Most interesting and inspiring artist on the business.
Great interview! I appreciate you make this kind of material that inspire and helps people in the same path :) Thanks 80 level and Mats! Ancelmo Toledo ;)
Thanks for sharing.I found a lot of interesting information here. A really good post, very thankful and hopeful that you will write many more posts like this one. - earn to die
Specially for 80.lv environment artist Aiko Shinohara from Japan talked about her amazing environment, inspired by the Bodleian Libraries from the University of Oxford. She talked about the choice of tools, the production of assets, beautiful lighting and materials. Aiko is a very experienced designer, who worked on Street Fighter 4, Bayonetta, Vanquish and Final Fantasy Type-0.
I learnt CG in school for a year and went on to work in the game industry after graduation.
I started working in contract-based production for about 5 years. There were times I visited the studios, as well as joining the development teams on site. I was involved with the environment production in games such as as Street Fighter 4, Bayonetta, Vanquish, Final Fantasy Type-0 etc. Eventually I joined Cyberconnect2, Platinum Games and currently I am working as a freelancer.
Despite being an environment artist, I was always involved with management and pipeline creation/assistance which resulted in me not being able to make much art. My exposure to pipeline and workflow creation started me thinking how to create quality graphics, optimized and yet not compromising on the quality.
Since I was working in companies with rules and contracts, coupled with the fact that I did not have the chance to make much art, I did not have a portfolio that I could show publicly. I decided to pick up on the things I learnt at work to make my personal portfolio.
The aim was to make something that I could show and use as a portfolio.I also really wanted to use Substance Designer & Painter to work out a new workflow for creating assets.
I managed to get a friend, Yap Kun Rong, who works as a concept artist to help and advise me in my creation as I was not fully confident of the art side of things. He was a colleague whom I worked together with at Platinum games, who offered a lot of help during production at work.
We started discussing what kind of level I could make to achieve my aim of using Substance Designer, UE4 as well as creating a relevant portoflio for myself.
There were many things that we discussed and in the end I settled for creating a high quality proper finished environment. I think it is really easy to get lost in the process and end up with half finished work. With advice from him, I managed to understand the basis for creating a successful picture, mainly composition, shape language and lighting.
Could you talk about the mock of the scene? What software did you use to build the whole thing? How did you approach the creation of the elements for the scene?
The library is rendered in UE4, modeled in Maya in Zbrush, texture mainly done in Substance and Photoshop for some designs and tweaks.
This is the starting point of reference for the library scene.
This is a photo of Bodleian Libraries University of Oxford. I used this as a base for the look and feel and then went on to gather further refences for individual assets, parts and details.
I started making modular assets and tried to work out the layout in Maya. I did not split up the modular sets specifically during this process, only did a quick block out of the pillars, floor, bookshelves etc and used instance for laying them out. With this mesh in Maya, I exported it into UE4 straight and started adding a quick pass of light, checking the scale as well as spatial balance of the whole scene.
White box in Maya
Once I think the scale and general aspects are in place, I started importing the individual modular parts into UE4 and made the same layout in UE4. During this step, I started to make admendments to the modular parts so that they could be better used in the scene.
From there, I started working out the block outs for the props required for the scene.
Even though the library scene was decided, I did not think much about the time and era that the current scene will be portrayed as. After discussion with Yap I decided that this could be in the 1950s. With this in mind, I could start gathering resources and references for the smaller props and assets to to be placed on the table.
UE4 White Box
During the White Box process, there was not much details so I started adding details to the models at this stage. Simple breaking up of the big forms such as adding the indentations on the arches and ceiling, the way light reacts to these make the scene come more to life and also allows me to control the distribution of details and where I want the light and shadows to be. This gives a clear glimpse into the possible final outcome of the image.
Once I am happy with the Grey Box the next was starting work on the individual assets.
UE4 Gray Box
I felt I made a mistake here by not having the base color in much consideration at this point. Looking back at it I think I would have paid more attention because the Base Color of the models highly affect the lighting which ultimately impacts the overall look, feel and atmosphere of the image.
As a result, the impression of the scene changed when I started working on the textures. For my next project I will definitely put more emphasis on the Base color during blocking out or even use Substance Painter to make some quick simple textures closer to the final outcome so that I can work more smoothly on the assets without changing the impression/direction of the scene as I progress.
How did you create the 3d models for your scene? How many elements did you have to make? How did you approach the optimization of your 3d asset pack?
By adding more things to the scene, it may get better but I wanted to avoid making more than required and instead try to push for quality in a realistic and practical way. I fixed the main compositions and from there decided on the props to make.
Once the Grey box progressed to a certain level, it became easy to spot the required missing assets that would aesthetically enhance the scene.
