Kerosene Lantern

Kerosene Lantern

.....

.....

This is a piece that was done for a class titled "advanced lighting". Strange name for a class that was really about modeling, but I'm not a school admistrator, so what do I know. In either case, the objective of this project was to teach us the process of modeling "high", and then taking it "low". This would include blocking out a mesh, detailing it, casting normals, then texturing the low poly object. Since I already had a decent grasp on this whole process, I decided to do a little extra at the end.

.............................................

.....

.....

.....

Like all my other projects, research is key. Research for this was a little different in that the lantern we were working on wasn't a specifc kind/brand. It was just sort of a generic kerosense lantern, so I found mixed resources. The different types I found helped give me an idea of how they worked, what the different parts were, surface detailing, and the general shapes of everything.

.....

.............................................

.....

.....

.....

1. I like hi-poly modeling a lot; I think it's fun, cathartic, challenging, and rarely gets old. I think the freedom of hi-poly modeling makes the process more enjoyable, but you still have to overcome obstacles and challenges of modeling. There are lots of tools at your disposal to make a render quality object look as real as possible, before even adding textures. In the "hi-poly, low poly, normal-casting" process it's obviously a critical step since it's what the low poly will receive it's normal map detail from.

.....

Here's my blocking mesh of the lantern. Almost the entirety of this process was shift-extruding edges up and out to form the shape of the object. Since this is almost entirely a cylindrical object, it's very symmetrical and the process that much easier.

.....

About the only deviation from the shift-extrude method was when I made the wire handle and the wires around the glass object. These were created with splines that were made render-able with only enough definition to get the main silhouette value across. I then used an FFD modifier to bend them around a bit and give them some mild distortion.

.....

.....

2. Several of the parts of this lantern are straight forward, symmetrical, cylindrical pieces, so describing the process for modeling them is unnecessary. I'll just go over the different stuff.

.....

Doing anything to a curved surface always presents it's own set of unique challenges, so the holes I decided to add to the upper portion of the lantern were my first "bump in the road". I feel like I've done this enough at this point that it's a pretty simple process, but for whatever reason little kinks always find their way in.

.....

In stead of using the technique showed in the tutorials, I've normally elected to use the method Michael Eudy showed for chamfering out a vert, connecting the edges of the new square poly that's made, then using the graphite modeling tool "geopoly". I then cut to eliminate n-gons.

.....

I knew I wanted 20 holes, so I made a 40 sided cylinder, isolated on part, cut the hole, then arrayed it out to make the cylinder whole again and welded. I didn't want the piece above or below this one part to be hampered by 40 edge loops, so I made it separately. The only problem with this is that you have to make really certain that it lines up with the pieces above and below, or it'll look separate in the render. I made sure by zooming in reeeeeeeeeeally close, then using the dialog box for the scale tool to put in very small increments of scaling till it matched almost perfectly. I never actually did match EXACTLY, but it's soooo close on such a tiiiiiny level, that it simply is not noticeable from a normal viewing distance (or in the render.) I'm sure there's a better way to do this, so feel free to shoot me an email and yell at me, I'd love to learn.

.....

.....

3. This piece is similar to the bugger just mentioned, but different in its approach. Instead of starting with a cylinder, I used a box to create the slats and the spaces between them. After they were made, and the necessary edge loops to define them in subdivision, I added the turbo-smooth.

.....

AFTER the turbo-smooth, and I must emphasize AFTER, I applied a bend to pull it around and together. Th reason for this being that the subdivision creates geometry that allows the bend to have the topology flow smoothly over the curved surface. If the bend were applied first, the edge loops constraining the form of the object for subdivision would not get bent properly, and the resulting turbo-smooth would make things look pinchy and ugly.

.....

.....

4. The smaller handle for the top of the lantern I made with a spline that was given some thickness. I often find splines to be immensely helpful for a great many number of things, and I really think more people should get comfortable with them to learn their power. There are many things in modeling that are just far easier to accomplish with splines, a wirey type handle like this is a perfect example.

.....

To go along with the handle, I made a small clamp that's "attaching" it to the top of the lantern. I didn't just want it stuck in the geometry, or just floating in some odd fashion, so I created that small piece, gave it some bolts, and applied a couple of bend modifiers to it so that it conformed to the curved shape of the top of the lantern. Could I have just made the top of the lantern flat and avoided the bending, sure, but I didn't think it would look as good or be as fun to make.

.....

.............................................

.....

.....

.....

