3D Modeling: Exercises

on

29.8.2022 - 19.9.2022 (Week 1 - Week 4)
Chung Yi Ki / 0345014 / BDCM
3D Modeling
Exercises

Quick links

Lecture

Week 2 / Introduction to 3D Modeling
Blender user interface 
Top to bottom: 
-Top bar (menu, workspace, scene and layout)
- Area (area under workspace) 
- status bar 

Editor: Area where user edit the object 
Outliner: Work similar with layers 
Properties: Show property option for the objects
Timeline: Used to playback and keyframe animation

You can change the workspace name by double clicking the workspace tab. Workspace name can be changed as well by first selecting preset and renaming it. 

To split window: hover to the corner of the screen until a “target” icon appear and drag the window to the side to split
To rejoin window: drag from the corner of the window until an arrow icon appears, and then drag to the window that you want to rejoin to 
To change modes of editor: click the button beside “object mode” 

View axis
In blender, the Z axis is the one that controls the height, this is called z-up system. y axis controls back and forth, x axis controls left and right.

In the 3d view axis, the circles with fill colours are positive values, the ones with outlines and lowered opacity circles are negative values.
Note: Front view of object is at -y value

To rotate the viewport, can click, hold, and drag the middle mouse button. To track view (pan), hold down shift and click middle mouse button to drag. To zoom in and zoom out, scroll middle mouse button. To have smooth zoom, hold down ctrl and click, drag middle mouse button. 

3d viewport 
Fig 1.1 Interface of 3D viewport

There are two types of view angles in blender: perspective and orthographic 
Orthographic view: top, front, side, no vanishing point  
Perspective view: has depth or vanishing point
Quad view fills the viewport with 4 windows showing different angles (top, front, right orthographic and user perspective). To open quad view, press ctrl + alt + Q on keyboard.  

Viewport toolbar 
From top to bottom: zoom, track (pan), camera, orthographic view

Maximize/minimize orthographic view
To open property shelf, press N on the keyboard or drag the little arrow beside the view axis control. After that go to quad view > uncheck lock rotation > mouse over to any of the 4 windows and press ctrl+alt+Q to maximize each of the windows in quad view. Note that quad view option will only show up when quad view is open. 

Object selection: left mouse button > click on object to select the object 
De-select object: left mouse button > click anywhere on the 3d viewport except the object
Multiple selection: hold shift + left mouse button and click on any object
De-select object in multiple selection: hold shift + left mouse button click on the object (may need to double click at certain times)

In class activity
After Mr. Kamal gave us a lecture on the interface of Blender, we followed along Mr Kamal’s demonstration of modelling simple 3D objects. We modeled a snowman and a table with fruits using simple shapes. 

Fig 1.2 Snowman 3D model

The snowman body is made up of spheres, its hat and hands are made up of cylinders and its nose is made up of a cone. The shapes are deformed from its default look using the transform tool. For example, the hat is a cylinder scaled down and up, same goes to the hands but with rotation for the fingers. The sphere is later applied shade smooth to smoothen out the default edges. 

Fig 1.3 Table with fruits 3D model

The table is made up of cylinders with their scale adjusted, same goes to the fruit plate. The fruits are made of different sizes of spheres. 

Week 3 / Rendering and Non-Destructive Polygonal Modeling
Viewport shading and rendering

The default render engine in blender uses Open GL, the same as workbench. It carries out viewport rendering, or in other words, what you see in the viewport is what you get when you render. Another render engine is Eevee which carries out real time rendering. The third render engine is Cycles which is
CPU based rendering. 

Steps on assigning material to objects
  • Select the object
  • Click material properties in the property editor (the 2nd last button)
  • Click new to create new material 
Viewport display settings in material assignment menu
Color - changes color 
metallic - adjust the metallic appearance of the object from a scale of 0 to 1
Roughness - adjust the shine of the object from a scale of 0 to 1, 1 indicated non-shiny
material 

Assigning the same material to different object 
Method 1
Previous created material can be assigned directly to different object by clicking the icon
next to new (aka: material library) Material name can be changed by double 
clicking it in the material window 
Method 2
Material can be linked from one object to another by first pressing the object that 
you want to assign material, next press the object which has the material you want,
press ctrl+L on keyboard and select link material 

Changing the camera angle and frame in camera view
View the object in camera view by pressing the camera button in the viewport toolbox or
by pressing "0" on the numpad of your keyboard. Then, press "N" to bring up properties shelf, after that go to view lock and check camera to view. Uncheck it after you're happy with the view and don't want to change it.

