Unreal Engine for Archviz Animations: Pros and Cons

András Rónai is an architect and 3d artist, UE4 specialist from Budapest, Hungary. He helps architects and archviz studios with creating high-end quality architectural visuals using UE4, from engine compatible modeling to walkthrough scenes and cinematics.

This article appeared originally on Andras Ronai website. Visit it to download free ebook "A33 Office Archviz Project Using The Unreal Engine 4 - Case Study"

Okay, so I got this question many times from my fellow 3d artists: is it worth creating archviz animations using the Unreal Engine 4?

Despite the fact that I have created two short films over the last year with this engine (both were paid projects), I started searching for articles to find out what my colleagues think of this question.

I was suprised a bit because there wasn’t any articles on this topic (or I just didn’t find them).

According to the survey results of CGarchitect, Unreal Engine is becoming the fourth most commonly used render engine for archviz after V-Ray, V-Ray RT and Corona. So it’s worth discussing this topic as Unreal might play an even more important role in the near future.

Luckily enough, I have managed to get some experience with UE4 for animation purposes. So my answer to the question is: it depends. There’s a couple things you should consider before you decide to use Unreal for your next project.

In this article, I will present my own personal perspective on this issue. I’m going to make a comparison between the two technologies: the Unreal Engine vs. frame rendering (such as creating animations in V-Ray or other render engines).

I hope you’ll have some insights after reading this article. My goal here is to make this topic clear so you’ll be able to make the best decision when it comes to chose between the two technologies for your next project.

So let’s start with the…

Pros

1) The absence of render times (broadly speaking)

Having insignificant amount of render times are one of the biggest advantages of using UE4 for animation purposes. It has a lot of great consequences.

Render times, because there are two rendering processes going on when you create animations. Insignificant, because they are measured in hours only. But to make my point clear, let’s go a deeper here.

Unreal supports three different light setup mode:

1) The first method is using static lights

Basically, you bake the lights. In this case lightmaps are generated for every object in your scene.

This is the first rendering process. With Swarm, you can have multiple computers rendering your scene. My most complex projects took only 18-20 hours at most on two computers so this process requires way less time than frame rendering. (Now imagine what 10 PCs that are equipped with Intel Xeons would do…)

2) The second method is using real-time lights

In this case there are no lightmaps in your scene: lights are calculated real-time by the GPU. In the end, you don’t have to deal with any render time regarding to light calculation.

3) The third one is to use both of them at the same time

This is a hybrid method.

I mentioned before that there are two seperate rendering processes when you are dealing with projects in UE4. The second one is exporting the animation sequences into movie clips. Good news: it takes only minutes to render them out, so it’s basically nothing compared to frame rendering.

2) It doesn’t matter how long your animation is

With Unreal, you don’t render any animation frame by frame. Your lights are baked into the scene (or calculated real-time) so the only rendering process you have is exporting shots into movie clips.

For this reason, there is no difference between a 30 second or a 10 minute long animation in terms of render time. That’s one another huge advantage compared to frame rendering.

3) Modifying camera paths are totally flexible

Thanks to the technology, you can modify camera paths (basically) whenever you want. Even if you are in the video editing phase, it is still possible to make some tweaks in a fairly reasonable time – if that’s what your client wants.

4) It might take less time to conclude projects

The absence of render times has an awesome ‘side effect’. My experience is that it’s possible to finish animations faster than using frame rendering. It can be measured in days if not weeks.

Undertaking projects with shorter deadlines might give you a competitive edge over other studios, especially in the archviz business.

5) Visual quality comes near to V-Ray and Corona

The Unreal Engine certainly has the potential to create high-end visual quality content with it that might reach the quality level of V-Ray or other render engines. You just need to have the know-how to do so.

My experience is that the untrained eye can’t make any distinctions between the two methods (game engine vs. frame rendering). So even if V-Ray is slightly better in terms of quality, it’s only you and other fellow 3d artists who will notice any difference.

Here’s an example of an architectural animation that I created in Unreal with NODE Visual (my former archviz studio):

6) It might reduce your costs

In case of huge and complex projects, sometimes you just can’t avoid using render farms for the rendering process. This part can be completely eliminated from your workflow and you don’t have to spend hundreds of bucks for such services.

7) It’s easy to go up to 4K and 60fps

Thanks to the technology, it doesn’t really matter what resolution and FPS rate you choose for your animations. The difference between exporting movie clips in full HD @30fps and 4K @60fps is measured in minutes only.

