Industry Tools For Game Cinematics Pipeline

Life Updates

Hey! It’s great to be able to share more with you guys about what’s been happening. It has been very busy since the last article. 

Two major releases in fact! Both Forza Horizon 4 and The Division 2 have been challenging and worthwhile since there’s always so much to learn from the production itself, and everyone involved in it. I am certainly not at the level 80 yet but I can say I have leveled up for sure!

Besides, I now have the opportunity to not only work on environments featured in real-time context such as in-game, but also pre-rendered movies that allow for a more cinematic final product, a different way of looking at things, new techniques and tricks learned during the process, and a more direct look at the storytelling/narrative of the environment compared to simply dressing it up for it to look lived in. Both have their lessons to learn from and this knowledge results in a more complete understanding of how to create interesting visuals, so I love having the chance to work on something like this!

The more important additions to my artist arsenal have been a more in-depth understanding on how to serve the ultimate purpose of the environments I create, which is to not just follow a concept to build a 3D version of what has been drawn, but to work backwards from what the viewer is supposed to feel and think when experiencing this to what resources I can use to convey those (lighting, use of color palettes, camera framing, etc) and more concrete things such as: the state of the world this environment is in and therefore, its level of wear, what types of objects are in this area, the logic behind their placement, etc.

Also, the fact that I get to work in a team that has a lot of combined experience not just from games but from the feature film industry and is always more than happy to share that knowledge with you, makes projects even more interesting to work in, as you can absorb so many skills. Goes without saying but the best way to improve is to surround yourself with that kind of people!

Another addition is the inclusion of Houdini as one of my main programs used for a variety of tasks, including prototyping little tools and utilities, assembling scenes, managing shaders, and more. Even though I am only getting started with it, it gets you to think in a very different way than traditional 3D does and I would recommend exploring how it can fit into your workflow and help you out on a day to day basis. The steep learning curve is definitely worth pushing through.

Blender Advantages

I had been using Blender for a while now (2.64 around 2013, if memory serves) but had switched to different programs during the years (Maya, then Max, then MODO).

As with a lot of people, the promise of completely revamped UI and features in the big updated 2.8 build piqued my interest, and so I spent a few days toying around with it, customizing it and stress-testing its capabilities.

I have to say that almost instantly I knew that I was going to switch to it full time, including at work. The array of tasks I have to take care means that I have to switch programs very often, but Blender greatly reduced this, meaning I can focus on being creative instead of constantly breaking the concentration flow by switching from one program to another for modeling, sculpting, texturing, lighting lookdeving, etc. It’s all just there.

Another very important side of this is that for me (and anyone who is comfortable with modeling, shading, and general 3D production pipeline, who is not starting to learn 3D from scratch) it has made the creative process more fun than ever. I can have a pretty realistic preview of what I’m working as I do it, which allows for very fast exploration and iteration. My Director and I could sync up on the look we wanted to go for in a matter of hours, as opposed to weeks.

Furthermore, the awesome community has been putting out addons that I don’t think I could give up working with: HardOps, BoxCutter, Meshmachine, and many many more. Shoutout to them, absolutely titanic work.

Can Blender Be Used in Game Production?

I feel like the conversation about Blender not being production-ready is going to be changing very soon.

Probably, the main concern at a studio level has revolved around the fact that tools and utilities written for 3D software all artists are locked up to would not apply to other programs, and thus the time and resource cost of switching would not make sense for them.

However, since the release of 2.80, the program is getting more attention from people and investors that can easily change this (making great suggestions and writing add-ons in the first case, and helping out financially in the second).

The potential issues that could arise from using Blender in production are being discovered as people test it, then brought up and tackled at a very fast pace. As long as Blender foundation has the resources to quickly iterate on what’s being requested, there will be no competition in terms of 3D software (in my opinion).

Its versatility, great features not often found in similar software (e.g. that delicious viewport), huge and great community, very good dev support and the fact that it’s free, far outweighs most of the issues people have with it. It’s just too good not to use.

What Software To Pick?

Luckily for us, artists at Massive, we can use any 3D program we want. We have people using Blender, Max, Maya, even Lightwave! So there’s always discussion around the pros and cons of the different programs.

I think the main takeaway is that if you have a solid understanding of 3D production, the program you use really doesn’t matter. You know what you want to do. The difference is that tools might have slightly different names, menus are in different places, geometry, modifiers, object history may be handled differently. After having used different programs you learn what adapts to your way of thinking, but the core ideas are the same.

If I were to compare MODO and Max, I would say that since Max had been one of the industry standards for such a long time, the community and resources such as scripts and plugins are more varied than for MODO. However, the fact that it hasn’t really had too many meaningful updates for a long time, it means that it lacks features that others have implemented. I personally love the way MODO presents itself with a very clean, streamlined UI layout, the way hotkeys work, how it intuitively handles geometry in its scene outliner (in folders and item containers) and other features like MeshFusion, Round Edge shader, and very straightforward UV tools.

If instead, you are a new artist or a hobbyist looking around for the best software in absolute terms or to start with, my biggest piece of advice is to simply pick one, learn with it, and most importantly, don’t fall into the trap of sticking with it because you’re comfortable, or else you will be limiting yourself to the tools and way of approaching things you've learned, and will be missing out on A LOT of content that will make you grow as an artist.

