Back in may 09 Mr. Sameer Sahasrabudhe (Project Manager at IITB*, Center of Excellence) contacted me for his work that requires help with Blender. I landed at Mumbai on 5th July. Upon arrival, I got his call, he asked me to come over to his house to discuss few things with his students Chirag and Nitin.Continue reading “IIT Bombay – blender walkthrough”
Unlike my previous tutorial on modeling the Toyota Celica car this article is not a step by step modeling tutorial for a beginner. What I have tried to tackle in this article is some useful approaches towards modeling the car body in blender. Although it looks and feel like a walk though I have tried to tackle some common problems while modeling a car in Blender which I hope will be useful to another car modeler in blenderverse.
While it has been emphasized often in many car modeling tutorials, the need to gather or ‘research’ various data including photographs of cars and its various parts to help get the modeler and ready reference during modeling.
Much before I start modeling I try to remember the shape and unique features of the car body while going through the references and if possible the actual car, this helps me to consciously identify the possible problems in the mesh while I model, of course having the reference images ready helps to solve any doubts that I may have during modeling also.
Needless to say, they are the very basis of accuracy that you want from your model. more accurate they are the more possibility of you being able to making near to the accurate model. I said near to accurate, because Blender does no offers measurement in real life units, even if you convert them to Blender units somehow there will always be a possibility of ambiguities in calculations. However if you are close they will be hardly noticeable.
While many users prefer to have the blueprints setup in a typical box-like fashion. I find them blocking my way while I model. So it’s mostly a personal preference. Choose whatever you feel comfortable with. We have carried setting up blueprints in a box style in one of our earlier issues, you can consult it if you like to use blueprints that way. I feel comfortable with using Blenders ability to show images in the viewports. For the viable scope of this article, I am assuming you know how to do that already.
There are various shape and sizes of blueprints available on web so first step is to edit them in an image editor like GIMP or your preferred application. We are using the blueprints of Nissan350z that can be found on the site http://www.the-blueprints.com.
There is often a problem that most blenderheads have to deal often and those are
- Different proportions in different views and
- Nonalignment of blueprints in different views.
Proportion problem occurs as most users just cut the blueprints without taking in consideration that Blender view-port somehow distorts (long back Pafuriaz had written an explanation for it) the proportions depending upon the resolution.
Step 1 Fixing the blueprint proportions problem
Make a single square document in your image editor with the maximum width being that of the blueprint. We choose square as we are giving Blender all images(views) of same size thus making sure that all are displayed in the viewport in the same proportion.
First copy all top, side and another view in separate layers in the image editor. While we are in the image editor we can do the required transformations of different parts of the blueprint so that they can be represented incorrect view ports in Blender.
I always choose to have the top view of blueprint in the vertical position as it compliments both front and back view nicely in the viewports. So if your top view is horizontal you might want to make it vertical by rotating it.
Now create a new layer on top and draw a line of single pixel in the middle of the document. Taking this as reference you can move the top, front, and the back view shots in the vertical middle. Now again draw two reference lines one above the car and the other at the bottom of tires. This will give you the reference for fixing the position of a side, front and back. If you did this step nicely you can have the four different reference setup like in the images below.
|Illustration 1: Back view with guides.|
|Illustration 2: Top view with guides.|
Step 2 Alignment problem in view-ports.
If you have followed the last part properly you will have a good chance to put this up correctly. Now all you have to do is split Blender’s windows into four sections and include the respective image(views) in the viewport like the Fig 1. (The screen shot has been taken in Blender2.4 RC1).
You can enable the viewport name from the preferences menu int the ‘View & Controls’ tab (View Name). So you can see in which view projection you are in. Notice the Offset numerical buttons, these are the keys for you to align your image, here the reference lines we placed on the blueprint image in last step help you to get a good alignment.
Now just to make sure your alignment is working all you have to do is insert an cube primitive and start extruding it in top view if the extrusion is outwards then it should b outward in front as well as back view and towards the user or the side view. See Fig4 for example.
