How To Prepare Metal & Leather Textures For a 3D Camera Model

Introduction

My name is Ewa Wierbik-Ziąbka. I’m a Senior 3D Artist from Poland. I graduated from the Silesian University of Technology a few years ago. I studied Biomedical Engineering.

My first encounter with 3D happened many years ago when I was just scrolling through the internet and found a program called Milkshape3D (I was 12 years old, but my computer was definitely not prepared for adventures in 3D). Then, I had a really long break and I started using Blender at the University.

I’m self-taught, so I started with tutorials I found online. After some time I realized I was really enjoying the whole process, so I started to look for a job in the industry, and since then, I have made 3D almost every day. I had a talk at the Blender Conference in 2023, which you can find here:

I made many renders, but I think the ones I’m most proud of are King Kombi (based on a concept by Alejandro Burdisio), skeletal phone, and the Peanut, the Bone Collector. 

About the Leica M2 Project

To be honest, it was supposed to be a quick project to add to my portfolio. I like photography, and I was looking for a nice prop to model and texture, so I finally decided to make an old camera like that. I love the pictures of the old Leica M2, especially in black paint, with visible damage and many years of usage.

I tried to find as many references as I could, starting with a blueprint to keep the correct size and proportions. Unfortunately, the only blueprint I found was a Leica M3 one, so I had to adjust everything to the rest of the photos. I’m always trying to find references showing the objects I’m modeling from every side, even from the bottom if possible.

For putting all references together, I used PureRef. 

I did the same for the lens, but the reference I found was the R-type mounting one (where I made a mistake because I totally forgot it should be M-type, but I corrected it in Substance 3D Painter later — you can see a little R on the front of the camera — it was later replaced by M).

Modeling

Thanks to many references, the modeling process wasn’t that difficult. All modeling was done in Blender 4.0.2. Some of the elements of my blueprint had other sizes than in the model I wanted to make, but the main shape remained the same. I think the biggest challenge was to model the top part of the camera and the dials with a diamond-like pattern, and create all the text textures (both on the body and the lens).

Similarly to the rest of the model, I modeled the small parts using the blueprint. I’ve tried to match the sizes of all the small elements to the references. Initially, I planned to make only the camera, but then I saw a few pictures with the lens, so I looked for even more references. I knew it wouldn’t be easy to make all these numbers and letters on the textures, but who doesn’t want to challenge themselves from time to time?

In general, I’m always using basic shapes to block out the shape (mainly to see the size of the object I’m modeling), and then I just leave a bounding box showing me the maximum sizes of the object. Then, using some primitive shapes (like cylinders for the body and the top part), I was able to make a little bit more detailed blockout.

I used the Knife tool for cutting geometry to create desired shapes, as I decided that I didn’t need to go with perfect quad geometry for this project. This time, I didn’t sculpt any details on the model, but sometimes I put a little bit extra work into creating the high poly model too.

Most of the time I’m trying to put as many details on textures as possible. The main reason why I changed the geometry on a body from a tull of triangles to a quad was to put a Displacement Map there to give the body more dimension. 

Unwrapping

Perfect topology was not the main goal of this project. Most of the models I make are not intended for gaming use but for scenes and rendering. On the other hand, I try not to put unnecessary vertices where they are not needed. So I put bevels on some edges just to make the baking process a little bit easier.

As in most of my projects, I’m using some tricks to hide the visibility of the connection of two objects that should be merged together. Data transfer modifiers with Shrinkwrap are my favorites.

I decided to go the traditional way and apply two materials to my model — one for the camera, and a second one for the lens (not using UDIMs for this one). After that, I started to mark the seams of the model in the less visible edges. For some parts, it was impossible, and I knew I’d have to hide a few seams manually in Painter, but in general, I tried to avoid putting seams on the areas of the biggest focus. I tried to straighten up all the circle elements to make painting the symbols/numbers easier and also to save some UV space.

ZenUV helped me to relax on some of the islands. I used UVPackmaster 3 PRO for packing my UVs. I’m usually setting the padding to ~0.004 but it depends on the texture size. If something is wrong after the initial bake, I’m just changing this value until the result is satisfying. 

Texturing

My texturing program of choice is Substance 3D Painter. For most of my projects, I usually use the material library I already have in Painter as a base, then I add more details, filters, and generators. I almost never merge the mesh to one object, because it’s easier to have control on the mesh if the UV is broken (at least for me).

If the top of the camera has the same material as some smaller parts, then I’m trying to create only one material (like Body and Lens&Elements), and then use some extra generators or just paint more details by hand. This way, I’m assured that they’re using the same material, but I have full control of adding even more details. 

As you can see in the video below, I’m using a lot of masks and anchors:

Details

As I mentioned previously, sometimes I am actually sculpting the details, but for most of my projects, I'm painting the details in Substance 3D Painter. The main challenge was to create all the lens indicator numbers and shapes and match the wear effect as close to the references as possible. 

First, every time I’m making a model, I’m thinking about its past. How old it is, where it was, how much was it used, and how it would look like now. Basically, it’s like telling a story about the object. Every material wears off a little bit differently, so it was important to show the difference between the metal parts and the leathery body. I didn’t want to make it look very damaged, that’s why only some of the most used parts are scratched and washed out.

For the text on the top (the logo and additional writings), I used an image found on the Internet and made an alpha image in Photoshop out of it, so I could use it for painting later. The numbers on the knobs were added in Painter, using its text tool. It was tedious and I didn’t match the font perfectly, but thanks to adding some dirt and imperfections, the final effect is looking okay.

For all of the lens writings and numbers, I looked for the best matching font from DaFont (I used 100% free-to-use font). The back film-type indicator graphic was made by overpainting the real image and making it bigger to show more details. Rest was also painted in Photoshop and then used as an alpha in Painter. 

Glass surfaces were made using the material mix in Blender — I painted some details, dirt, and scratches in Painter, but then I mixed them with a glass shader from Smouse’s Shaders Plus. 

Rendering

For the final model, I decided to go with a simple background. I’ve tried a few colors, but I found blue was the most appealing and let me emphasize every element I wanted to show. I used one HDRI for most of the shots, but I rotated it for different cameras. Also, a few of the cameras have a different light setup, helping to draw the attention where I wanted, so I just kept them in different collections and turned them on or off depending on the chosen camera. For some of the shots I had additional area lights or mesh lights, but in general, I tried to have neutral lighting.

For in-Blender post-production, I’m using a Photographer addon and I’m mostly adding the distortion, vignette, and film noise, and then, after saving the result, I tweak the colors and sharpness of the image. I’m trying not to overprocess my renders, but using Camera Raw is inevitable. I just like to emphasize the details even more.

Advice

I think the best advice is to be patient. Making realistic props is time-consuming but worth the labor. Definitely using a lot of references also helps, because it is just easier to visualize the prop and add realistic details (even when you think you know how the object looks). And also, just do it, if the first model you make isn’t that appealing, then collect all the mistakes you made and do it again. Consistency is the key.

Ewa Wierbik-Ziąbka, Senior 3D Artist 

Interview conducted by Theodore McKenzie