To set the size resolution, go to properties editor > output properties (3rd button). We can also choose a preset by clicking the list icon button beside format.

To display the rule of third line in the viewport, first select the camera. Then, select object data properties (the camera icon in the property editor) and go to viewport display > composition guide, check third. 

Shading and rending look setting 
In the shading options, click the down arrow button to access the shading settings. Make the viewport is in solid mode before doing this. 

Lighting settings
Studio - Default setting that controls lighting with applied materials 
Matcap - provides preset materials to choose 
Flat - gives a flat appearance 

Shading settings 
Click material under color to view color from the applied materials. 
Check shadow under option to add shadows. To customize the shadow, click the cog icon beside
it. Click and rotate the sphere to change the shadow direction. Shadow shift adjust the coverage of the shadow and shadow foc asjust the shadow blending to look soft or sharp. 

Final output rendering 
First, disable show overlay to hide all display except the model. Then, go to view > viewport rendering to render the final image.

In class activity
While presenting the lecture, Mr. Kamal demonstrated how to assign material and render a scene in Blender by using the snowman model that we made in week 2 class. We follow along the demonstration using our snowman file. 

Fig 2.1 Snowman viewport display

Fig 2.2 Rendered snowman scene in studio lighting

Non-destructive polygonal modeling 

Apply simple deform modifier 
Before applying simple deform modifier, the object needs to be subdivided for the modifier
to work the right way. To apply a modifier, click the modifier properties, which is the spanner icon, in the properties editor.

When applying subdivision surface, choose the simple option and set the subdivision level viewport to the desired number (usually 2 or 3). Subdivision modifier should be applied before simple deform modifier

To view the object wireframe at viewport display, make sure overlay is turned on and click the drop down button and check wireframe under geometry. Optimal display should be turned off in the subdivision modifier too to see the wireframe of the subdivided geometry. 

To apply simple modifier, select the object, go to properties editor > modifier properties > add modifier > deform > simple deform. 

Four settings in simple deform: twist, bend, taper, stretch

Week 4 / Continuation of polygonal modeling - non destructive method 

Array Modifier 
Used to create a series of duplicated objects along a path. In the settings of the modifier, choose fixed count in fit type and in the count box, key in the number of duplication.

Plain axes can be added as a point to set origin for simple deformation. So, direction of deformation
can be changed by changing the direction of the object.

When using simple modifiers, first apply the subdivision surface modifier to an object and apply the chosen simple modifier. 

Problem with using bend in simple deform for plane
Bend would only work on z-axis, x and y axis will have no influence on a subdivided plane. To solve this, first rotate the plane up to 90 degrees on the axis, facing the front side. Second, apply all transformation by selecting the object and go to viewport menu > object > apply  > all transform 

Using array along a curve
First apply array modifier on the object, set the count number. Then, apply the curve modifier and select the curve you want the deformation to be based on by using the eyedropper tool next to  curve object

Create curve 
To create a curve, go to viewport menu > add > curve > Bezier 
To manipulate the curve, go to edit mode (shortcut key: tab) and use the curve point to position and curve handle to shape the curve.
To extend the curve, select a curve point and press "E" to extend it. 

Set the radius setting for the curve 
Select the curve, go to edit mode, select a curve pont, press "n" to open the property shelf and set the radius value in the transformation setting 

Lattice 
First create a cube, then scale the cube to a desire size. Make sure the subdivision is 
enough. Then go to viewport menu > add > lattice to create a lattice object. Scale and move the lattice to wrap around the transformed cube.

After that, more subdivision need to be added to the lattice object. Select the lattice object and go to object data properties in the properties editor (2nd last button). Set the resolution count to U:3 V:3 W:6

Then, apply the lattice modifier to the transformed cube by selecting the cube, go to modifer > lattice and pick the lattice object using the colour picker tool beside "object". Finally, modify the cube using the lattice in edit mode.

