Monthly Archives: March 2016

Post Mortem; The Dragon Garden

Overall this project has gone very smoothly; timewise, I have been much less under pressure than I was for the Fool’s Room, and I had much less Zbrush to deal with which I always find stressful. I have a lot more bad than good to say about this diorama, but I’m going to keep it short and just mention the main ones.


I’m pretty happy with the product so far, and have a very minimal list of things which need to be dealt with in polish time. So small, in fact, that I can actually type it myself;

  • Birch tree texture needs to be improved.
  • Blackbird needs unwrapping, texturing, and animating.
  • I need to add variation to the large flat stone face, particularly where the roots cover it.
  • Runes need to be added to the pillar faces.
  • I’d like to create a bumblebee somehow.
  • I need to add ‘tufts’ of scales around the dragon’s beak and face.
  • Music/sound.
  • Add dirt for the roots (current placeholder).

Even with some additions this list will remain manageable. I’ve posted my work to Polycount to get some opinions from more people. So far, a lot have people have said that they like my style which is great. Hopefully I’ll get some constructive crit soon though.

My thoughts on the diorama so far are as follows.

I’m generally pretty happy with the diorama overall, and with how smoothly it’s gone in general. There was much less new stuff to deal with for this scene. I’m probably least happy with how noisy and hard to read the scene is overall, and also there isn’t really much going on composition-wise. I don’t feel like anyone would know where to look because you’re just assaulted with a mish-mash of noisiness. A lot of this comes down to the lighting, which despite an obscene amount of fiddling, serves to just add another layer of noise to the noise. I prefer how the scene looks when unlit.


I’m quite pleased with the colours, however they’re definitely not what I was initially going for- I’d wanted the scene to be evening and almost golden/sparkling, with much more saturated/rich colours and darker shadowing, but I definitely needed to lose that if I wanted the scene to be a success. I like how I have complimented the green/orange of the scene with small areas of pink flowers which are reflected in the dragon as a later addition also. The white of the birch trees additionally works well to break up the darkness of the scene. I really like the combination of pink/blue/green/yellow with some white to break it up.


I spent a lot of time playing with the dragon texture, taking it through practically every colour combo you could imagine before settling at the current one. I also revisited the dragon’s face just last week to bring it up to a better standard- I’d had a comment or two that it looked a little blobby. I’m not totally satisfied with it. Were time permitting I’d totally start from scratch on the dragon, but that won’t be possible.


I’m  unhappy with my ability to sculpt rocks, and though I can hide behind the pretense that they were meant to be so simplistic for the sake of stylisation, I had actually wanted them to be much nice looking and slightly more detailed and realistic. My failure came down to both an inexperience in creating rocks, and partially the low resolution I worked with. I’m going to be taking much more care with the rocks for my next diorama, but obviously I can’t change my process too drastically for the sake of continuity.

Generally, I’m not proud of any of the assets I created for this diorama. They look decent as a whole when viewed from a distance, but if you were to view, say, the tree bark material on a sphere as a singular thing. Well. I wouldn’t let that happen. This is mostly my fault; I did actually find a tutorial on how to create really nice tree bark in Zbrush, but due to fear of running out of time on my project and also my aversion to Zbrush, I didn’t try it even though it would have drastically improved my work.

As I’ve said before, in the future I’d like to put some more serious effort into actually creating good, hero-level models/textures/shaders as opposed to an overall nicely composed and considered scene with largely rubbish/rushed assets in them. I think my strength is fully hand painted textures without normal maps, and I’d quite like to explore that some more and then get more into Zbrush. I think that’s what I’ll be taking from my FMP; it’s time to actually get good at hand-painting textures rather than hiding behind an overall scene where you can’t really scrutinise the actual quality of an asset.

I don’t think I’ll be able to bring this diorama up to a level that I’m super happy with, but I’m still relatively satisfied with how it’s turned out. I’m looking forward to bringing it to life with sound later on.

