRay settings, don't tackle me on the
aliasing please. :)), for completeness.
Some other renders of the same scene, with different camera's and varying amounts of DoF:
173
174
175
176
Making of... the Shelby Mustang GT500 Wheel

Summary
I was browsing Dieselstation.com the other day to look for some reference photo's of silver
cars (I am working on a project for a car manufacturer at the moment), when I came upon the
following photo of a wheel of a Shelby Mustang GT500, the 2005 GT500 that is, not the '67
one. I liked this photo a lot, thought it was beautiful and thus a couple of days later decided to
model it for myself. I didn't start out with the idea of making an exact copy, but more my own
interpretation of it: it would obviously look a lot like it, but I didn't intend it to be exactly the
same. So in the end my wheel and tyre don't look like the one on the photo lighting and
texture wise (my rims are more metal-like, whereas the one on the photo is more silver paint-
like), but it's still a nice picture I think.
What follows is a sort of step by step description of what I have done, without going into all
the details of the texturing and lighting bit, because it would become too much if I did that.
And besides: I don't like describing every single setting in a material. There's no point to it,
especially since this is not specifically about making a silver paint texture for instance.
All the screenshots tend to be rather large at full-size, but that's because I work at a
1650x1080 screen.
177
Modeling
I started out with making some reference splines: first a 10 pointed star...
178
... then some additional circles for each fase of the rim: the center bolts, the center cap, the
center part which goes inwards and the flat bit outside of that.
179
Then I drew some reference lines for the 'forks'; they're not rally accurate, but I already knew I
wouldn't be making an exact copy of the reference photo anyway.
180
The real modeling bit started here: I created a cilinder which had the exact size of the entire
center bit, gave it an additional segment and deleted all the thickness out of it, so that all I had
was what you see in the screenshot. Just a flat bunch of polygons shaped as a piece of pie.
Added a couple of Symmetry modifiers, a technique explained in an earlier tutorial: Polygon

Modeling 8. Car Rims.
181
Made a tube, which had the same slice as the first part I created, and that will be the outer
part of the rim.
182
Added a couple of Connects in the edge sub-object. And extruded some more to make a start
of the 'fork'.
Connected the center part to the outer part, with the fork.
183
More refinements to the single part using Connect; the screenshot shows it with all the
symmetry modifiers on. This whole bit is still completely flat, so I still need to edit it in the third
(the depth) dimension.
184
And that I started doing here. It's all by feeling: I didn't really measure anything, or actually
tried to make it look like the original, it's just a rough equivalent.
So some extruding of edges and moving of vertices later...
185
... and with all symmetry modifiers.
186
I liked the shape so far, and thus I started chamfering all the edges that needed to be sharp
after they had a MeshSmooth modifier applied to it.
187
Applied the MeshSmooth modifier. This is starting to look the way it should, I think!
Created a tube that will function as a base for the tyre.
188
Added all the text (always Text with a Bevel modifier and then a Bend modifier too) on the
wheel. Since one would see it in so much detail, I didn't want to use a bitmap, but actually
model everything.
189
Here's the tyre base with all the text on it and the 'GoodYear' and 'Eagle F1' texts (which I
made in Illustrator and imported).
190
Now with everything unhidden and a camera angle that matches the photo reference better.
Close-up of the wheel with added thread.
191
Lighting, texturing and rendering
So now that all the modeling is done (I always model everything first, then start lighting and
texturing), we can start rendering. Just for giggles sake, I made a clay test render. I use Brazil
R/S for almost everything I do, so in this case I had access to the Render Pass Control, and
selected Off-White Plaster (80% grey) to be able to spot any modeling errors and general
surface properties.
Looks pretty good, but obviously it has a lot of flaws:
- The 'E' of GoodYear has some double faces (that causes this type of rendering artifacts)
- The Eagle F1 and GoodYear don't bend with the base of the tyre, they bend along with the
shape, but not with the 'bulginess'.
- The Eagle F1 and GoodYear are too small, they should be stretced.
- There are no rubber pointy, cilinder thingies, the ones that are always on brand new tyres.
192
- Some of the texts are too 'high'.
- No valve to add air.
All these things were fixed (but I somehow screwed up the camera angle on this one, no idea
why):
A couple of Rubber tests: first one was a tyre material I had...
193
...second one was a test with more reflectiveness....
194
...third one with some additional lights...
195
...and a combination of two and three...
196
...and how it would look on the text and on the thread.
197
Test of the materials for the valve.
198
First test of the rim material and a center Shelby logo.
199
Second test of the rim material, different center Shelby logo, with bump map, and more
contrast in the metal, different inner black paint material.
200
It took me a while to realize that I had made the 5 holes for the nuts in the wrong spot. So I
rotated them 36 degrees into place. Also again a different black paint material for the inner
bits.
201
And with everything unhidden. The metal is too flaky, the outer rim has some weird (and I
mean very weird) UVW mapping. Too little contrast in the rubber too.
202
I want more contrast in the image, as in the reference photo. And thus I added a lot more
light, both illumination wise as reflection wise.
203
See this setup, with two reflection cards.
204
Test render of this setup for the rubber alone. I wanted more contrast from the front of the
image to the back, and this result is already a lot more like it.
And with everything unhidden.
205
More tweaking.
206
Added some more detail on the tyre (the ovals at the right top) and made all the text more
smooth.
207
After editting the last render in PhotoShop:
208
209

