Awesome breakdown. Good designers often have similar processes.
Found it here: https://exoside.com/quadremesher/, just in case anyone else is looking for it.
The link at the end is pointing back to the article. Couldn't find the Quad Remesher and I would really love to test it.
We’ve talked with Gamera Interactive about the tech they are doing prototyping, working with environments and assets for an innovative RPG Alaloth.
Gamera Interactive is up & running since September 2016. The studio has been founded by industry veterans who spent the last 15 years working with big publishers (Midway, Eidos, Bethesda among all) and development studios (Milestone). Alberto Belli is the CEO of the company and his portfolio includes games such as Kane & Lynch, Stranglehold, Fallout 3, Age of Conan, Mortal Kombat and others, ranging from production to marketing, PR, licensing and business development. He decided to go indie at some point and Alaloth was something he was thinking about for a long time. The chance to get access to tools that years ago were not accessible for a small team, but long sessions of market analysis convinced us it was a good moment to go on right now. Anyway, Gamera Interactive released two console games in just 15 months (Unit 4 and Fearful Symmetry & The Cursed Prince on Xbox One and Steam) and we have two new releases for 2018. Meanwhile, we ended the pre-production of Alaloth which stays as our main project for years to come. The studio has been built around a delivery-oriented philosophy.
Our pipeline is very efficient, and this is the reason why we had a good pre-production in the back. Pre-prod is when you can test processes and make mistakes. No turning back once you’ve entered production. So, basically, we have graphic assets/environments/models and everything world-related that moves independently with its own scheduling. Then we have gameplay which means iterating on daily basis on skills and combos and testing IA which is something completely separated in the flow. When a new environment is ready, a new skill is implemented or a new enemy is working, we know for sure they are going to work in the merge 90% of times. Once we’ve done with this we proceed to check any possible problem in terms of look & feel because anyway there is no way to fix stuff (or even notice problem) without seeing everything running in a game-ready shape. There are things that work perfectly alone but could not fit with the general vision of the game. So, no blocky environments or characters. Everything is nearly final when it’s put in the project, checked, iterated, eventually adjusted.
All the environments in Alaloth have their own sketch and key art, made using references put in a mood board built during the pre-prod. With the key art ready, 3D scene is built following world lore was written by Chris Avellone and the design bible with all the levels detailed. Whit the 3D ready, we apply the iso camera jumping in with a char to check any problem with scripts and eventually fix it. The whole scene has to be adapted for the camera of course, so we take some time to change details if something is missed with the iso-view or to facilitate action in the area itself. So, we record a sort of mini-run with a placeholder char and if everything is fine we move on editing the scene, adding new assets (sometimes the navmesh needs to be fixed, textures tweaked, proportions corrected and so on). Then we add lighting to fit the mood and the first in-game integration. This is when we check for small imperfections or something we could have missed or decide to add new ad-hoc 3D stuff to fill the scene perfectly. Once we’ve done with all of this, we are ready for render (RGP, normal, ambient occlusion, depth) and overpaint phase, where we add the lighting effects accordingly to the daytime version we are working on and all the little custom details that really make the difference when it’s up to create a deeply characterized and living world.
The whole development pipeline is completely modular. The game itself, by design, is a huge combat system but everything around that could be defined modular and the same is for any asset we use. We have 4 different kingdoms and races with different styles in terms of environments, dresses, props and so on. We have a Layer 1 of contents, let’s say “Tavern” which is created with a couple of variations in terms of design but it’s basically the same for each race. We have tons of props we can switch transforming the “” in “the elven tavern” creating a brand new amazing env using the appropriate assets in the right place. Even the mood is different from race to race so it helps to give back a great result to players, using everything we have in a very production-wise way. When it’s up to characters, we have a tool named “Dressroom” which is basically a stand-alone scene where we import all the characters and creatures to check everything, test effects, weapons, spells and so on to see which adjustments we can work on in real time.
Well, the most realistic materials and textures are created directly from real photos (from our archive or directly taken from us here and there) with software like CrazyBump for example. The others are usually created and modified using Substance Painter, Photoshop or Illustrator.
Unity is a great engine as any other engine available now for developers. This is because results depend on the people using the engine itself. When we started the development, we thought about the chance to use Unreal or even CryEngine but we decided to go with Unity for very practical reasons. First of all, the entry level to use it is a bit lower than others, so it’s easier for us to bring people in without relocating from abroad (Italy is not the best place to develop games actually). Then we were sure that to achieve a good result in terms of visual, considering the isometric nature of the game, it was enough to do what we had in mind. We even considered that (with the right proportions and staying humble of course), Obsidian made something great with Pillars so at least it was possible to try this way. When we started prototyping the game it was fully 3D then we decided to move to pre-rendered stuff to stay safe in the long run, with optimization for a console especially. We are constantly improving post effects and would be great to add new features such as a true dynamic day/light cycle in the future. We have it now but is something happening on the game map only so when you enter a city or a fighting area, you can find day or nights, following the game clock which is connected to events happening in the world.
When you saw the game at GDC we have just switched from 3D to pre-rendered and it was the first public showcase of Alaloth. We invested a couple of months to achieve the results you are talking about (since then we kept going on in obtaining even better visuals of course and we are pretty satisfied with that at the moment). All the cool details and the majority of the lighting job (light-sources, shadows, general light conditions) are obtained working on the 2D image that represents every single environment as we said previously. This way the artist’s touch and light modulation (color, brightness, even the size of the light range for each light source in the scene) is the key. Of course, we also have a bit of 3D light (ambiance), but it mostly helps the look of the characters and details on their armors/bodies rather than the environments themselves.
As I wrote previously, the game is born with a full 3D soul. Then, considering the idea in the back that was staying in between classic and modern action RPGs, we thought that switching on a Pillars-like style with a real-time hardcore combat sys could have been a KSP for the game. But we did it even because of tech reason because of course optimizing a full 3D game is harder than working on something that runs only models/effects in real time. This is especially something we thought with consoles in mind. Moreover, doing like this the weight of the production is more on artists than on coders. And finding a good artist is always easier than finding a coder following our experience.
You can see the progress the team made by comparing these screenshots from the older versions of the title…
…with these image from the most recent build. Instead of full 3d setup, here the team used pre-rendered 2d environments. It looks even better in action.
To be honest, we didn’t mention 2018 officially. It was something press said following the unveil we had on IGN last year. The original plan has always been with a 2019 release in mind, considering a self-publishing option. The point is the game raised the interest of a few big names in publishing so we are trying to understand which option is the best for us. The problem usually is that when dealing with a publisher, you have to work on request features for pitching purposes. And we don’t intend to work on something with a never-ending session of pitching for months. Because this means working on nothing and if no one is going to sign a publishing deal, you have simply wasted your time. Of course, a publisher could be a massive opportunity to raise the bar in terms of quality and contents, so we are still on with a couple of companies and we had interesting bids on our desk. This means that the original plan could change. We could think about an Early Access option by the end of the year (that could be great IMHO but it’s something that I can not manage without someone in the back) or even for a 2020 release changing the scope of the whole project with someone paying for this. So, sticking to our plan, it stays 2019 but everything could change, E3 and GamesCom will be crucial to know more about this.