Tagged , , , , , , , , , ,

Slowing down

As the title suggests, things are really starting to slow down on this diorama. It’s beginning to look pretty much like a finished product and I’m working shorter days as a result. It seems like while there will inevitably be a few things I need to polish later, there will be a much smaller list than for my first diorama. Not that I’m not agonising over stupid, tiny details though. There are a lot of things I want to add that aren’t really worth the time and I’ll try to force myself to ignore. Currently, my polish list is;

  • Increase the resolution of the dragon’s face, add details.
  • Add moss to dragon texture.
  • Add scale/feather tufts to dragon’s face.
  • Improve the birch tree texture.
  • Moss on tree roots.
  • Rock face variation.
  • Dirt for roots (replace placeholder).

But, I have also considered adding;

  • Blue flowers.
  • White ivy flowers.
  • Shrubs.

However I think ultimately it’ll just make the scene even harder to read than it already is, and they won’t be worth the time.


This week has been a mixture of new asset creation,  but also a lot of polishing, playing with lighting, and general fiddling to respond to critique.

The newly-added pink salvia flowers, heavily inspired by The Witcher, help bring some much needed colour variation into the scene. I’ve also added a similar pink to the dragon’s spines and even a rose tint to the birch trees to compliment it. The large white ox-eye daisies also help to break up the rather green/orange palette of the foliage.

The pink/blue of the now-fixed sky contrasts the orange/green of the diorama nicely too, I think.


I’ve been quite good for keeping on schedule with this diorama, and I feel happy that I’ll get what I want to get done, done.

There’s a lot of little things I’ve done this week so I won’t list everything- you can see that in my time schedule spreadsheet above, but I’ll run through a few bits. On Monday I created the black stone pillar at the back of the scene. In the books I am basing my work on, these pillars act as portals between other stone pillars throughout the realm. I also began the creation of the large fallen tree trunk which sits next to the stone pillar. I realised that I would probably need to scale down the trunk size, otherwise it would be absolutely huge in the scale of the diorama and look a little weird.


Though I liked how it broke up the scene, the trunk’s height near-matched the stone pillar and looked wrong. I decided that to save time and frustration, I would take one of my existing trees and modify it to have a large root system. This worked well, and the two tall trees frame the pillar nicely.


I also made the decision to re-add the birch trees, which I had previously removed because they were throwing off people’s idea of the scale of the scene. I modified the leaf texture to make the leaves smaller so that they matched the scene better, and placed them mostly on the raised section behind the pillar to make it come forwards more. Without them, the darkness of the pillar blends in a little too much with the dragon. I also feel that the white of the birch trunks breaks up the darkness of the foliage/dragon and adds interest. The trees are also all now animated to sway in the wind.


The above image shows a subtle mosquito particle I made. I think it grounds it for the viewer a little- everyone at some point has had to walk through a gross cloud of flies. I find myself thinking back to warm, sunny days when I’ve had to duck through fly swarms, and it brings the feel of those days into the diorama scene for me.

The final addition to the foliage I made was the ivy.


The main reason I chose to add ivy was to create the sense that the dragon has been laying where it is for a long time. It’s really difficult to create the sense that the dragon is stone, while still maintaining all the colours of the dragon as if it were living. Some things don’t work as well visually as in writing, and this is one of them. The ivy definitely helps, but it doesn’t completely solve the problem. Also, the ivy is so noisy that it just makes everything confusing to look at.

The whole damn thing is confusing to look at. It’s frustrating.

I wanted to get the feeling of the scene being very overgrown and alive, but I think I’ve just ended up making the whole thing feel really messy. I’m not really very happy with a lot of aspects of it unfortunately. I think it looks good as a whole, which is what I wanted, but it just looks awful when you start looking at it up close because all the assets aren’t very well-done. For future projects, I’d like to take the time to create really nice-looking hero assets for my portfolio, as I’m lacking that in my portfolio.


Currently, I still actually prefer my diorama as an unlit scene. Since this is week 5 of this diorama, I doubt that’ll change. In the unlit scene the colours are just nicer, and without the dappled lighting adding more noise it’s much easier to look at. I’ve tried to reduce the contrast between the light and shadows by upping the sky light brightness and bringing down the brightness of the source light, but it’s impossible to completely ease the nosiness created by the lighting.