The cool thing is that the Unreal Engine even supports 8K and 120fps (that’s good news if you are creating VR content as well).

But you should always keep in mind that choosing 4K requires using high resolution textures to achieve a decent visual quality.

8) You can use up a content for multiple times

If you decide to use the Unreal Engine for your production, it requires a relatively small investment to create walkthrough and VR ready scenes out of your content that you created for the animation. You might want to give it a thought as it might be quite beneficial businesswise.

9) It makes the design process more flexible

Architectural design is an iterative process. It’s never finished, rather just ended due to the deadline. Architects often modify the hell out of the projects which can be pretty challenging for 3d artists.

Animations tend to be much more ‘rigid’ than still images, in terms of changing different parts of the content. You might need to finish the modeling phase a couple weeks before the final deadline. But the abscence of the rendering process might free up some time for last minute modifications.

Okay, now you know why it’s a good idea to use the Unreal Engine for animation purposes. Let’s take a look at the downsides of it!

Cons

1) It requires a different approach and a more complex workflow

You have to do things differently to create content for the Unreal Engine. And it might be a challenge.

First, you have to use 2 softwares, not just one. It means that you have to export the model from your 3d application into the Editor. This process usually results in making your workflow more complex. Especially, if you have to reimport certain parts of the entire model many times.

Second, you have to make your model and textures compatible with the Unreal Engine. Creating content for game engines is different than for frame rendering. Paying attention to poly and object count, slicing up objects for proper lightmaps, unwrapping objects will definitely make your workflow more complex.

Third, if you have to create still images from the same content, you have to prepare the model for both still image rendering and for Unreal as well. And it requires more time to do so.

2) You need expertise to create high-end quality content

The learning curve for the Unreal Engine is pretty steep. It means you have to invest a lot of time to learn how to use it well. This is one of the biggest downsides of using UE4.

For example, it took me about 6 months to achieve a decent visual quality. My first results were pretty bad. I didn’t really know how to use the Editor properly although I had years of experience in creating content for game engines. Furthermore, at that time, there wasn’t too many tutorials on how to use Unreal for archviz purposes (I think it’s still an issue).

After 6 months of neverending practicing and learning, I managed to create the A33 Office that was a milestone for me. But it contained lots of errors and I did things pretty wrong in it. It took me about 3 additional projects to learn how to use it more efficient in terms of visual quality, creation time and optimizing scenes and resources.

Maybe the steep learning curve is the reason why only a few 3d artists are able to create high-end quality archviz content with Unreal, compared to V-Ray.

3) It might be risky to use without know-how

Sometimes we run into technical problems during a project that we need to solve. Although Unreal has an extensive documentation and a great AnswerHub support, you might face problems that take up a lot of time to resolve. If you’re not very experienced with this technology, it might be pretty dangerous in case of paid projects and strict deadlines.

4) It requires (a powerful) gaming graphics card

As a 3d artist, you’re probably equipped with powerful hardwares: fast CPU, 32GB or 64GB of RAM and SSD.

Unreal Engine requires a high-end graphics card as for archviz purposes we tend to use lots of high-poly objects and high resolution images for the textures. It makes things ‘worse’ if you use movable light sources that are calculated real-time. Even if you’re working with static lights, changing one parameter of your lights might slow down your viewport a lot. Working at 10-15 FPS rate is no fun, it’s rather stressful (I’ve been there).

If you plan to upgrade your graphics card in the near future, stick with GeForce. You could use a Quadro, but it’s simply not designed for gaming purposes and you won’t get better performance for your money. I think purchasing a GTX1060 is a good starting point, but you should go for the GTX1070 or GTX1080 if your budget allows it. The better GPU you have, the smoother your workflow will be.

The conclusion

Okay, here’s the bottom line.

Thanks to the technology, it is possible to make the whole creation process much more flexible with Unreal (in terms of animation length, camera paths, resolution and FPS rate) than with the frame rendering method. All of this with a visual quality that comes near to V-Ray.

But you have to pay the price for that. You need 1) expertise to achieve the desired visual quality, and 2) experience to deal with a different workflow and any technical challenges that might come along. You need to invest a lot into developing your skills with Unreal.

If you have created a couple test projects so far and you managed to achieve a great visual quality, it’s a good starting point. If you feel you can handle projects in Unreal in a reasonable amount of time, then it might be worth creating your next project using UE4.

I think creating archviz content with Unreal should be considered as a seperate branch of architectural visualization as it requires different skillsets.