3ds Max - Tutorials - 20070606 - Onnovanbraam

Uploaded by

Copyright:

Available Formats

3ds Max - Tutorials - 20070606 - Onnovanbraam

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

3ds Max - Tutorials - 20070606 - Onnovanbraam

Uploaded by

Copyright:

Available Formats

Onno van Braam - www.onnovanbraam.

com - Tutorial Collection 2007

3Ds Max Tutorial

Onno van Braam

Copyright © 2007 Onno van Braam, EVOlution WebDevelopment

Warning and Disclaimer

Polygon Modeling 1: Introduction

- EP: Editable Poly

- VSO: Vertex Sub-Object

- ESO: Edge Sub-Object

- PSO: Polygon Sub-Object

- MS / MS'ing: MeshSmooth / MeshSmoothing

Alrighty, enough of the philosphical / psychological mumbling.

Additionaly ESO is very handy for creating new surfaces.

Polygon Modeling 2: Vertex Sub-Object

be available in the Edge sub-object (and be very very useful).

Deactivate the Soft Selection again.

They are all independent.

to do. Believe me.

Polygon Modeling 3: Edge Sub-Object

Split: A sort of detaching, utterly useless.

Polygon Modeling 4: Polygon Sub-Object

Slice: Performs the slicing as described above.

Select by ID: Select the polygons that have a certain ID.

Smoothing Groups: Mainly used to create creases or pseudo-discontinuities on low-polygon

Polygon Modeling 5: Techniques I

The Essentials and 2 Examples

Let's redo the example above, only in a slightly different way.

as much as the hole edges. Then I apply a MS with 2 iterations (3).

Polygon Modeling 6: Techniques II

The steps will be:

Problem 1: Crossing creases

Problem 2: Crossing Creases II

Problem 3: Creases / chamfered edges on bend surfaces

Polygon Modeling 7: Car Body

Finding a Place to Start

So in the end it looks like this:

Polygon Modeling 8: Car Rims

The Rim and Symmetry Analysis

Start: Outline and Basic Shape

Polygon Modeling 9: Car Tyres

Reference and Interpretation

Polygon Modeling A: The Process of Modeling a Car

The modeling process

The actual modeling process I usually see in a couple of stages:

3. Rims and Tyres

3. Rims and Tyres

Oh, how I dream on of my next project... :)

Corner: Non-adjustable vertices that create sharp corners.

Spline Car Modeling

Or else this tutorial will go way too fast.

Then check these links:

Then you can proceed with this tutorial :)

And here the final result:

Step 2. - Go to the Front View

way with more UVW maps and more materials.

A map channel associates a map with an object's mapping coordinates.

Step 2. - Go to the Front View

Blend/Mask and Diffuse map for the Pirelli Logo.