TLOU INSPIRED DIORAMA

TLOU INSPIRED DIORAMA

CONCEPT

It’s been a time since the last breakdown but never say never.

So…..It all started with the concept art above and with the desire to make a big environment using UE5 and its new features like Lumen and Nanite. But before I would embark on a personal project of such scale, I decided to enroll to CGMA’s Intro to UE5 with Chris Flynn (Dekogon Studios) in order to fine tune my skills in UE5. Furthermore, upgrading my skills with UE5 knowledge will also be helpful for MADCHAMO’s 3d services.

BLOCKOUT

After gathering a TON of references and shaping the idea in my head it was time to block out the scene. In this phase, like on every project I did, it’s very important to know from the very beginning exactly what you want to create. Gathering references and brainstorming the idea it’s the most impacting step which will have either a negative or positive influence on your project. We all know that failing to plan is actually planning to fail. So, if you fail to gather enough REF you will find it extremely difficult to block out the scene, because REF photos are supposed to art direct you. In the end, your art will be as good as your references, so make sure to pick not only informative stock photos, but also a similar project that you are trying to achieve.

So I gathered REFs and I had final image that I wanted to achieve in my head. Crystal clear, but there is a catch and many 3d artists tend to fall into the same pit like I did.

Normally, all the geo from the block out is throw-away geo, so it’s very tempting to keep adding different assets, to set dress the scene to match the reference and easily you step into the danger zone where you bite more than you can chew, meaning that you fill the scene with too many distinctive assets that would require too much time to finish. It’s very important to set a personal deadline because otherwise, you will never finish the environment. There will always be something that has to be modified and refined.

So, the golden rule is to reuse and recycle assets to match the REF. Try to achieve the same result but with feuer assets. This way, not only that you free yourself a TON of time of modeling, UV, texturing…….but you will avoid “project fatigue”. If you work too much on something, eventually you begin to get tired and lose interest.

With this diorama I have been in the situation where I went a bit crazy with the number of assets, in the end, having to delete some of them and use the recycle and reuse strategy. The CGMA course was 6 weeks in total, so I had 6 weeks to finish the environment that I planed, but I failed miserably at this stage because, yea….too many assets and I wanted to model and texture everything. Lesson learned I guess.

ASSETS

Moving on, I had all the assets ready to go into Unreal, to recreate the Maya blockout. It would have been a lot better to start the blocking phase inside Unreal and avoid this importing and recreating step.

After that, the fun part begins. UVs. For this step, the best advice I can give is to arm yourself with plenty of patience and coffee, at least, this was my main workflow. I UVd all the assets in Rizom which was an absolute blast. Many tend to treat this step superficially and try to get by it very quickly, but trust me, a good looking 3d model with bad UVs it’s a total no go.

For the ground, I wanted to have concrete plates that have been worn out or partially destroyed, giving the environment should be in the universe of The Last of Us, but at the same time I wanted to achieve a certain level of flexibility. So I sculpted a concrete base in Zbrush and also manually divide it into blocks along the sculpted cracks, to increase the number of combinations possible, when placing this asset into the environment. This way, you can achieve many variations using a limited number of assets. This asset pack will be featured on Artstation Marketplace and also on the future Madchamo 3d store.

I didn’t had modularity in mind for the Wall asset and I sculpted it as a single mesh with high details. Taking advantage of Nanite, I wanted to push the system as much as I can. In the end I would still recommend to decimate your meshes before importing into UE5. 300k – 500k polys is kind of the perfect spot to be with assets like this.

TEXTURING

For texturing I used Mari for the Dodge Charger and Substance suite for the rest of the assets. For texel density I decided to go off the rails a bit and use a TD of 40.96 to give me enough texture space for details and resolution quality.

To give more flexibility, for some of the assets like the shelves I had modularity in mind. This gives a lot of freedom in creating and shaping up the post-apocalyptic garage diorama that I was aiming for. Nothing in the real world is perfectly placed and aligned. Every object has a story on how it got there and what has happened to it when it got there. In such environments, rust plays a special role in setting the general tone of colors and composition. Almost everything made of metal can’t avoid the rust building up, especially when it’s exposed for a long time to rain, even if it’s coated with paint. The paint starts to crack and leave the metal exposed to natural elements.