So, this diorama is nearing completion. Over the weekend I’m hoping to create a blackbird to sit in a tree somewhere, and perhaps a bumblebee later. And of course I have my polish list. I’ll be taking this work to Polycount and to university tutors for crit, so I’m sure that list will grow. I’ll write a full post-mortem in the next week or so. 🙂


Tagged , , , , , , , , ,

Nice Trees

I’ve been making the main trees in my diorama this week, something I’ve been looking forward to for a while.

I was heavily inspired by trees in The Witcher 3, and did some research to find that the trees I wanted to base mine around are called Scots Pines. They’re very cool, twisty and gnarly trees with extremely textured bark. They come in all sorts of shapes and sizes, and some really remind me of trees you would expect to see in an African bush kind of environment.


I really wanted to get the trees right because in the past I’ve always been disappointed with how my trees have looked, particularly with regards to how I texture the trunk and branches. Slapping a tilable on there looks awful, but trying to texture the whole thing as a unique is time consuming and ends up looking worse because I just rush. To reach a compromise, I decided to create a set of textures, both tilables and uniques, to quickly give the appearance of the tree being uniquely textured across the board. I’m so, so pleased with how it turned out. Like, I wanna wave these trees in people’s faces because I’m so proud of them.

I’m going to add some additional secondary and tertiary roots to the foreground tree, as the root systems for these trees are quite extensive. I’m also hoping to simplify the canopy texture so that it renders better at a distance. As you can see, the pine needles become very blobby from even a slight distance.



You may have noticed that I’ve pretty much completely rejigged the lighting of the scene. Initially, I wanted to make the lighting evening so that it was very warm and golden, however I find that evening lighting is very difficult to get right in UE4. It looked a little weird I realised, and was making the scene quite difficult to read. I’ve made the shadows more blue/purple, and changed the lighting over to a more midday-esque feel. This brings out the intended colours in the textures more, particularly the dragon which was losing the blueness I had wanted it to have. I’m much happier with it and will probably keep it this way.


How the colours look in Unlit mode.

As a result of changing the lighting, I’ve started playing about with the background colour again. It’s a tough one. The golden rule of pretty much anything visual is to not put things on a black background, however I actually quite like how this looks on black. For some reason, the sky colour is refusing to change according to the sun position like it should. It looks ok on an orangey background, but I’m going to fiddle with the skybox and test it on a blue sky later, as this will probably make the most sense to the viewer. I feel a blue sky will drown the blue of the dragon however.


I made a couple of changes to the water such as making the area covered by water larger, and also discovering a tickbox that makes the water actually render reflections properly, if subtly (screen space reflections).



I’ve also made the addition of more foliage! Lilies and lily pads, cattails, and ferns. The moss is starting to look a little weird now, so I want to edit that.

I have a few concerns with the dragon aspect of my diorama, and I’m not yet sure how to deal with them. I’ve had a few people look at the dragon and say thing like, ‘are you going to make the wings more real, like with sub surface scattering?’ or ‘why don’t you make the dragon breathe?’ and so on. I feel like an idiot when I explain for the fifth time that the dragon is meant to be made of stone but actually looks real, colours and all- it’s just stone to the touch. I’d totally have done all those things otherwise, I swear. But stone doesn’t breathe, doesn’t normally have SSS, and so on. I’m toying with the idea of just making it move like it’s alive, or assuming people will read the book excerpt I’ll include in the ‘game’.


Additionally, the dragon’s head is too low-resolution for my liking, and I’m not sure what to do. It looks a little out of place- literally like I’ve blown it up from a smaller size. I’m thinking maybe I could just take the head aspect of the dragon, increase the texture size and re-texture it, and then apply the head to a different material ID so the head only is much higher res. It means there is a head-sized section on the dragon texture which is being wasted and abandoned, but I don’t really see an alternative. There are a couple of further changes to make such as making the wings more detailed, and curving the horn on the dragon’s head to match the spines.

I’m starting to feel like the end is in sight, and I’m liking the scene a little more than I have in the past. I’m growing a bit concerned about the ever-growing polish list, but it’s easy for me to forget that I have roughly 4 weeks of polish for all of my dioramas. Time shouldn’t be a problem. I have a lot of gripes and frustrations about this scene… some things just don’t feel right… but I’m hoping I’ll overcome these problems by the end of my FMP. I’m feeling relatively optimistic. 🙂

Tagged , , , , , , , , ,

Didn’t want it anyway.