Cylinder might be hard to transform, start off with spheres or cubes. 

Week 5 / Destructive method polygonal modeling
Non-destructive modelling involves the use of modifiers, while destructive modeling involves 
editing the vertices of the object manually in edit mode, without the use of modifiers. 

Type of select mode
In edit mode, there's 3 select modes that can be chosen beside the "edit mode" dropdown menu. 
From left to right:
Vertex select - Select a vertex of the object 
Edge select- Select two vertices 
Face select- Select a face of the object 

In edit mode, the vertices, edge and face of an object can be scaled, rotated or moved around. 

Polygon types
There's 3 types of polygon, 3 sides polygon is called a triangle, 4 sides polygon is called a quad, and polygon with more than 4 sides is called a "n" gon. Quad polygons are the best surface to model in. 

Note that the object always need to be subdivided to see the quality of the surface. That is, model that isn't subdivided enough will have limited number of components or LOD (low level of details), known as low poly models. When the model is subdivided to have enough number of vertices, more details can be seen on the model, this is known as high poly models. 

When modeling an object, always start with low poly modeling using primitive objects. When a new object is added in Blender, the amount of segments, rings or size of the object can be customised on the pop up menu.

Symmetrical model 
In blender, the x and y axis needs to intersect the object at a symmetrical plus sign where each space around the plus sign has a symmetrical amount of faces. For example, the lowest amount of segment for a cylinder to be symmetrical is 8. 

Always start with planning a symmetrical model and move the object above the floor grid so that the view of the object will not be obstructed when modeling. Remember to always understand proportion when modeling, sizing of the object is not as important as it can be adjusted later on. Other than that, always model the object where the front side of the object is facing the front view of the camera. 

Loop, cut and slide
When modeling on primitive objects, the objects can be manually subdivided by splitting the component using loop, cut and slide. Go to edit mode > edge> loop cut and slide > scroll mouse to increase or decrease the loop cut. Short cut key for loop cut is "ctrl + R". Then, left click to confirm and right click to disable slide. Though, this can only be applied to quad surface. 

To select an entire edge loop, alt + single mouse click on a vertex to select the connecting vertices that make up the edge loop. 

To select a round of edge ring, ctrl + alt + left mouse click on one vertical edge to select the connecting vertical vertices that make up an edge ring. 

When scaling to shape the object, remember to scale the object uniformly to get a symmetrical outcome and avoid scaling in one axis only.

Extrusion
Another method to manually split the component is by extruding the surface. First, select a face of an object and press "E" on the keyboard to extrude and scale when needed. 

Smoothen the object surface 
After you're finished with modeling the object, remember to smoothen the surface of the object by using shade smooth. The degree of smoothness can be adjusted by first selecting the object and go to object data properties > normals > check auto smooth and control the degree of smoothness. 

Then apply subdivision surface modifier. The shortcut key for adding subdivision surface modifier is pressing ctrl + the number of subdivided levels. 

Creasing the corner 
Sometimes, the object can appear too smooth and this leads to loss of detail. One way to sharpen the object is by creasing the corner. The first method is by adding poly count and the second method does not involve the addition of poly count. 

Adding poly count
Add a loop cut and slide it near the other loop cut on the edge that you want to sharper. The nearer the new loop cut is to the other loop cut, the sharper the edge. Usually, adding additional loop cut above and below the designated loop cut will work. 

Without adding poly count 
Select an edge loop on the edge that you want to sharpen, and adjust the min. crease slider in the property shelf. 1 is the maximum amount. 

The pros of adding poly count is that it's friendly to other 3D software while using the min. crease method may not work in other software.

Minimize "n" gon bad appearance 
To minimize the "n" gon appearance at the base or top of the object after modeling it, first method is to use inset. Select the face you want to inset, press "I" on keyboard and drag until the inset surface is as small as possible. 

Second method is to use the knife tool to convert the "n" gon into a quad or a triangle. The shortcut key of knife is "K". Once the knife tool is chosen, click on any edge to freely add edge to create edge loop. Or click to any vertex, to add edge connecting 2 vertices. Press enter to exit the tool after you're done connecting your vertices or edges. 