Modeling Car Body
There are two approaches towards polygon modeling a car in Blender one is ‘Box modeling’ and another is ‘Plane modeling’. They are completely opposite in nature, one attempts to create shape in blocks and then approach towards detailing while the later attempts to do it directly but in parts.
Step 1 Plane modeling: I have no experience with box modeling so I will go about explaining few approaches which could save you time and effort while modeling a car body using ‘Plane modeling’. The best way to work is usually to work with parts like the front fender or the bonnet, see illustration 3 for more.
We can start immediately by adding a plane in the side view. Before doing that position the cursor as seen in the Fig 2.
Position it over the front fender and reduce its size so that it can cover the contours of the fender just above the tire as in Fig 4.
Step 2 Constructing geometry: You should enable Subsurf of the value 2 for view-port if your graphics card can handle it. One thing you need to keep in mind while you are modeling with blueprints is that you will have to model and recheck the mesh in three views. Three because once you are modeling for example the front part, the back view will not be useful for any references. If you are not doing it often enough you will be ruining the geometry of the mesh. So immediately checking and modifying in other views as soon as you make some major changes in the mesh in one view will allow you to keep your mesh in clean state.
Following this now get into top view and move the top vertices a little back and extrude a set of another vertices towards the bonnet of the car to cover the mesh fas seen in the Fig 5a. Now move the vertices as seen in Fig 5b following the contours of the front fender.
TIP: While modeling a car body in SubSurf it is advisable to initially use only very few extrusion to chalk out the basic outline of the shape or contour of the part. For example in Fig 5a we have use only three extrusions to make the side of the fender mesh. While we will need more cuts in it to create more detail but that can come later. This method will allow you to have a great control over the mesh complexity.
Following the method we will now immediately switch to side view again and adjust the contour or shape there as can be seen in Fig 6 the result of which can be seen in Fig 7 and Fig 8.
Step 3 Detailing: Always add more vertices or extrusions in which you can see the major portion of the part for example we can sideview shows major portion of the fender. Switch to it and add more extrusions to cover other portions in this view. And following the step 2 correct the geometry in top and front views. An initial and corrected version can be seen in Fig 9 and 10 respectively.
a) Bends or contour lines: The shapes of car have often pointed surface guides or bends like near the fender rim and a slight elevation near to the bonnet in this car(Nissan350z). These unique bends can be seen in the Fig 10 as highlighted orange edge lines.
In Subsurface you will need approximately three edge liken nearby to bring out sharp bends. In Fig 10 you can see that the fender rim looks pretty sharp as in the real life images of the car. TO achieve the sharp blending it will help if you keep your mesh in clean state by only allowing to have quad faces near the place where the bend or sharp contour will be placed.
Placing an bend is as easy as using the loop cut tool at the required part. Check Fig 11 for the results.
Fig 10: Emphasizing contours.
b) Clean edges. All the body parts of the car are bent inside to created a clean edge look. we have to replicate it in our mesh too to get quality models. Here all of the outermost edges are selected and extruded once. Now the movement of the extrusion should be downwards for the part that are at top and it should be inwards for the part that are at side and again inwards for the parts that are at front or back.
To understand watch the Fig 13 carefully. You will need to select and move the extruded parts separately.
Fig 13: The edge details after extrusions.
TIP: Snapping. It is a good idea to enable snapping from the preferences menu, as the car modeling is all about detailing and when you move vertices of edges in perspective view the snapping could help you a lot, like while extruding to create and bend it will give you nice known results.
Good looking ’tillable’ wood textures are often hard to come by, and that too if they are needed to be suitable for a particular job. The usual way is to modify the image(photograph) based textures, but in most cases they are not up to the mark. As photographs mostly come with lighting distortions, which, unfortunately, make them unusable or unattractive for tillable textures.