Very often you will see people “fanboying/girling” online and arguing about the best software - I say that is valuable time you’re not spending on improving your art! I like to compare it to speaking languages: there is no better or worse language, and each culture has a different way of thinking and speaking about things. Knowing more means you “unlock” these fresh outlooks and add them to your pool of knowledge.

Before Starting a Large-Scale Environment

When it comes to massive-scale environments, it is very important to keep a very consistent design language across the scene, and be smart about how to use elements such as lighting, fog and composition to give variety in shapes and tonal values so that it doesn’t become a noisy mess of high-frequency detail everywhere you look. Your eyes need places to rest after all!

I had the chance to build the environment featured in The Division 2 game title’s sequence - a good example of a large-scale environment (see above). Notice how the different layers create depth and separate the image into values so the brain has an easier time simplifying what you see (lush green and blue summer sky against more muted, concrete-like colors of DC in decay).

Remember how the player/viewer is supposed to feel like when looking at this. What are the most important elements that contribute to the atmosphere that your Director wants to express? What elements take more space on the screen? Which objects are closer to the camera?

After figuring out what's important for your scene in terms of the art side of things, you need to break down your references and concepts into objects and textures that will need to be created. This is the technical side of things. You need to work out what is the most efficient way of getting everything done within the time limit. Of course, bigger objects, hero and close-up assets take priority, then modular shapes that can be repeated around to cover big chunks of space, and so on.

Being clever about reusing assets, breaking repetition in silhouettes and textures, finding creative solutions to performance budget constraints, and showing the interaction between different elements in the world and how they affect each other/are affected by inhabitants or forces of nature are key to produce high-quality huge-scale environments.

How Cinematics Are Produced

The process of creating cinematics can be different depending on who is producing it, but there are always common points that the pipeline goes through.

It always starts with the need to present part of a story, characters, places or feelings that would be less effective or not possible during the normal gameplay loop. It can also be used to abstract the player from their character for a bit, giving them some time to rest after a tough battle, to prepare before a big event, to reward them for taking the time to explore, and more.

From the information that needs to be presented to the player, a script is made that should contain the story, characters, locations, etc. The format is similar to a film screenplay and should make it easy for everyone involved to be on the same page about what the team should produce.

Once this script is approved and locked down, artists can begin storyboarding the sequence dialogues and situations, experimenting with what composition works best for the mood each shot/sequence should have. Once the shots have been chosen, an animated version of the storyboard is made, which consists of very simple animations to get a feel of how long each shot and the entire sequence are, how well shots flow into the next one, and how the characters move, so the animators can start thinking of how to approach the work.

Environment concepts are then produced from the most important shots in the storyboard, to have a more clear picture of how these should look and field. It's still a very experimental phase, and very important since it will define the look of the cinematic. The same is done for characters and special objects that might not exist in the regular game.

When there is a clear vision of the direction we are taking, a 3D environment blockout will be made, so we can introduce cameras, replicate what was designed in the storyboard, and see if some adjustments need to be made to maintain the shot composition in a 3D space. A refined version of this will also be used by the motion capture team and animators to add the first pass of character movement. Since there might be walls, doors, objects and other things that the actors lean on, interact with, carry around, etc, they need an already established 3D space to make their actions believable.

At this point, the animation side of production will keep iterating on the character acting using the 3D blockout that won’t drastically change, while the actual environment is being refined, built upon, textured and dressed up with all appropriate assets. Lighting is also explored at this stage but mostly set to the mood the concepts show.

Due to the nature of cinematics being experienced from a set camera, this creates a situation where, as opposed to an environment the player can roam and look at from all angles, the camera dictates a much narrower scope in terms of the content needed to produce. This results in not needing to spend time on things that won’t be seen directly such as the backside of buildings, for instance. It also means artists can put extra love on what we do see, elevating its quality.

From here it is a pretty straightforward process that requires regular iteration, feedback, polish and adding layers of detail to the scene.

Once all the content is there, the entire sequence is pushed to the render farm, and then comp and color grading stage, so that the final touches can be added.

Whilst this is happening, VFX that were nothing more than blockouts to this point are finalised and final renders are created, then all the grading work applied to them, and it’s final render time!

If the cinematic is real-time and not pre-rendered, then we don’t have to worry about render times, however, performance needs to be taken into account, so that all platforms can play the sequence properly.

The most important thing, in my opinion, is to be proactive experimenting with the elements you’re working with, whether that is environments, characters or anything else before they’re locked for production. You may find something that is better than originally thought, give more variations to your Director to choose from, and bump into the occasional happy accident!

Afterword

The fact that games keep growing in scope and actual size means that it’s not wise to throw more and more people at projects and expect that the results and quality delivered would be directly proportional.

In my opinion, non-destructive, procedural workflows and tools have gone from an interesting novelty to a necessity to accomplish ambitious and high-quality content. Not only that, but it keeps iteration time shorter and makes it easy for artists to make any change required, so they can focus on putting their creativity to good use!

I encourage any artist to delve into the technical side of our craft, that being coding (python as an example), Houdini or any others, because you can turn it into a powerful skill that will, in the long term, allow you to create solutions for yourself, help make certain tasks easier and faster, and make you a more complete artist.

Here at Massive, we are constantly improving our workflow and are encouraged to bring new ideas to the table, test them and maybe implement them, for the benefit of the entire team. We thrive as a family of artists, so I speak from experience when I say that the best way to improve is to put yourself in a similar situation where people will challenge each other to move forward.

Abel Dopazo, Cinematic Environment Artist at Ubisoft Massive

Interview conducted by Kirill Tokarev