Modifiers 
Bevel modifier is an alternative to bevel
Solidify modifier is an alternative to extrude.

Destructive method for solidify is to select all the faces and in the viewport menu, go to face > extrude faces along normal 

Bevel 

Fig 3.1 Examples of bevel types
Source: https://cgian.com/2022/05/how-to-use-blender-bevel-tool-and-modifier

Used to smooth sharp corners to give the object a rounded corner appearance. Go to edit mode  
and in the viewport menu, go to edge > bevel edges. Move your mouse to see the bevel and
scroll the middle mouse button to subdivide more into more rounded shape. 

Week 6 / Hard Surface Modeling 
In deciding number of segments for cylinders, always think in addition of fours, the minimum amount of segment is 8. 

Planning for symmetrical modeling 

Boolean tool 
The Boolean method is creating different shapes from two intersecting shapes through difference, union, intersection and slice. Boolean method can be applied using the bool tool in the Blender add-ons, or using modifier.
  • Union: joins two intersecting objects as one object 
  • Difference: removes the object and the intersection point, creating a gap. The last object to be selected will be retained
  • Intersection: only leaves the intersection point 
  • Slice: similar to difference but the intersection point is kept and can be moved separately from the main object 

Auto Boolean is a destructive method while bursh Boolean is a non-destructive method. Brush Boolean apply the effect as a modifier instead. 

Make sure that one object is larger than the other object and completely intersecting to use Boolean tool. 

Eliminate n-gon in Boolean 
Select 2 opposite vertices and connect them to form quads by going to viewport menu > vertex > connect vertex path, shortcut key is J 

Another way to create a hole other than Boolean is to scale a cylinder and delete all the faces until there's one face left, and then inset the face and delete the insetted face, and then extrude 

Instructions


Exercise 1: Modeling from primitives objects
For our first exercise, we are tasked think of any objects that can be simplified as a combination of primitive shapes and model them in Blender. We are required to apply shading using the viewport shader and compose the scene in the camera to output it as a 1280px x 720px PNG image. 

Visual reference 
Fig 1.1 Visual reference #1
Source: https://www.pixtastock.com/illustration/74436028

Fig 1.2 Visual reference #2
Source: https://www.researchgate.net/figure/A-sandglass-used-for-timing-a-one-hour-interval_fig4_252061612

Fig 1.3 Visual reference #3
Source: https://www.pixtastock.com/illustration/71097902

Since I thought we had to finish this in class the day we were briefed with the exercise, I went to a random object generator website to get some ideas and one of the things it suggested me is a sandglass. I thought that's doable in a short time so I just went with it (I did continue to make the model more interesting by adding details later on after class). Before I start to model the object, I went to search for some image references on the structure and details I can add to a sandglass model. 

Modeling process
Fig 2.1 First attempt at modeling - perspective view

Fig 2.2 First attempt at modeling - perspective top view

I started with using two cones as the glass of the sandglass, cylinders for the base, cover, poles and metal gold rings. The sphere is there to add some details to the poles and the same goes to the smaller squashed cylinder at the top of the sandglass as well. 

Further development of modeling

Fig 3.1 Video showcasing the further detailed model

Fig 3.2 Further detailed model - Perspective view

Fig 3.3 Close up of sand falling down structure using cone and cylinder

Fig 3.4 Further detailed model - Top perspective view

Then, I added some more detailing to the sandglass model with the gold material with reference to the image references. The sand in the glass is also made with two cones with different scaling, though for the part where the sand is falling down, it's made with a cone at the top and continued with a cylinder which goes through the bottom pile of sand. The base and cover of the sandglass is modeled to have a oblong shape rather than a perfect circle to give some stylisation to the design. The, glass, wooden base and cover of the sandglass is shaded smooth to remove the edges. 

Fig 3.5 Scene setup using planes

Fig 3.6 Scene setup using planes (without wireframe)

After that, I used two planes to create a dark blue backdrop as the setup for the render scene. Dark blue is used as it contrasts with the colours of the sandglass, thus drawing the attention to the sandglass more and accentuates the shine on the gold metal material. 