- Step1 You can choose your own resolution as usual, but here I have chosen 512×512. First off we will create bands for the wood. Go to Filters>> Render>> Clouds>> Solid Noise. To get bands(vertical here) increase the ‘X size’ to 16 and reduce the ‘Y size’ to 0.6. You can copy the random seed value or choosing the value of your own.
|Image 1. Generating the wood patterns.|
- Step2 Next step create a new layer, rename it to Mahogany color. Select the color #5f0909 from the color picker. And fill the new layer with this color. Change the layer mode to Multiply. This gives us the very basic texture.
|Image 2. The Mahogany color.|
- Step3 You would have noticed that right now the texture have very subtle color radiations. We will increase the saturation so that they are more visible. Select the bands layer and press Layer>> Color>> Levels. Either copy the values from the image3 or make sure the ‘Preview’ is checked and modify the black and white level until you are satisfied.
|Image 3. Increasing saturation.|
|Image 4. Applying noise on banding.|
- Step4 The bands are too plain to be convincing. We will make them a bit dirty. Go to Filters>> Noise>> ScatterRGB, un-check the independent RGB and increase the value to 0.46. The independent RGB created colored noise while unchecked one gives black and white noise and that what we need, you can use that also and after applying use Layer>> Color>> Desaturate for approximately same effect.
- Step5 The noise addition destroys the plainness but is too coarse to be usable. We will not apply motion blur vertically. Go to Filers>> Blur>> Motion blur. Change the length value to 20 and angle to 90 press OK button to apply. The noise patterns will get washed vertically creating interesting patterns in the bland wooden bands and the over result is much better in comparison to the plain bands we created in step1. But is still lacks the chaotic variation of the natural wood.
|Image 5. Applying Motion blur.|
|Image 6. Before distortion.
|Image 8. Hard banding for a new image.|
- Step8 We want more banding lines in our texture and the ‘Solid Noise’ tool does not give much control over this feature, so we will compress the new image to half in width only and leave the height intact. Then select all (Ctrl+A) and copy the new texture.
Image 9. Compressing the image width.
|Image10 Layers position.|
- Step9 Paste the copied texture in the older as a new layer on top of all. Now go to Layer>> color>> Invert to invert the new layer. change the Layer mode to ‘Burn’ and bring the opacity down to 20. There you have it the Mahogany tile texture.
- Step10 The texture although complete is not perfect until its tillable. Save the Image copy now and flatten the image now use ‘Ctrl+Shift+O’ and change the offset to 256 in both X and Y. Start editing the texture using the clone and smudge tool until you deem necessary.
|Image 11. The seam problems in the middle.|
|Image 12. Tillable Mahogany wood texture.|
Creating a vase in Blender
Part 4: Scene Lighting
- Step-20 For lighting the scene we will add two more lamps. You can either add new lamp from the Toolbox [Space bar] or duplicate the original lamp. To duplicate first select the lamp then press [Shift+D] and move the duplicate object to desired position. Get in front-view [Numpad1] duplicate lamp once and move it down to nearly bottom of the original lamp.
- Step-21 In top-view [Numpad7], duplicate from the last duplicated lamp, move it opposite but equi-distance between Camera and the Vase. Keep it slightly off the straight line with the camera, as this lamp will be used for highlighting the side of the vase. Save your work.
- Step-22 If you render the scene now, you will see two bright spots on the floor, caused by the two duplicate lamps. We don’t want that and here is where Blenders ‘Layers’ feature comes in. The layers button are visible on the Menu-panel of the 3dView. Two layers will be sufficient for our scene, to add a layer keep pressed [Shift] and click on the second layer.
- Step-23 Select the floor and the original lamp then press [M](move to layer), in the layer pop-up push the second layer button as you did in last step. Why you may ask that we are using more than one layer. Well Blender allows lamps to effect either ‘local’ layer or the whole scene, using this feature we can remove those two bright spots in the render. While we want the two duplicated lamps to affect the highlights of vase.
- Step-24 Select the original lamp and in ‘Shadow & Spot’ tab enable ‘Ray shadow’ to enable shadow generation, remember shadow casting should be enabled only for original lamp. Select any duplicate lamp, press [F5], in ‘Lamp’ tab click ‘Layer’ button also click ‘No diffuse’, repeat that for other duplicated lamp. Save your work now and if you wish to do the test render.