When texturing the assets I wanted to avoid relying only on Substance procedural masks and instead I opted more for hand painting details and a lot of photo projections. I used procedural masks as a base layer on which I manually painted/masked effects like dust, wear, rust and many more. Texture projection, I find it as a must, when it comes to texturing and asset that has to look in a certain realistic way, because no matter what generators you use and combine or how detailed your manual paint can be, it cannot be as good as a real life photo of the asset that you are creating. The amount of detail and variation is unmatched. So I carefully projected textures or certain details from ref photos and tried to blend them as best as possible. As a general advice, the pictures that you want to use with projection workflow have to be high res. For example I took my camera and took some 6k ref shots of rusted sheet metals and used them. Worked perfectly well.

Working on texturing the assets was a bit on and off, as I was constantly shifting my attention to the scene in Unreal as sometimes I was a bit tired of working only on textures. A good method of not losing motivation when working on bigger projects is to switch tasks to avoid burnout. Repetitive task like UV’ing and sometimes texturing and asset for days or weeks can have an impact on your performance and motivation. The best strategy that I found was to switch the task, to start working on something else and then come back and finish the asset. Often times, taking a pause and revisiting a work-in-progress asset gives you a new fresh start and really helps on finishing it. I call this strategy pause & comeback and has helped me maintained a constant pace when working on projects.

THE SCENE

This environment will be in the universe of TLOU and the general idea/story behind it is that Joe (our fictional character) was working on his charger in his garage, minutes before the outbreak. He had to panicky leave his workshop and his work and run. After 10 or 20 years, nothing much has changed in our little workshop, but time began to show its mark on all the objects and materials inside. Green elements started to take over, as the garage door was left open. You have to invest in the story as much as you invest into creating the whole environment because it will guide your whole design process and the end result.

LIGHTING

Lighting plays a huge role. I consider lighting the most important factor in any artwork that I do. As I was playing a little bit with lumen , it was a bit tempting to put more lights in the scene because it was looking good and obviously because of the real time performance of lumen. BUT, using my strategy of pause & comeback, I felt like something was wrong and my lighting was not on the same page with the story. In such environments, most of the lighting comes from sunlight and the interiors are mostly lit with indirect lighting, since electricity systems are down. So I tried to approach everything focusing more on the story. After deleting the additional lights, I had a problem that my diorama was poorly lit, mostly dark and I wanted to showcase all the assets that I modeled and textured. What I did? I just went further with the story and decided to collapse the ceiling, to let more light through the cracks and openings. I used a custom HDRi that I made to capture nice reflections and a directional light. But still, some of the assets or corners were too dark and lack detail. The secret ingredient here was to place spotlights, aimed at areas of interest, with disabled shadows, in order to art-direct what the viewer sees, which areas have points of interest. I tend to use light as a guide, to guide the viewer where I want.

HERO ASSET

Obviously, the center of attention was the Charger, on which I spent a considerable amount of time. I used Mari on it because I wanted to manually paint details with 8K resolution of textures for the final rendering and I find Mari a bit better when it comes to high res painting, compared to Substance. I already had this car modeled and UV’d as part of another project for the 3D shop and marketplace. One thing to mention is that I didn’t need the interior, so I deleted everything which was not visible for performance and made a temporary opaque glass material with much dust on it.

POST

One of the many mistakes that I did in the past was to leave my renderings untouched after the final rendering, assuming that the work was done. In reality, post-production represent the other 50% side of the project. For color grading and general post-work I used Davinci Resolve, Fusion and Photoshop. I opted for a cinematic presentation of the project and for this I used 80mm lense for rendering. I wanted the renderings to be up close & personal. For post-fx I added film grain, chromatic aberration and some lense dirt. Post work can make all the difference between a good quality image and one that just looks professional and great. It has that “wow” effect. For this step I spent quite some time reading cinema techniques on books online. Below you can see the difference. So, as advice, do not skip this step, it is as important as the whole project itself.

FINAL RESULT

It’s very difficult to call a project done, but as a general rule: A finished project it’s much better than an unfinished one that is just collecting digital dust on a forgotten folder. Thank you very much for reading and share with others. I hope that I have inspired you into creating something awesome.

You can see more high resolution renders on Artstation, just search for Madchamo. :)

For more breakdowns like this you can visit my website www.madchamo.com.