Old TV Camera – Prop Breakdown – Petteri Torvinen
Hi everyone, my name is Petteri Torvinen. I am a Freelance Hard Surface Artist from Finland. I was introduced to 3d graphics at the university, where I studied 3d modeling, 3d animation, and 3d graphics programming. After graduating from university I worked several years as a 3d modeler in different 3d research projects and in product visualization. About three years ago I found Game School Online and started learning how to make 3d game assets along with some other game dev-related stuff.
Before this project, all the game assets in my portfolio had been weapons or other military-related objects. So I decided to make something else this time. It is much easier to model and texture something that you have seen in your own eyes, so I decided to visit a local museum to see if there was some nice old gadget that I could choose as my next portfolio piece.
The object that appealed to me most at the museum was an old tv camera that was used by Finnish Television in the 50’s & 60’s. It had lots of parts and little details and many different materials. What also helped me choose the old camera was the fact that my father had worked as a tv cameraman for over 40 years. After checking out that there wasn’t other similar artwork in Artstation I made the final decision.
I took about 150 photos of the tv camera at the museum trying to catch every detail I could for both modeling and texturing. However, the camera at the museum didn’t have all the parts I wanted to model like the cable so I had to search for some pictures from the internet too. For texturing it is also good to get some additional references from the web.
I like to use ACDSee Photo Studio as my photo browser and PureRef for reference boards.
Blockout modeling is not one of my favorite parts of the game asset pipeline, but really important in my opinion. The main purpose of the blockout to me is to figure out the proportions and dimensions of different parts. In my experience, it saves time to get those right at this early stage of the project.
After the blockout, I always make the high poly model. Most of the high poly modeling is done with basic sub-d modeling in 3ds Max. In sub-d modeling, I like to use manual support loops but sometimes use the chamfer modifier or creases too. A few more complicated parts I exported to Zbrush and used live booleans, dynamesh, polish crisp edges, and decimate to make the high poly.
The only parts where I used Zbrush sculpting were the wooden legs, where I added some edge damages and some organic shape to the end parts of the legs.
The next stage of the process was to make the low poly model. It was mostly about making copies of the high poly parts and deleting every vertex that didn’t affect the silhouette. I also knew that I wanted to render close-ups of the tv camera so I had to add some geometry to the corners/curves to prevent a faceted low poly look. What I learned at this stage was that to save time I should have planned my low poly model better at the high poly modeling phase. In the case of a simpler model, this would not have been so important but this camera has over 400 parts of which about 250 are unique, so planning ahead better would have saved me a lot of time.
For the UVs I made sure to straighten the islands as much as possible and leave enough space between them to prevent bleeding. I used two 4k UV maps, one for the camera and one for the stand. The UV islands of the wooden parts were rotated so that the wood grain would go in the right direction in all of them in the texturing.
It is always important to save as much UV space as possible. The camera stand is a tripod, so I made unique UVs to only one camera stand leg and stacked the other two legs UV islands over those. I also tiled the UVs of the cable.
The way I like to make UVs is to first assign smoothing groups to my low poly parts in 3ds Max. Then I use the ”Flatten by Smoothing Group” function of Max and export the model to RizomUV where I finish the process. In my experience, RizomUV is by far the best solution for UV unwrapping.
As a result, I got a texel density of 21 to all my UV islands. To further optimize the UV space I made the UV islands of camera parts that would be more or less hidden smaller. I gave them a texel density of 16. Also, I knew that there would be really small text on the camera objectives and decided to give those UV islands more space and texel density of 26 to ensure the small text would not be too blurry.
Marmoset Toolbag is my number one baking solution. To be able to use baking groups and to adjust the cage and the skew is great and saves time and effort. For this project, I baked the normal, ambient occlusion, and color id maps in Marmoset. The rest of the maps I baked in Substance Painter. I baked all maps in 4k and used 16-bit bitmaps.
I began the texturing part in Substance Painter by making all the height details and exporting and re-importing the normal map after that. This made sure I didn’t have to use too many anchors in the texturing phase to make the generators recognize the added height details.
The camera has almost 50 different materials, so the texturing was a lot of work. I decided to use the ACES color profile I found in Substance Share to make sure the textures would look the same both in Substance Painter and in Marmoset Toolbag when doing the final renderings.
One of the hardest parts was the edge wear. There was lots of it and there was no way I could get the right look using generators. Most of the edge wear is hand-painted using some nice alphas I have. I also made some alphas/textures from the photos I took to be used as alphas/stencils. This gave the texturing some authentic look. I used Photoshop to make those and also the many decal alphas that were used.
In the wooden parts, I took some artistic freedom since I didn’t quite like the look of the original wood. In general, reality is not always super interesting and for in-game assets, it’s sometimes good to exaggerate things a bit to make assets look more appealing.
After I was done with texturing I added a ”Final Adjustments” folder at the top of the stack and set the blending mode to ”passthrough”. In that folder, I added hsl_perceptive, contrast_luminosity, and sharpen filters to make some final tweaks to the texturing.
Adding story elements
At some point in the middle of the texturing process, I thought this piece of art needed some story elements to make it more interesting. I didn’t want it to look like an ordinary product rendering. My father(ex tv camera operator) told me that they used to keep a roll of tape hanging on the pan bar of the camera. I also added a second number plate with a coffee cup stain on top of the camera.
I like to make the portfolio renderings of my game assets in Marmoset Toolbag. The first thing after importing the models and textures is setting up the cameras for each shot. I tried to figure out camera angles that would look interesting and would show all details of the model and texturing as best as possible. After that, I tweaked the render settings and start lighting each shot. I used a grayish indoor HDRI as the main light to prevent coloring the model too much. Then added lights to highlight some parts that were too dark, to show off the roughness variation, and to create some rim lighting. Depending on the shot I used 4-11 additional lights. I increased the diameter of many lights to avoid too harsh shadows.
In the rendering, I used ACES tone mapping and a field of view of 23. I didn’t use any effects. I rendered the shots with alpha and added the background and some final adjustments like brightness/contrast in Photoshop.
Asking for feedback during the asset creation process is really important. I have a habit of showing my work in a Discord group at the end of every phase. Showing your work for critique might be a bit intimidating at first but you usually get some really good hints and tips to make your work look better.
The first time I asked for advice was after finalizing my blockout model. I wasn’t too happy with the way my camera cable looked. The feedback I got was that the camera cable didn’t look very interesting and that it looked a bit too thick. I made corrections to the cable in the high poly modeling phase.
In the end of the texturing process, I got some very good tips too. The advice was to emphasize the color variation of the camera’s main body with contrast/luminosity filter and to add some subtle hue variation to the different parts of the camera’s main body. These kinds of easy tricks at the end of the texturing can make a big difference and make a ”dull” looking asset look good.
Big thanks to the Games Artist crew for giving me the opportunity to tell about my artwork. I hope this article will help somebody in his/her asset creation process somehow. Also thanks to the people of Game School Online for critique/feedback.
Feel free to visit my Old TV Camera Artstation page to see more renders: https://www.artstation.com/artwork/XnVRnw
To see my other artwork, visit here: https://www.artstation.com/petterit
Thanks to Petteri for allowing us to have such an in-depth look at his process. If you liked this prop breakdown and want to see more like it from other inspiring artist’s make sure to follow us on :