Creating still images with Unreal Engine’s real time ray tracing – the end of V-Ray and Corona?

Is Unreal Engine’s real time ray tracing capable of replacing V-Ray and/or Corona when it comes to creating still images?

Is it possible to create high end quality still images with using Unreal Engine’s real time ray tracing only?

This is one of the hottest topics these days in the field of archviz, especially that RTRT became production ready and the latest series of RTX graphics cards are out there (mainly the 3080 and the 3090 but the new Ampere based RTX A6000 ‘monster’ – formerly known as Quadro – is worth mentioning as well).

In order to give you an answer to these questions, first, let’s see what the advantages and disadvantages are of using real time ray tracing for creating still images. This way you’ll have a deeper understanding of the answer given at the end of this article.

First, let’s start with the…

Pros

High end visual quality

My short answer to the second question above is: absolutely, you can achieve pretty high end visual quality with real time ray tracing, thanks to the physically accurate reflections that was a missing ‘feature’ of real time – up until now. 

And this is why it might be pretty competitive to V-Ray and Corona.

A pretty basic test scene that I did real quick – I used real time ray tracing and Photoshop to post process a bit

If you put two visuals next to each other (one is created with the Unreal Engine while the other one is rendered with either V-Ray or Corona), there might be some minor differences in GI, shadows and reflections. In addition to that, offline render engines tend to produce more accurate and clean renderings but the ‘laymans’ usually won’t notice these differences.

So if you are creating high end quality visuals for the ‘untrained eyes’, real time ray tracing might be a valid tool for you.

Let’s see the next huge benefit which is…

(kinda) no render time

Although both V-Ray and Corona offers GPU based real time rendering features, they are still much slower compared to real time ray tracing. You have to wait tens of seconds until you end up having a decent looking image (that is free of noise) from the actual camera angle, depending on your configuration. Additionally, if you go for a higher resolution, render time might even go up to minutes. 

In case of production quality, you will end up having hours of render time, depending on your hardware specs (mainly the CPU power) and resolution.

In the Unreal Engine however, at low to mid quality settings, you will have instant results. For example, if you put a new light source into the scene, you will see the light setup changing within milliseconds. If you want to capture mid size screenshots (up to around 3K), it will take only a couple seconds to do that. So RTRT is pretty fast and quite flexible, especially at draft settings.

Render time only appears when you generate really high (production) quality screenshots. In this case, you might have to wait even 5-10 minutes for the Movie Render Queue to capture a screenshot, depending on the resolution, your GPU specs and real time ray tracing settings (such as ‘Bounces’ and ‘Samples Per Pixel’). But I think it’s still pretty manageable compared to V-Ray and Corona.

Comparison between low and high quality settings in real time ray tracing (click to enlarge)

This leads to the next benefit which is…

Greater flexibility

Basically, your scene is updated instantly when you adjust any parameter (depending on the resolution and RTRT’s quality settings). 

This goes for mainly the lights: you can fine tune the entire light setup in real time by placing the light sources wherever you want and by tweaking any of their parameters. You’ll see instant results.

It gives architects, interior designers and CGI or archviz studios greater and more spectacular interactivity when they are checking with their clients. Because of the instant results, live sessions become much more productive and efficient.

Real time ray tracing might also help you with having a…

More simple workflow

From an artist and creator standpoint, the whole pipeline (from modeling to post production in the Unreal Engine) becomes more simple for real time ray tracing as you are free to skip certain steps.

First, you don’t have to deal with unwrapping as there are no lightmaps – that’s definitely a huge benefit. 

Second, for this reason, you can use objects with higher poly counts and you don’t have to prepare those 3d models in order to unwrap them properly. That’s another great ease!

Third, you don’t have to pay attention to texture sizes. Basically, you can import whatever maps you want.

If you have created any content in Unreal that included lightmap baking, you pretty much know how skipping these steps affects the overall workflow.

Let’s move on!

Easy to create animation from

If you set everything up and you’re ready to generate still images, you will be just a couple steps away from creating animations and short films for your archviz scene. 

You just have to put some cinematic cameras in there and set up their paths and parameters. After that, you are ready to start capturing the frames – within a couple hours (depending on the specs), instead of weeks.

That’s another huge benefit of real time ray tracing, compared to offline render engines.

But real time ray tracing has a huge business related benefit as well…

It has high perceived value

Although that’s not really a practical value but for a lot of users, architects, interior designers, CGI and archviz studios (even for filmmakers), it’s quite important to use cutting edge technologies in their businesses as they demonstrate high perceived value. 

Trying out and experimenting with new things to make our products and services better, more efficient, more valuable, equals innovation. According to many marketeers, innovation is absolutely a must for businesses. Innovation enables us to get ahead of the competition and to charge more.