- Step-25 If you did a test render you can see the render with nice shadow. Blender has a GI or global illumination feature called, ‘Ambient Occlusion’ this behaves much in the way natural light behaves, i.e. it also takes into account light bounces among objects, and this makes the scene more realistic. But for some reasons this make its computation intensive so if you have slower processor you might have to sit for long time as the rendering completes.
- Step-26 We will now go to World buttons to enable AO. Press [F8] and in the World buttons goto ‘Amb occ’ tab and enable the ‘Ambient Occlusion’. Here enable ‘Use distances’ and increase the ‘DistF’ value to 2.000. Press ‘Both’ button. Set ‘Energy’ to .80 and ‘Bias’ to .010. Render the scene. Most probably you will see noise in the rendering when it is finished. Depending upon your system capability you can increase the ‘Samples’ in AO tab. for this scene it was set to full at 16.
- Step-27 Our work is ready for rendering. We just need to tweak the rendering options to get good results. Press [F10] and in the ‘Render’ tab enable OSA and increase the value to 8. The OSA or oversampling or anti aliasing make the jagged edges in the render smooth and increase the quality of the image. If you want to increase the size of rendering then go to ‘Format’ Tab and input the values manually. Also here you can change the format for saving the rendered image. After you are done press [F12] for final render. After rendering press [F3] to save the image, and remember to input the proper extension, as Blender does not add the extension automatically.
Creating a vase in Blender
Part 2: Modeling
Part 3: Material & Textures
Part 4: Scene Lighting
Page 3 of 4
Part 3: Materials & Textures
- Step-12 Now we are going to apply materials and textures to the vase and the floor. We will start with floor first. Select the floor in object-mode. Press [F5] to bring up Shading & Materials features in the buttons-window. Press ‘Add-new’ button to assign a material to the floor. As soon as you assign a new material the ‘Add new’ button gives way to ‘material name entry-box’ click it and type ‘Floor-mat’ for easier identification.
- Step-13 After assigning a material you can now add textures to it. For floor, we are going to add a wood image texture. Press [F6] , the Texture-buttons window opens. In ‘Texture’ tab click ‘Add new’ to add a texture channel. Immediately a ‘Texture-type’ drop down list appears from it select ‘Image’. Again a new tab called ‘Image’ tab appears next, click the ‘Load Image’ button and browse to the texture wood-tile.jpg, select it and press [Enter]. In Xrepeat: and Yrepeat: increase both values to 4.
- Step-14 After adding the wood image texture, we will now modify the Shading parameters of the floor material. Press [F5] and got to ‘Mirror Transp’ Tab next to Shaders and press ‘Ray mirror’ and increase the ‘RayMir’ slider to .20. And finally Depth 3 (under Ray Mirror button), that’s all is needed for the wooden floor material.
- Step-15 Following the procedure in step-11 add a material for the vase. Then add three image textures namely vase-tex.jpg, vase-tex-spec.jpg and vase-deco.jpg in three different channels. You can rename the texture channels by entering text in the “Add new” box which by default will show Tex(#). Rename all of them suitably.
- Step-16 We had added three image textures for vase and now we will modify the shading property of the vase material. Press [F5], in ‘Shaders’ tab change ‘Ref’ (diffusion property) to .400, next change Spec (specularity) to 1.700 and Hard to 400 also Depth=3. ‘Lambert’ and ‘CookTorr’ are the algorithm’s user to calculate the diffusion and specular property of the textures.
- Step-17 Now we need to give proper mapping values to the three texture channels. In the Texture tab(far right) you can see the three texture channels being visible. The ‘Map-input’ and ‘Map to’ tab will change the value of only the channel which is currently pressed(vase-tex). So to modify the properties of other channel you have to first select them. Go to ‘Map input’ Tab select Tube mapping for all three channels.
- Step-18 In Tab ‘Map to’ the first texture channel need not be changed. For sec channel disable ‘Col’ and enable ‘Spec’ and ‘Emit’ buttons. Also, change the Var slider to .500.
- Step-19 For the third and the last channel change the texture blending mode to ‘Multiply’ from the default ‘Mix’ and ‘Col’ value to .500. Try out the render now to see the materials you applied. Also, save your work now.