5. The next pieces I went at were the large arms on the sides of the lantern. In a few different versions of this lantern that I saw, the arms had these neat simi-circle protrusions going around them. More to the modeling that just making a cylinder, sure, but again, more fun to make and I think it looks better in the end.

.....

To make them I used a combination of a couple of previous methods. To get the shape of the arms, I used a flat surface again that was turbosmoothed and bent. With the small object made (the blue one on on the left) I capped one side to get a face that I would use for "extruding along a spline". Again, splines to the rescue. Any number of tutorials on making pipe-like objects will refer you to splines, so that's exactly the mindset I adopted for this.

.....

.....

6. Okay, so, again, I like to find ways to avoid just sticking geometry into other geometry. When all other options are out the window, I'll go that route, but I want this thing to at least have some level of realism to it, and metal pieces never just jut into one other; they're either welded or screwed together.

.....

So for this, I opted for a welded look.

.....

To get the weld, I created a boolean object (first frame) that was an intersection of the upper piece of the lantern, and the arm that was intersecting it. This allowed me to have an edge that followed the surface of not only the upper part of the lantern, but the curved surface of the arm. With that edge, I "created a shape from the selection". That shape, a spline, was given some thickness and a noise modifier. Getting the weld to look the way I wanted took some tweaking, but I think I got it close enough.

.....

I think viewed up really close, it looks a bit wonky, but it's pretty convincing from a normal viewing distance (in my opinion, of course.)

.....

.....

7. Next, I made the glass, and the metal piece it's attached to; the one that brings the oil up to the cotton wick and allows the lighting process to happen.

.....

Nothing fancy about the metal piece, just added some bolts to it, since it's connected to the lower lantern body somehow, right?

.....

The glass was simple to, but I shelled it to allow better use of the arch and design thick glass material. It looks better than the thin geometry option.

.....

.....

8. I made another weld for where the bottom of the arm meets the oil basin. This weld was a little more difficult to get looking right, as the shape was definitely more complex that the upper one, but the same general process was followed.

.....

.............................................

.....

.....

.....

9. Next I made the pressure dial with another spline, and the cap and hole that are used to put the oil into the bottom basin of the lantern. Now, as far as how this part sticks into the bottom of the lantern, I opted for a simple circular tube. I used this instead of another weld because several of the reference images I saw looked like that. Either the weld made was just really clean and smoothed/rounded out, or it's one whole piece manufactured that way.

.....

.....

10. The next pieces I made were the metal outcroppings that the main handle of the lantern attach to, and the main handle itself. I saw several versions of this handle, but I like this one the most.

.....

.....

11. The next part was fun. After looking at several images, and analyzing what that wire around the glass was really doing, and how. I came to realize they acted as a type of clamp to stabalize the glass. They're clamped to the arms for tension, then ran underneath and around the glass to grip it onto the lantern tightly.

.....

The metal outcroppings from above were used again and altered slightly for this lower part. The wire clamp was made with ... you guess it ... A SPLINE! It's one whole spline too. It starts at one end, loops under, up, around, and back again. It was a complex spline to make, but I'd hate to make that any other way.

.....

.....

12. Finally, the finished, hi-poly, model.

.............................................

.....

.....

.....

.....

.....

The above images were renders I produced to show the end result of the hi-poly modeling. It wasn't required to make either of these sets, but I think presenting your work well is an important layer of polish. Both were textured with Mental Ray Arch and Design shaders. In the clay wire renders I used a method I learned from Jeff Patton from The Gnomon Workshop.

.....

In the second image I also used AnD materials; what I used was a sort of off-shoot of a car paint material for the main painted area, along with some metallic and glass shaders for the other parts. I used a threeway lighting setup for the object, as well as an HDRI that was generated in a different class from the library at school using a chrome probe.

.............................................

.....

.....

.....

This is the setup in 3dsMax of the scene used to generate the render seen at the top of the page. Using meta data from the image, I ascertained the correct focal length of the camera used to take the backplate photo, then matched that in Max's camera. From there I placed planes with the MR Matte Shadow Reflection shader appropriately and staged the camera. I used the same HDRI mentioned above. I used some lights as well as self illuminated materials for the inside of the lantern. In used the "lens_glare" shader in the "output" portion of the "camera shaders" in the render settings to get the glow. I later found a better way to do it in post to allow for more control in the comp, but I didn't get to implement it here. The major issue with using the lens shader like I did is that the specularity in the glass ended up getting affected by the glare too, and the result ends up looking a bit blown out. I did render this in passes with EXR embeded render elements, and then combined those using Nuke.

.............................................

.....

.....

BACK TO CONTENTS

BACK TO THE TOP