Okay, with that being said, let’s see the ‘dark side’ of using real time ray tracing for creating still images!

I'm on Facebook!

I post smaller piece of content about real time technology and archviz that might be really helpful for you:

Click here to check out my Facebook Page!

Cons

Hardware limitations - the great kryptonite

This is still one of the biggest downfalls of using real time technologies for archviz (and for filmmaking), especially when it comes to real time ray tracing. 

Static light setup and screen space reflections are relatively cheap on the computers’ resources (meaning, when you optimize your scene properly, it can even be run on mid range computers without any problems). In general, there’s no light and reflection related calculations going on in real time – both lightmaps and SSRs are pre-rendered.

But that’s not the case with real time ray tracing.

Global illumination and (physically accurate) reflections are calculated in real time, thus RTRT is much heavier on the GPU. Not to mention that you need to have an RTX graphics card – although the GTX series support real time ray tracing as well, these cards are much slower as they don’t have any RT cores.

In addition to that, the amount of GPU memory (VRAM) tends to be a bottleneck as well – especially when you capture screenshots: it’s relatively easy to exceed 6GB or even 8GB of VRAM. It means you shouldn’t go below the RTX 2070 if you’d like to work with Unreal Engine’s real time ray tracing. 

Night version of the test scene – unfortunately, you need a high end GPU to work with real time ray tracing

Good news is that the RTX 30 series cards will be equipped with sufficient amount of VRAM. Even the RTX 3060 will come with 8GB of VRAM, and according to Videocardz.com, the RTX 3070 and 3080 will be available with two different VRAM sizes – 8/16GB for the RTX 3070 and 10/20GB for the RTX 3080 – which is pretty awesome!

Let’s see the other downfall of using real time ray tracing for still images, which is…

Another workflow to learn

Learning and adopting new workflows is usually another huge downfall when it comes to real time. 

In general, architects, interior designers, CGI or archviz studios and freelancers have their well-tried pipelines, tricks and techniques about creating visuals. If they want to learn new ways to create and other workflows, it will usually require them a lot of time, energy and money to do that – it tends to be such a huge investment that they don’t really want to take the challenge.

Even if you learnt how to create visuals using a static light setup in Unreal, it takes weeks to explore real time ray tracing and to experiment with it until you produce a similar visual quality that is very close to static.

The next one is…

You might be far away from static light setup

It’s tempting to believe that if you set up your scene for real time ray tracing, it will automatically mean that you are just a few clicks away from creating nice looking static environments.

Unfortunately, that’s not the case. In fact, you might be far away from it – remember when I said your workflow becomes more simple for real time ray tracing? 

Well, you will need the steps you skip for RTRT as they are essential for a properly working static scenario. You will still have to deal with optimizing object count and poly count, unwrapping static meshes and optimizing textures. 

In addition to that, you will have to build up the light setup in a completely different way. The lights you set up for real time ray tracing is not ‘transferable’ to a static environment.

Difference in post process

You will have to take a different approach for RTRT images when it comes to post processing – mainly because the Unreal Engine generates different render passes (or elements) compared to V-Ray or Corona.

Although I wouldn’t consider it as a huge issue but I know a lot of 3d artists like to work using render passes (or elements). It’s their default workflow to enhance the images during post processing. 

So for them, it might be a challenge to adjust the pipeline accordingly. For anyone else, it won’t be a noticable difference.

The conclusion

Okay, by now, you probably have a firm understanding of the advantages and disadvantages of using real time ray tracing for still images. So it’s time to answer the first question I asked at the start of this article.

I think RTRT won’t replace V-Ray or Corona but it will definitely have its place when it comes to creating still images for your projects. 

I’d say the following:

If you don’t need to spend months to implement a new workflow in your business (meaning you have experience in using the Unreal Engine for archviz) and it’s very important for you to harvest RTRT’s benefits (such as no render times, instant results and greater flexibility during the design phase of your project), you might want to spend some time with real time ray tracing as it may help you a lot in the future.

So that’s all for now!

Agree? Disagree? Have any suggestions?

If you have something to say, feel free to drop a message in the comment section below!

If you think this article might be helpful to others, please share!

What is 'real time' archviz?

What Are The Key Benefits Of Using It for your business?

‘Real time’ technology is undeniably here. It has become, and will be more and more affordable in the near future.

If you are open to new innovations in your business, have heard about ‘real time’ archviz but not sure what it is and how it might be helpful for your business, this free ebook is for you.

Click here to download the FREE pdf ebook!

andrasronai_profile_2020

About the author, Andras RONAI

I help people with unleashing and bringing their visions into life with real time technology (using Unreal Engine) so they can make this world a bit of a better place by creating something great.

Click here to read my full story!

Leave a Reply