Texturing: nDo2 vs Crazy Bump

Today I downloaded the texturing script for Photoshop called nDo2. It works in conjunction with Photoshop and can help generate normal, diffuse, Ambient Occlusion, and all that. I did a little experiment with an unprocessed photo in my personal library. I'd normally tweak images before attempting the creation of normal maps – This involves manipulating the grey-scale values on several adjustment layers and painting/adding and using the existing image to create a height map. Even explaining it takes long.

In nDo2 I picked from a list of typical texture types – "Sharp Brick" in this example. After that it took about a minute generating loads of layers in Photoshop to create a normal map – The default was pretty good, and guessed well for the specular and gloss even. Great! After it has the normal it can be adjusted and refined (working within Photoshop linking to the active file), a bit like you can in Crazy Bump (with a lot of copy & pasting). There was a little difference in quality between the two.

nDo on the left and Crazy Bump on the right.  This is the preview after opening a plain old photo of a wall.

Also I tried converting the normal map into a diffuse – This actually worked out a lot of extra detail from all of its equations creating the other maps. Feeding this back over the diffuse and overlaying gave a much more interesting and rich texture. This would be handy for adding detail and variety to maps easily as a basis to work off. The whole process is so much easier, although I do have a moment of sitting around idle (was a large image to process). nDo2 is also prone to crashing – any cancellation of the process messes it right up – So experimenting can be time consuming. To compare the two isn't really fair – Crazy bump is a dinosaur in comparison, although is pretty quick and easy to pick up. Using the two and converting different maps between each other might be interesting, I did give it a go but I think Crazy Bump really needs a height/normal map to be of any real use. nDo2 cuts the height map process out somewhat, and there are loads of tricks I haven't even explored yet.

Here on the left I combined the diffuse generated from the normal map settings I had from nDo, on the right is the unprocessed photo .  Noticed lots of the subtle detail is more apparent?  Getting these selections manually would be improbable (I would probably paint it) and this is a nice quick way to acquire and customise detail, with the added bonus it will match great with the normal bump map.

When comparing the two you really have to consider the availability and pricing. They both have a free 30 trial, crazy bump for students is £30, nDo2 is £60(£48 on offer for Christmas). I have no problem paying £48 for nDo2 as it is worth it just for the few features I have used today. Creating normal maps can be a drag – You can project geometry within 3DS max, create & convert height maps in Photoshop etc, and project/bake from within 3DS Max, possibly importing from programs like Zbrush. 

nDo 2 can be useful for all of these examples – Enhancing these maps and even creating pretty decent diffuses to start painting on. When you cut out the monotonous and time consuming processes everything flows a lot better. There are no short cuts for many processes, and you have to really know stuff before jumping ahead and not really understanding the fundamentals. For example I used the preset "sharp bricks" in nDo, but I'm aware of how to create my own presets and my learning so far will help a lot.  On top of that I will have to be creative with the layers that are created still, I just now have more options thanks to nDo2 and more time to do other things.

Composition: The Eye of God

Hi, I was thinking a lot about composition recently. Mitch Small posted applications of a compositional tool Facebook recently.  It was the "root 5" grid... I'm not going to get technical, it's basically the golden spiral/rule of thirds rotated 4 times around the image, so there are four event horizons/infinity points are whatever you wanna call 'em... I personally like using the word "infinity point" - It kinda makes more sense to me.

Here's a picture of Mitch's overlaying the root 5 grid.  The colour wheel has a Gamut triangle -This defines the available palette (before he whacked on some adjusments).  Here I believe he used it to colour areas to match the planes how the light & shadow would fall (green from above, blue shadows from the front/dropped by the green source).

 part of what I'm writing about - not the numbers or theory - Not the how, but maybe a little about why.  I looked around google and found this site, I personally don't work with mathmatical formulas, I see patterns - This is what it's about, the calculations were written after the fact to try and understand how our mind's eye sees.

If you have the chance watch the film Pi (CLICK ME) if you haven't already.  Even from the advert you can get how it links to composition and is actually really interesting.  I think in a way the premise is on to something - Maybe there is a universal pattern or something?  In the film the some religious people call the pattern/formula "the true name of God", which I thought a nice spin.