Fig 3.7 Properties of the material used

Fig 3.8 Settings of the viewport shader

As for the material choice for the shapes, the gold metal look is created by increasing the metallic slider, the glass look is created by decreasing the alpha slider for transparency and decreasing the roughness slider for the smooth, reflective look. Metallic slider is increased a bit to give it a bit more shine. The materials for the base and the poles are set with varnished wood in mind, and the roughness slider for the sand material is increased to give it a matte look. 

For the viewport shading, rim light preset is used to give the sandglass a grand look and the lighting direction is adjusted very far to the back to make the shine on the gold metals more prominent. Backface Culling is enabled to give shading transparency to the glass material, and the outline is in dark red to give some subtle sense of volume to the objects. Other properties are adjusted to control the brightness of the scene. The scene is later composed according to the rules of thirds and exported as PNG image for submission. 

Final primitive object model 
Fig 4.1 Final sandglass model

Exercise 2: Non-destructive modeling
For this exercise, we are tasked to refer to animal drawings that show the basic form construction of the animal in primitive shapes. Then, we are to model the animal based on the drawing using primitive shapes and non-destructive modeling tools such as subdivision surface, array, curve, simple deform and lattice in Blender. The final model would need to be assigned basic materials, composed in the camera and output as image for submission. 

Visual reference

Fig 5.1 Main image reference
Source: https://drawpaint.art/dogs/

Fig 5.2 Additional image reference (dog drawing)
Source: http://conceptdesignacad.blogspot.com/2012/05/animal-anatomy-with-jonathan-kuo-1-spot.html?m=1

Fig 5.3 Additional image reference
Source: https://www.deviantart.com/seadre/art/Sketch-dog-358922101

The animal that I chose to model for this exercise is a dog. Fig 5.1 would be my main reference throughout creating the model, while the additional references are for me to refer to different angles and shape construction of the dog, since Fig 5.1 just shows it at one angle. Fig 5.3 is mainly used to reference the tail and ears.

Modeling process
Fig 5.4 Blocking out the shapes and pose

I first started with roughly blocking out the form of the dog using shapes, modifiers and lattices, so that I can get a sense of its proportion and it would be easier to refine the model later on. Mainly cubes are used for the model, except for the neck, joints and legs, which are made of cylinders and spheres.

Fig 5.5 Modeling the tail

Then, I proceeded to model the tail using bezier curve and a cube with array and curve modifiers. Subdivision surface modifier is also added to the cube with a level of 3 so that the tail would look smoother. The radius of the bezier curve is late adjusted in edit mode using the radius slider in the objects property sidebar so that the tail would go from thick to thin. 

Fig 5.6 Modeling the ears

I also added ears to the model so that it would resemble a dog more clearly. The ears are made using a cube with the subdivision surface, taper and stretch modifier added. The scale of the cube is adjusted while using the modifier so that it would form a triangular shape cuboid. 
Fig 5.7 Modeling the paws

Fig 5.8 Adjusting the shape of the paw in edit mode

For the paws, they are modelled from a cube with multiple simple deform modifiers added. The cube is first subdivided using the subdivision surface modifier, and then tapered and stretched with a restriction to the Z axis so that it won't stretch vertically. The scale of the cube is adjusted as well to match the shape of a paw. The vertices of the cube is also adjusted in edit mode to make the front face more slanted without affecting the other parts. 

Fig 5.9 Lattice object for head and mid-section of body

The shape of the head was first modeled using stretch modifier and the body is modeled using a combination of stretch and taper modifier, and then both are fine tuned using lattice. For all the other parts of the body, a combination of stretch and taper simple modifiers were used to create the needed shape, and the vertices of the objects were later fine tuned in edit mode to make the shape more accurate. Though, the front section of the body is done only by adjusting scaling of the cube and its vertices in edit mode. The thigh of the model on the other hand, is created using a cube with subdivision surface modifier in Catmull-Clark option with 2 levels, and with scaling adjustments. 

Fig 5.10 Refined tail shape

The shape of the tail is later refined as well to match the model. 

Assigning material and composing in camera 

Fig 6.1 Video showcase of model 
with and without material and wireframe

Fig 6.2 Still images of model in different views

Fig 6.3 Settings of materials used in the model

Fig 6.4 Model colour reference
Source: http://www.vetstreet.com/dogs/doberman-pinscher

After finishing with modeling the dog, I then assigned materials to the model. The colours are chosen based on the Dobermann dog breed colours (Fig 6.4), as I thought the form of the model kind of suits the look of the breed. The materials setting are assigned with the model being a toy dog in mind, made of brown metals and dark brown wood. 