As mentioned last week, this entire week has been dedicated to the creation of the sleeping stone dragon in my scene. it was a good job I gave the whole week to this task too; it was practically the perfect amount of time that I needed. I told myself to make some shrubs this week too, but I haven’t got round to those yet and that’s not the end of the world. Also, the dragon’s wings, spines, and various other sections are not yet finished and I will work on them tomorrow, rather than taking my scheduled day off *sigh*.



Apart from the obvious addition of a dragon into the scene, I have made very few serious changes beyond procrastination and finnicking. I copy-pasted another tree in there. I added another ‘waterfall’ (trickle, more like). I downsized the grass meshes and upscaled the trees to better reflect the scale I want, though I still need to work on that.

The background is now a flat purple, generated by a sphere with an unlit emmisive material on it. I prefer the gradient generated by the skybox in last week’s screenshots, however to have that breaks the lighting and makes the shadows become too desaturated after building the lighting- the skybox colour influences the shadow/indirect light colours. I’ll probably texture the background at some point to have a gradient, and give the background in the Fool’s Room a similar treatment (I had similar lighting issues with that scene too, it was just less obvious). I’m glad I’ve worked out a resolution to a long-standing issue.

For the dragon I did a fair bit of research trying to find the best technique for creating, in particular, the scales. I considered techniques such as hand placing scales in 3dsMax (too time consuming), fibermesh in Zbrush (scales were too irregularly distributed, like feathers), Zbrush micromesh (confusing, beyond my skill level), Zbrush Insert Multi Mesh (didn’t want the stress of learning such a new technique in scary software at this point), Zbrush Alphas/Stencils (don’t like the look it gives). I ended up just going to hand-sculpting the scales, and I’m pretty happy with how it looked. I much prefer the control I get from hand painting the scales- I can change the size, position, and orientation of each scale however I want for the desired effect.



Sculpting was probably the smoothest part of creating this asset, if a little tedious. I set the bar quite low for what I wanted to achieve- I didn’t intend on creating a full anatomically correct and beautifully designed/sculpted dragon, otherwise I would have spent 5 weeks on that alone (and probably still done it badly). The dragon will ultimately be quite concealed by shrubbery, flowers, leaves, ivy etc. so I didn’t want to go too overboard. A week was definitely enough for this asset in this particular project.

I ran into a few annoying issues with baking and importing into UE4, where something in the mesh seemed to have corrupted and so importing into UE crashed the editor. Exporting as a .obj and reimporting to Max fixed that, but then the smoothing groups didn’t seem to want to work properly; I wanted the back spines on a separate group, but it wasn’t affecting the mesh in any way. This weird problem in turn made the bakes on the spines go wrong. Detaching and reattaching the spines sorted the problem, because 3dsmax is stupid, but the normal map bake on the spine still had some annoying artefacts I couldn’t ignore. I decided to just abandon the subtle normal map for the spines. I didn’t want it anyway.

I’ve been jumping in and out of watching Jonah play World of Warcraft recently, as opposed to caving to this horrifying addiction myself. He takes screenshots to show me sometimes, and today I walked in and really liked the area he was in and asked him to take a shot of that too (top).



I find the woodlands of WoW so inspiring and depressing. I particularly like the top screenshot which shows the lovely texturing of the tree with the lichen on it, and also the incredible ground textures. It’s inspired me to go back in and have a go at painting some more detail such as lichen onto the rocks, which will hopefully help improve the scaling of the scene. It’s also given me some ideas for how I can clutter up the floor of my diorama. Looking at the grass in my work, I’m happy with the texture I made but really disappointed with how low resolution it is after imposing it onto the large base island. It looks a bit (very) shit, but I’m not sure what to do about it. Vertex painting the grass texture in would make sense if the model had more verts to support that… but it would go way over the specifications I’m ballparking. I don’t know. It’s made me sad looking at the screenshots of my work and the WoW art. My work really isn’t good.

I wish my work was better.

I am sad.

See you next week when I’ve hidden my ugly dragon and ugly grass behind more ugly assets.


