3dxnixah - DLSS | Neural Rendering - Seline

I’ve prepared a few renders of Seline for you here. In the post below, you’ll also see Gwen.

Simply put: DLSS is an AI-powered rendering technology from NVIDIA. It can analyze an existing 3D image and use additional image information during rendering to make the final result look more realistic. DLSS brings out facial features particularly well. It helps the overall appearance of the characters come through much more clearly.

In my projects, this does not mean that an AI suddenly creates my images or animations from a text prompt. That is not the case here. The scenes, characters, poses, animations, camera work, and lighting are still created normally within my 3D workflow. The AI only comes into play during the rendering process, where it supports the visual presentation.

Neural Rendering [NR] mainly affects lighting, skin, materials, and fine details. This can make the finished image look more natural and, in some cases, significantly more realistic.

Getting everything to work properly required quite a bit of work. I built a new DLSS rendering workflow and spent a long time testing and refining it. Over the past few weeks, I’ve created several 4K renders, tested different lighting setups, and have now also rendered my first animations with it. In my opinion, the results are definitely worth seeing.

Some adjustments were also necessary for the characters themselves, especially their skin, because it can look quite different under Neural Rendering compared to my previous renders. So far, Seline and Gwen are around 90% finished. You could say: this is probably quite close to what Seline and Gwen would look like in real life.

I’m of course aware that this more realistic look may not appeal to everyone, because it differs quite a bit from my usual comic style. Still, I want to keep up with new technology, try new things, and continue improving my own skills. For me, that also means exploring new possibilities and seeing how far I can continue to improve the quality of my projects.

I also want to show you a small comparison: a facial image of Gwen and Seline from my early days and one from today. Especially in the faces, you can clearly see how much my style, skin rendering, lighting, and overall image quality have changed over the years.

With [NR] Neural Rendering, I feel like I’m taking the next step in that evolution.

My next project will be the first one where I plan to use DLSS and Neural Rendering throughout the production. The tests so far have been very promising, but testing individual renders and animations is one thing - producing an entire project with a new workflow is another. So I can’t say yet that everything will work perfectly from start to finish. There may still be things I need to adjust, improve, or learn along the way. But that’s also part of trying something new. I’d rather take the next step, learn from the experience, and see what is actually possible than simply stay with what I already know.

I’d really love to hear your opinion: What do you think of this more realistic [NR] look? Feel free to let me know what you think in the comments.

From now on, you’ll see the abbreviation [NR] in the filenames of my new videos and images. It simply means: This render was created using Neural Rendering.Seline - Mini Image Set [NR] 10-2026

Sunny