Fig 6.5 Scene setup

Fig 6.6 Camera composition

Fig 6.7 Viewport shading settings

I then used two planes to create a scene with the colours chosen to look like an outdoor setting. The outdoor.sl lighting setup preset is chosen as well to create a outdoor looking scene. I composed the model at a low angle as it seems more interesting and make the composition look less flat. I align the dog's head to the center line of the camera to create a focal point on that spot. 
Fig 6.8 First attempt of dog model

Fig 6.9 Reference picture

After receiving feedbacks from Mr. Kamal, I resized the front legs to be a bit bigger to make the proportions better. I also searched for a reference picture to refer to the size of the legs. 

Final non-destructive modeling model

Fig 7.1 Final non-destructive modeling dog model (updated)

Exercise 3: Modeling from cylindrical shape

For this exercise, we were tasked to model any different objects made of cylindrical shape using tool such as extrude, loop cut, bevel and inset. After that, we need to compose the objects in a scene and render it with light and shadow using viewport shading, and export the render image in 1280px x 720px resolution. 

Visual reference 
Fig 8.1 Reference images used and source link

I started with finding reference images of the objects that I want to model for this exercise to use as blueprints for my modelling. I tried to find different shaped cylindrical objects to give variety to the composition. The Coca-Cola bottle picture was given by Mr. Kamal for our in-class activity. 

Modeling process
Fig 9.1 Modeling Coca-Cola bottle following the reference image

Fig 9.2 Modeling the other objects following the reference images

Fig 9.3 Modeling the plate using reference image as visual guide

Fig 9.4 Bowl modelling from in-class activity

The Coca-Cola bottle, glass mug, champagne bottle and glass, and vase were modeled directly following the reference image. Since the reference image for the plate is in perspective, the plate was first modeled directly following the image and then further adjusted and resized to make it look more like a plate in all angles using the image as a visual reference only. All of the objects were modeled from a cylinder with 12 segments (except for the glass mug) mainly using loop cut, extrusion and inset for the base. For the Coca-Cola bottle and the champagne bottle, the bottle cap and gold tin foil were modeled separately from the body of the bottle. 

The models were then shaded smooth and subdivision surface is applied to make the model look smoother. The subdivision surface setting was changed between using Catmull-Clark and simple depending if the model needs more edges details or a rounded smooth detail. Solidify modifier is also applied to plate, vase and champagne glass to give the models a thicker appearance. 

An additional bowl was made for an activity in-class by cutting the sphere in half. The base of the sphere is then moved upwards to create a flat base by selecting multiple loop cuts and transforming them. The thickness of the bowl is done through a destructive method by selecting all the faces on the surface of the bowl and extruding them by clicking faces > extrude faces along normal in the viewport menu. 

Details of modeling glass mug

Fig 10.1 Basic form of glass mug using cylinder with 8 segments

Fig 10.2 Applied 1 level of subdivision surface to create more vertices

Fig 10.3 Creating the pattern groove on the glass mug using inset

The glass mug was done with a bit more detail than the other models. I started with modeling the glass mug by directly following the reference image using a cylinder with 8 segments, loop cuts and extrusion. After the basic form of the glass mug is modeled, subdivision surface modifier was applied to the model to create more vertices and the indented groove pattern along the glass mug was created by selecting a row of faces at once and identing them. The handle of the glass mug is done by selecting faces at the side of the model and extruding them while applying the necessary transformation. The end of the handle is joined to the bottom part of the glass mug by selecting the faces that needs to be joined and using bridge edge loop by pressing F3 on the keyboard. 

Composing the objects
Fig 11.1 Resizing the object to correct proportions

Fig 11.2 Composing the objects according to rules of thirds and center lines

Fig 11.3 Applying materials and setting the scene of the composition

Fig 11.4 Materials settings for the objects

Fig 11.5 Viewport shading settings

After finishing with modeling the objects, I then put all the objects in a line on the same axis and resized them until their proportions make sense. After that, I arranged them on a plane by moving them around to create a composition. Different materials are then applied and light and shadow of the viewport shading is then adjusted. The studio light preset was used for the shading. 