These are 2 images before and after the props are decided.
I added things that would positively impact the overall look of the scene yet are not technical hassles. For example, the gold fonts, books on the shelves, green panels on the ceiling etc. I also changed the direction of the Directional Light slightly and this whole process was done with the help from Yap, offering his advice on composition and lighting, bringing out the points of focus in the scene.
Up: Before adding the props Bottom: After adding the props
Next, I worked on the camera for the video. I did the layout of the props with the main fixed camera angles in mind.Having this idea of a fixed camera show-off spots of the level, I could concentrate on making better decisions to improve the aesthetic quality of the scene without making time consuming assets that might not be have as much value-addedness to the scene.
As this was not made for a game, I did not particularly put in extra attention in optimisation.
I made a master material which I then used for most of the assets. I added a normal map with a cloud mask to give the reflections slight variatons when light hits the surfaces to try and imitate the realism in nature.
The Master Material Node I used
As it is a library there are tonnes of books. However, I only used 2 meshes for all the books. Design of the books could be changed by the material settings, which I used to make it look like there are more variations more than there actually are.
As for the layout of the books, the thickness and colour are randomised to make them look less alike and more natural and realistic.
Blueprint made to place the books naturally
I did not have a lot of experience with the blueprints so it took me a while to figure out, but it proved a time saver when I managed to get it working.
Node of the Blueprint
You will notice that if you look at the scene, it does not contain any complicated assets. As such I made alot of the high poly models in Maya. I made the meshes such that I could add details using Substance without much fuss. For things that were hard in Maya, I used Zbrush instead.
Base made in Maya with added details in Zbrush
I baked the textures of the base data in Substance Painter, and made the materials for the respective assets. I looked at alot of references when making the textures. I paid extra attention to the type of dirt and scratches and where they usually are found.
The book in Substance Painter
I used 2 wood textures for this scene. One is for the parts that make up the library itself such as the pillars, arches, and the other is for the furnitures such as the chairs and tables.
The texture for the architectural parts of the library was the first to be made so it was a process of trial and error for myself on how to create the textures using Substance Designer. I used Bitmap2material, basic photo of wood to get the base color, normal, roughness and used these to eventually create the actual materials used in Substance Designer.
Bitmap2material to get the base colour, normal, roughness and used these to eventually create the actual materials used in Substance Designer.
I used PlankWoodBase_B in Substance Share for the material of the props. I added some customization to this and subsequently used it for a lot of the assets in the scene. There are many parameters in the material and they were extremely useful for controlling the different materials for each asset.
A lot of the textures I used were from Substance and SubstanceShare. I changed and tweaked them to what I needed for my assets.
I used Lightroom：Interior Day Light found in UE4 Marketplace for reference with regards to the lighting. The scene used white planes to reflect spotlights shone of them from outside the window to light up the interior scenes. I did the same for my library, using planes to bounce off spotlights from the exterior into the scene rather than using direct light.
This is with reflected light only.
I used directional light to to add the light coming straight from the outside. With only Directional Light, I could not get as good a result as when I used it with the reflected lights.
With Directional Light added.
Furthermore, I added a spotlight to get some specular for the pillars and arches. With Castshadow turned on, I got some weird unnatural shadows other than directional light.
With lights for specular.
The thing I noticed when working on the lighting was reflection.
I put in a lot of Reflection Capture in the scene. The whole scene is dark with huge reflective surfaces such as the tables and floor so I added a Reflection Capture for each desk. I managed to get nice reflections by doing so and avoided using SSR as it did not give me the results I was expecting.
As this is a scene with relatively simple forms, flat reflections tend to make it look more CG than real. To solve this, I added layers of cloud normal map to the master material to break the flatness and uniformity of the surfaces.
With no added cloud normal map (only table)
With cloud normal map added. (only table)
To get faster in production it is best to spend time and effort on the white and grey boxes. It is understandable that one is tempted to start making assets quickly at the start of production, but spending time on the grey/white box will eventually save a lot of time on revisiting and dumping work later on.
As I don’t draw or make sketches, I think gathering references and using them to build the grey boxes has the same value as having the concept art. This time I had a lot of real life references to work from. I think for original scenes in future, I would do some simple sketches and follow the same process of working things up from the grey/white box.
This may sound embarassing but it is my first personal environment work that I have done. The first time is always the hardest and it took me a lot of time. I think with the experience I have now in doing this I can work more efficiently for my future projects. I think there is no need to hurry and rush into trying to achieve something, it is probably more important to achieve quality before trying to get faster with the process. With the ability to achieve a certain level of work I think speed will pick up with time and experience.