Anyways, I made my own "root 5" template in Photoshop.  I immediately noticed that it will line up to many different images.

Here I compared it to a model I'm making - I made up a load of stuff in my head, which may be real - explaining the use of negative space, weight, and framing to create the form...  My problem is I did this after creating the model, so either it is a coincidence or there really is something about this pattern theory.  I thought it was uncanny how many points matched up.  The infinity points are in red (which are also called the "eyes of God"... pretty cool).  I must say though, if you look for something you find it - I could probably throw this over an Atlas and locate Atlantis!... if you get what I'm saying

I'd like to think there is some kind of pattern which unites everything... I'm not religious, but I am maybe a little spiritual.  I was thinking that composition is normally quite personal.. It can be a reflection of what you want to "say", a tool like perspective or lighting...  It just kind of encompasses everything to a degree.  In a movie the directer is the composer, like the composer is the composer in an orchestra - They just work in a different medium.

Just maybe I have absorbed more then I realise?  I've drawn and seen a lot of pictures over my lifetime, and experienced a lot a stuff.

Evaluation

Today was our (year two)term one evaluation (queue imperial march music or maybe The Ecstasy of Gold)

I wasn't too surprised by the results, critical studies are on track and 3D had silly little , 2D went smooth as I can output a lot of work and is my comfort zone. I'll note the key points mentioned.


3D    (Good)

• Obvious use of reflected symmetry
• Use of incorrect map sizes (non-binary multiple)
• The building was higgledy-piggledy
• Was a gap in the level created in UDK
• No Collision meshes exported into UDK
• Didn't use normal and specular maps on every asset

In my mind I have rationalizations for most of the points… But rationalizations don't matter, the design document should more than cover these points. I think the root of the problem is trying too hard if that makes sense? I'm trying to throw in as much as I can learn, but miss the fundamentals. The grading for game production is very clinical and the actual artistic content is almost moot. The next project I am going to keep it relatively mundane but ensure that I get all the right boxes ticked and make a solid & clean asset.

Here I have started the optional 3D project, I'm concentrating on the fundamentals & applying recent teachings.  I'm keeping the concept simple - a basic cart and some nature, all of my efforts are to make a nice looking clean scene & maybe rendering in UDK for practice.

2D    (Very Good)

• Presentation needs improving
• Not always abiding to standard paint techniques – layering/building up colours
• Should post more personal work in 2D folder

Visual design went pretty smooth, no major concerns and quite a few positive comments, which is always re-assuring. I haven't been putting as much time into 2D as 3D but have more experience and confidence here. I think my fine-art and graphics experience, and age help me a lot here. I really should be aiming higher though, I'm finding out about new techniques and tricks and could be doing better at implementing them into my practice.

Here's a pic I did that Chris mentioned he liked, I focused on colour theory/harmony and translating how I personally saw the scene - The tree and foliage dominating my path... a natural gem among the concrete of Leicester, with it's surrounds being almost incidental.

Crits    (Very Good)

• More reflection & applying to self
• Comment on 2D & 3D work

Critical studies I was a little unsure of, I do pay attention at lectures and have a basic grasp on writing, but was uncertain of whether I was approaching the subject correctly. To be honest it's difficult trying to work out what not to write.

I'd be honest and say I wasn't sure how on track I was with the syllabus until this evaluation. I do put in a lot of work and time, but that doesn't always equate into good marks. I was only really surprised by the crits score being so high (cheers Mike). It occurred to me that I don't really promote my own abilities or experiences much, I've been drawing for nearly 30 years (on and off) and have to my own surprise retained a lot of knowledge along the way… But how the hell do you know that? I can write it, but that's like showing someone a cake – they can't taste my unique recipe, and barely get a sniff. So I think Critical Studies has helped loads… Not just for the information supplied, but by the encouragement to self-analyze and make larger artistic and personal decisions.  This has helped me be more confident in my own abilities and limitations.