Fig 12.1 First attempt of render

After receiving feedbacks from Mr Kamal, I went to apply Catmull-Clark subdivision surface modifier to the coke bottle, the champagne gold foil and the bowl to give them a smoother appearance without the visible lines. 

Final modeling from cylindrical shape models

Fig 12.1 Final cylindrical shape models (updated)


Exercise 4: Modeling a Karambit
For this exercise, we were given a reference image of a Karambit and were tasked to model it using the hard surface techniques tool that we've learned. The final output should be rendered in 1280px x 720px resolution with viewport shading settings. 

Modeling process
Fig 13.1 Given reference image

Fig 13.2 Base form of the model

Fig 13.3 Creating hard surface details using knife tool

I created the model starting from top to bottom using a hollowed out cylinder. The hollowed out cylinder was created by scaling the cylinder to the needed size and thickness, and remove all the faces until there's one front face remaining, then using inset to match the size of the loop and delete the inset face. Thickness is then added again by extruding the faces. The rest of the knife is then modeled by adding loop cuts and arranging the wireframes. After the base form of the Karambit is modeled, more vertices are added to create the hard surface details by using the knife tool. 

Fig 13.4 Duplicating a part of the knife for the handle

Fig 13.5 Reshaping the duplicated part to match the shape of the handle

Fig 13.6 Adding hard surface texture details using knife and adjusting vertices

I then proceeded to model the handle of the Karambit. To make things quicker, I duplicated the original model to extract the the handle part of the knife since its shape is similar to the actual handle. The extracted part is then reshaped according to the reference image to match the shape of the handle. As the handle in the image has an uneven surface texture, I tried to reproduce it on the model of the handle by adding more vertices using knife tool and adjusting the position and scale of the vertices to create the texture following the reference image. 

Fig 13.7 Placing spheres in screw area and using cubes to create the gap

Fig 13.8 Close up of the screws with material

Fig 13.9 Viewport shading and material settings

Fig 13.10 Video showcase of finished model 
with wireframe, solid mode, and with material

After that, spheres were used as the screws in the handle. The hole for the screws were made by using the difference brush boolean function between a sphere and a handle. The screw is a smaller scale of the same spheres used and the screw gaps were also created using the difference brush boolean function between cuboids and the spheres. Some tweaks were made on the vertices of the edge of the knife and the surface texture of the handle so that their surface look more hard and their topology doesn't look weird. 

Fig 13.11 Setting the scene using a plane

Fig 13.12 Composing the Karambit 

Next, I used a plane which is positioned close to the Karambit to reduce the harsh shadow made on the plane, and to make it look like it's laying on a wall or table. As shown in Fig 13.9, the outdoor lighting preset is used as it gives the scene a brighter look, and also the light source shines from the top. The scene is composed with the camera centerline in mind. 
Fig 13.13 First attempt at Karambit model

After receiving feedbacks from Mr. Kamal, I added a subdivision surface modifier using the simple option at a value of 1 for level views. 

Final Karambit model

Fig 14.1 Final Karambit model (updated)

Feedbacks

Week 11
Exercise 1 is good. For exercise 2, the proportion of the front legs kind of looks a bit unnatural, maybe they need to be a bit bigger, can find more reference pictures to refer to. But silhouette wise it looks good. In exercise 3, the coke bottle, champagne tin foil and bowl have vertical lines in the render that may be because of a problem of subdivision. So maybe check it as the surface should look smooth.
The glass is good. As for exercise 4, the black part of the handle seems to look weird, but in terms of silhouette it looks good.

Reflection

The exercises definitely helped me in understanding how to use the tools in Blender a lot, and also gave me a chance to learn on how to compose an object/scene in different ways using different lightings. The most challenging exercise is definitely exercise 4 as I've never modeled hard surface before (the most I did was cylindrical objects as I've watched a YouTube tutorial video way back when). I had to fumble around to understand the process and get it right, but once I'm familiar with it, it gets easier to figure how to to use it and is especially helpful going into project 1. 

Comments

Post a Comment