Tagged , , , , , , , , ,

Ticking Along

This week has been a vast improvement on the last. I’ve still been a little iffy, but I feel that I’m getting back into the swing of it. However, I’ve probably been a little easier on myself than I should have; despite the fact that I’m taking 3 days off to visit home this weekend, I decided to take Thursday off as well. This was because I finished my work scheduled for Thursday on Wednesday, but I should have just brought some work from next week forward. I feel guilty.

Time Schedule

The week started off well, and I just got lazy. Can’t afford to let that happen next week, even if I’m still doing ok for time! It’s just silly.

Diorama at it’s current stage.

Overall, I’m beginning to feel better about my diorama progress. I’ve spent some time trying to bring the atmosphere up to scratch, such as by playing about with the lighting, sky colour, and post processing. Obviously this’ll continue to evolve though.

Here is how I create the dappled light on the diorama; leaf planes with ‘SimpleGrassWind’ for movement.

Currently, my biggest concern with this piece is getting the scale correct. The dragon is going to be the big give-away to the size of the diorama, and I may have to come back in and change the scale of other assets such as the grass to reflect his size. I’ve decided that I will probably be trying to make the scale of the diorama a little smaller than in the concept, as the diorama begins to feel less like a diorama, the larger the scale. I’ll do this by making the dragon a larger component, and have him curl more closely around it. I’m sure adding smaller elements such as flowers will also help with this.

I’m really pleased with where the water is going. The purple colour of it, though unrealistic (well, this is stylised after all), brings some variation to the scene and nicely reflects the sky colour. I’ve been learning more about materials recently so I can make the water’s edge fade out when it intercepts with another mesh, and I can use a location-based opacity fade to make the water-fall areas fade as they fall further from the diorama. I’ve then added a particle effect over the top of the falling water, and a mesh with a panning texture for the foaminess of the water just before it falls off the diorama. I feel like I learned loads making this aspect of the diorama, and I feel much more confident with creating shaders and materials in UE4- there really is a way to create anything.

Water material setup.



You may have noticed that I’ve put in my birch trees from the style tests I did in the first week. This saves me a whole load of time, even if I do decide to come back in and modify them during polish time. They may just end up being a placeholder, but I’ll see. I’ve also added some more rock meshes, and moss to top them.
Moss was a funny thing that I’ve never really known how to create… I’ve tried before and always ended up abandoning it. However, someone on Polycount linked me to a Wikipedia with a series of tutorials. Lo and behold, one is for how the moss in Ryse was created ( It was genius! So I followed the tutorial and created some nice, fluffy(ish) moss. It works on the premise of using 4 mesh layers going from dark and opaque at the bottom, to lighter and more transparent at the top. It’s oh-so clever, and I think the higher polycount is totally worth it. I then used depth-fade to make the edges of the moss fade subtly into the rock and hide hard edges. 

Up close, you can see the 4 layers that create the softness of the moss.


The base platform was also finished this week with the final addition of the grass. I made a tiling texture for the grass, but then overlayed this onto the texture of the base itself and masked out areas in Photoshop for maximum control. The texture for the large base platform is 4k, so it holds up the resolution of the grass just fine for what I’m after and saves me having to bang my head against UE4s vertex painting tools and the likes. It was important for me to be quite regimented in how I arranged my Photoshop layers for that texture especially, so I can quickly alter, say, the tiling of the grass.
So I’m starting to like where this diorama is going, though there are still things I’d like to fix, improve, etc… such as the UVs on this grass which are causing silly floaty artefacts. Pure laziness on my part… a common theme.
Next week is the dragon! This will be a tough one, as I’ve never really made an asset of such scale, complexity, and importance. I’m not 100% sure how I’m going to go about it. I think a lot of people would say to use Zbrush, but I’m thinking a lot of it might be a 3dsMax job. I don’t know. Whatever happens, I’ve given myself a big chunk of time to do research and make mistakes, so it will work out. Time wise, things should be ok. I’d say I’m running at a similar level of completion as I was for the Fool’s Room, and the amount of repeating assets means that I will fill out the area much quicker than I did for the FR.
Once again, here’s to next week being even better. *sigh*

Tagged , , , , , , , ,