Game Engines & Optimizing

I have been thinking quite a bit about analysing and optimisation of 3D assets for game engines. To state the obvious, the less information processed the better the performance in game. There is the standard in PC games of 60 FPS (frames per second), and 30FPS is more common on consoles. Anything over 60FPS is really hard to tell the difference, although the BBC have been experimenting with 300FPS for sport broadcasts. The performance of any game relies heavily on the game engine; I had the experience of a poor port from console to PC earlier this year. I paid for the game Dark Souls and actually had to download an un-official patch which fixes the frame rate and you can manage the settings. The From-Software actually said 60FPS was impossible due to engine restrictions, but this can be reached on a high-end PC with this fix.

Anyways, more importantly and relevantly is optimization for the UDK engine. There is the fundamental way of making a strong library of assets which share much information, for example using the same mesh and applying a different material (re-skin), and shared textures in areas (a little like Fallout but not abused so much). More importantly for a game artist I have been looking at analysing and fixing models in 3DS max, as well as efficient texture mapping and unwrapping.

It may be interesting to know that the triangle count and texture size is a small amount of the information used from assets. Each triangle can be broken down into verts, so the number of verts counts for a lot, and can store more than position data. Here I have been using a 3DS Max script, which calculates these variables and displays it as a figure. I was surprised to find out the UV unwrap seams counts for so much – it's almost a whole other assets worth of information. Knowing this I could imagine how cheap tiled textures could be... I've gotta thank Mike Pickton again for sharing this information.

To help understand UDK a little better I recently purchased Unreal Tournament 3 on Steam. Not only is UDK packaged with the game to encourage community created levels, but it also has levels made from the same packages available to play with inside UDK. I have noticed in recent years lots of games have the Unreal Engine; this is because it is flexible, pretty all-encompassing and has a cheap licence… This all adds up to industry standard in game engines. More recent and becoming more popular is the Cry Engine, this is apparently a lot simpler to use (with less customization), but very effective and reproduce outdoor scenes. Also from my own experience playing Crysis the physics and destructibility is a big feature.

Staying Fresh

There is a common problem with anyone when learning and producing work. On one hand we can get "artists block", when we struggle to be creative. On the other hand we could be spewing out lots of work which really doesn't help to demonstrate progression… I'll call this over compensation. There's a fine balance between being creative and burning yourself out. Looking at the link you can see the first point is fear – people get too precious about their work and feel that it's not good enough, or that they are not good enough and struggle to reach their ideal vision.

This of course is very different for each person, and is more a reflection on their temperament then their work. One thing I do love about this course is that we are left to our own devices somewhat – If I want to do abstract paintings for digital backgrounds I can, If I want to learn how to use a 3D application and learn a bit, I can. Everyone has different experiences and works better differently, and the more you can draw from your own experience the stronger and more informed your work will be.

Personally I don't really love one thing, and like to try a bit of everything to keep it fresh. For the tripod project I concentrated on coming up with unique designs instead of a thousand similar designs. I then pushed one a unique design a little further and developed it. On one hand I have lost out on peer-points as I steered away from what everyone expects to see – Pointy limbs, illogical mechanical detail, organic and menacing forms… I did something like that on day one to get it out of the way.

It's worth noting that my day one picture got more Facebook "likes" than my main work. This doesn't surprise me, in a way I'm being selfish with my work. But that's what it takes to stay fresh for myself. I used the opportunity to try low-poly modelling; I learned lots in a small time about texturing and unwrapping, and have a lot more confidence in making quick/minimalist models quickly. See, this skill can help further enhance my 2D by incorporating this models and painting over them, and also reminds me of the good ol' days.

So to cap, I don't think is healthy to do the same exercise over and over when you gain little from it. For example I was drawing a load of 30 second figure sketches as an exercise. At the start it was great to help me work without thinking and gain speed, now when I do them it can be a challenge still but I don't feel that panic or split second decision making I originally experienced. For me this means it's time to move on, it would be beneficial to do them when I'm out of practice, and I now know somewhere to get excellent reference. Instead I may spend that time reading a guide or following tutorials or new programs like UDK and Zbrush.