Looks like I have to refresh my knowledge of rigging! MakeHuman can create rigs for characters but I’m finding them hard to use. Tiny bones, no IK, etc. I had a look at the rigify addon for Blender but it seems unnecessarily complex – lots of head bones that don’t seem to fit my basemesh very well. So, back to doing it myself.

There seems to be a neverending stream of skills I need to get on top of before I can create anything decent. I seem to remember complaining about that about 10 years ago, so I guess not much has changed, including my skill level. I’m trying to hang in there.

Tweaking, Tweaking, …


More tweaking, plus I used MakeHuman to create the nearest character. I plan to spend a bit of time with that program to learn how to create decent NPCs, so to speak.

I’m not sure how far to take this. I need a layer of grunge over everything, but I could add heaps more detail. I would like to get this up on BA though, and move on to something else.


The tutorial I’ve been watching on Gnomon has an insane amount of detail in the scene. It’s Creating a Sci-Fi Alleyway by Devon Fay. He has an electronic parts stall in the foreground, with shelves, boxes and baskets full of small items. He admits that some of the components he made for previous projects, and I guess if he’s a professional modeller he’s been making stuff for years. Be nice to have a decent library of components I can call on. One of the reasons I bought a couple of kits to kick-start the process. I must hunt through them to see if I can find some useful air conditioners and the like to use elsewhere.


So here’s my scene. I was getting a bit discouraged by the prospect of having to model a lot of stuff, but decided to keep this one a bit simpler. That’s a hardware shop right midground, so I’ve put a couple of baskets outside at least. Fortunately I remembered that there’s a rattan material comes with Substance Designer so I was able to create the necessary textures for those. I might be relying heavily on SD I think. I’m not making a huge amount of progress with procedural materials in Cycles, and SD has definitely got more options. Can’t use procedurally unfortunately, have to output as image textures.

I’m afraid boredom is going to be my enemy in this work, although I enjoyed reworking the house left foreground. Maybe it won’t be so bad. At least with the boredom I’m a bit less obsessive about the whole thing.

Let’s Do This


After wondering if I’m really prepared to take the time necessary to do a good job on this (or any other) image I decided that I am. I estimate a month at least, just working on details. I spent quite a bit of time this morning getting the detail around the door on the house on the left, fixing the bottom window and working on some better (but not final) materials. I’m hoping that every asset I spend time on I can reuse somewhere else.

I’ve been collecting more reference too, and PureRef is certainly a great app for organizing that. My modelling skills are actually not too bad. I’m finding booleans very helpful for certain jobs such as the doorway on house left. Blender seems to be able to handle them better nowadays. Also I’m using image textures but not UV mapped.

Another Day


I’m breaking out more of the scene into separate files. Should make it easier to work on them, and easier to use as assets in other scenes. I plan to spend a bit of time on each one, fixing a few things I don’t like and adding more detail. Spend a bit of time on texturing too.

That green shop on the right presents lots of opportunity for goods to spill out onto the street. I’m a bit inspired by a course on Gnomon at the moment, an alleyway with lots of boxes of stuff piled around. Other scene dressing too of course. I think it will take quite a while to finish.

More Work on The Street Scene


I’ve started to break out the buildings into separate files and link them back into the scene. I’ve also made them a bit wider – they were about 5m wide which is pretty small by any standard. Applied some basic materials, found a nice HDRI that I might stick with, added some sidewalks and generally finalised the composition. I don’t think the figures will make it into the final scene but I think I need some human content.


Some windows, a couple more people, minor changes to the composition. I’m quite pleased with how this is coming along. Now I need to work on the details, and I tend to find that a little boring. Let’s hope I can do a decent job of it.

That Street Again

I did a workshop in 2013 on environmental texturing, and we worked on a street scene provided by the tutor. Textures were painted (including photos) in Photoshop. Here’s the result:


I’ve revisited that scene a couple of times since, and as it fits the general scenario that I plan to focus on for a while I’ve decided to have another go and see if I’ve actually learned anything. Here’s the starting point:


My general goals are to make the assets modular and perhaps a bit more detailed than the originals. Also to add a couple of new assets. I’ll try to do the textures without UV unwrapping. And finally to add some interesting set dressing. As a general management strategy I think I’ll break each asset out into an individual file and assemble in the scene by linking.

One thing I haven’t been doing with my scenes is to find good lighting reference, something that most environment artists highly recommend. I guess I should put a bit of time into that first. It might be just a matter of finding the right HDRI, as different ones give very different lighting.


I’ve made some progress. Different composition, different lighting, and a couple of characters for interest and scale. Making progress.

Rewards of Virtue

Someone asked a question on BA yesterday about how to texture (with images) a large number of objects without having to UV unwrap them all. Someone suggested using generated texture coordinates, and I replied that object would be better to get a consistent scale. Which of course is the solution to my own problem – texturing an object so that the texture is independent of the size of the object. I just hadn’t thought that one doesn’t need to UV map an image texture.

Image textures have so much detail, and correct real-world detail, that is so hard to achieve with procedural textures. I don’t really have a problem with UV unwrapping things, but couldn’t see how that would work with parametric models that can change size and position of details (like door position in a wall).

So I might be putting the parametric models and the procedural materials aside for a while, despite the amount of time I’ve spent on them on the past few weeks. I just need to get something out.

Rethinking, Rethinking,…

I’ve decided to put the parametric aspect of my models aside for the moment. Just focus on what objects I need in the kind of scene I’m intending to create. Doing it parametrically just creates extra problems at the start.

So, street scene. Obviously some kind of road. Buildings on either side. Maybe sidewalks. Then set dressing of various kinds – light poles, pipes, wires, rubbish bins, etc. I’ll be working to a grid so I can create modular assets. I’m inspired here by the texturing workshop I did with Leigh van der Byl a while back, and also the prevalence of back alley scenes on various CG forums.

Developing my own textures will take a while, so I might do significant overpainting at the outset. I haven’t used my Cintiq much lately, so perhaps time to connect it up again. If EOS goes up a bit more I might sell it and buy that Z8 too. No hurry though.