All work for my Final Major Project can be found HERE
I first sought out some reference imagery of the grass I was trying to create. I found a post on patternsinnature.org called Patterns In Nature: Nature’s Symphony which contained some great images, one of which I used as a primary source.
I then developed a simple rig in 3Ds Max for making grass. The scene composed of two planes (one for baking onto) and a countless number of grass blades.
From this rig I baked out the Normal map, Diffuse colour and Alpha. I created the roughness map in Photoshop by desaturating and tweaking the Diffuse.
I then produced a simple low poly mesh for testing. The texture samples and mesh were imported into UE4 for a quick test and the results were okay.
There were a couple of issues that needed to be addressed. Firstly, as this grass was going to be used in conjunction with UE4’s foliage painting tool, it need to be less polygons. 2-5 poly’s per tuft of grass would be ideal instead of the current 56. The grass also feels far to dense, a good prune was needed. I also needed to produce 5-6 low poly variants, and then in conjunction with the foliage tool’s random scale feature give the illusion of real world grass. It also looked to dark in colour.
I looked back at my reference imagery and noticed that it grows in little tufts. This is something I overlooked in my first attempt.
I thinned out the grass and lightened up the colours, then baked out the same information as before.
I then produced four low poly meshes to be used in conjunction with the foliage tool in unreal. I arrange the UV’s of the low poly mesh to match the baked information on the texture maps.
After importing into engine I built the material in UE4.
I then made a material instance which aloud me to tweak the colour and saturation of the base diffuse to get a closer match to its surroundings.
The colour in engine still feels a little light, and perhaps to green. Adding a brown, slightly dried out look may feel more accurate to its surroundings.
After a tweak in Photoshop, and converting the grass materials diffuse texture to a parameter (so I could change the diffuse colour base with ease) I started to get closer to the result I wanted to achieve.
I left some of the grass with the greener colour as the two together looked and felt better.
After breaking down the scene further I knew what elements needed to be done and which were more important.
I gradually built up the scene over time. With each new element added I took screen shots of my progress. I often tweaked and played with the light to see what moods i could bring to the scene. Below is the work in progress screens accompanied by a turntable video. You will see that I added a dirt floor which I felt give the scene a little more context as to where you could find this type of mobile work space.
I used a lot of repetition in the scene but utilising the ability to change the overall colour of some assets broke it up a little.
The second asset I worked on was what seemed to be some type of communication device. The device was used many times within the concept as you can see.
My original idea was to make this one unique asset then repeat it the required number of times to fit the concept. During the development of the asset I noticed that there was many smaller boxes with in the main body of the asset.
However I still carried on making the asset as a whole. During a discussion with my lecturer Paul, he suggest why not make each box individually so you can have more variation and arrange them in different configurations for each box. I adopted Paul’s idea and moved forward with it. Below is a visual representation of the idea as it was a little hard to describe.
The reason for take the idea forward and not carrying with the way I was producing before hand was due to one main reason. The overall piece would look a lot more dynamic and although still repetitive it would look more like a real world situation. It would also allow me to have varying colours of the housing box, which again, breaks up the repartition of the piece.
Above is the box which was represented by yellow in the diagram.
Above is the box which was represented by green in the diagram.
I decided to add lights to the assets and using different emissive maps create the illusion that the machines are doing different things.
Unfortunately I didn’t get the 3rd box done, which would have been the one represented by red. I did make the high poly model, but didn’t find the time to bake it down and texture it onto a low poly.
I then proceeded to make the outer case to hold all these boxes. It was a simple model that didn’t take to long to make. I followed the concept as closely as possible.
I then updated the scene.
I quickly made a high poly ramp and baked it down onto a low poly mesh. I used the same texturing process for the ramp as I did the platform base. The same base texture was used as well as colours used for the details.
As you can see the mesh is just 6 sides. It could’ve been 5 side if I had removed the bottom, but I decided to leave it on as there may be times where you could see it slightly.
To get the proportions of the diorama correct I first blocked out the area in 3DsMax. I worked as closely as possible to the concept art as I didn’t want to loose the feeling portrayed by the piece.
As you can see I didn’t add the cloth to the block out as it didn’t change the overall form of the piece. The chair wasn’t added either as I didn’t feel it was necessary at this point.
Once I was happy with the block out I started to replace each piece for a more refined model. I started with the platform base which you can see below.
I used a high to low poly work flow for the floor. The details such as the electronics hatch were modelled on the high polygon model then baked onto a low using a normal map. Below you can see the high and the low poly models in 3DsMax.
I was happy with the results I got and decided to use this work flow for the rest of the scene.
For the creation of the diffuse I first used nDo2 to create a AO map from the normal I baked out. I then used a base metal texture and slowly built up a more realistic looking asset. You can view a little animation of the process HERE
I placed the scene into Marmoset Toolbag 2 which is where the diorama is going to be displayed. In the image above the metal trim around the side is white. With feedback from my peers I decided to make it a lot darker as you can see in the previous images.
For the first part of my 3rd client brief I was tasked with looking further into the Rabbit Heart project and developing an aspect of my choice. As I would like to pursue a career in environment and asset creation I chose parts of the project relating to them. There were a couple of areas which I really liked the look of. The first area that grabbed my attention was a sensory system the player would encounter before being introduced to Ululu’s Mech Rabbit companion.
The image below, taken the Rabbitheart blog shows the area mentioned.
I took the ideas above and moved them into 3DsMax where I quickyl blocked out the area.
I was happy with the quick block out and I felt the scale looked correct to the concept art. While I was developing this scene I came across another piece of concept art that I really wanted to develop and as Paul had already developed the space I was working on I felt it was influencing my work a little. With this in mind I decided to start work on the other piece of concept art that I had discovered.
My eye was drawn to this image for a number of reasons. Firstly I felt it would make a good portfolio piece as the concept had some narrative to it which I would like to portray. Secondly I wasn’t overly complex which I felt left me some room for artistic interpretation. I had recently been looking into diorama’s before coming across this concept and I felt it would be a great way of displaying it.