Finished Model

While I was shooting for my post production project Phil finished the model by filling out the hat and adding a bulge to the back of the head where the hair texture will be. We also added more bulk to the waist to fatten him up a bit.

I started the rigging process today by mapping out the reference curves for the joints. I decided to give the antennae 6 sections for flexibility. You'll also notice I only mapped out half the curves. This is to save time, as obviously the joints on the other side will just be mirrored.

bag modelling

Phil and myself spent a somewhat stressful day trying to get past this rucksack issue. I resorted to trying to model it out of seperate objects thinking it would be easier, but it turns out making the straps become even more of an issue using this method. So we basically took a deep breath and did it the old fashioned way of working into the mesh. We learnt a valuable gem of wisdom here, and that is that modelling something complex seems to be about creating triangles and then getting rid of them. Either way modelling does seem to be something of a jigsaw puzzle if you're aiming for a purely quadradic mesh. Still! we eventually got it done. Special thanx to Phil for this one for working solidly with me even though modeling technically isn't his responsibility.

Of course, it was only afterwards when I realised that a lot of time could have been saved here with more careful planning before starting the model. If I had thought about what primitive shapes maya had to offer for example, I could have started with a pipe transformed to make a strap and worked from there. Never mind though, live and learn!

We modelling only one strap, and then used this opportunity to half the mesh, and then mirror it to ensure an even model and also to save going through the complexities of the strap again.

Next steps are to fill out the hat a bit more, give it some depth. I also remembered today he needs a magnifying glass still. All in all though a good days work, and while it was a shame to miss out on the Musion studio visit to get it done I really think it was worth it for the solid distraction-free work time we gained.

texturing tutorials

http://www.3dmd.net/MAYA_texturing_tutorials.htm

just a quick link Phil found for texturing/uv mapping tutorials

further details

http://plissken.fatalunity.com/tutorials/subdiv/index.html

Using this excellent tutorial on modelling hands I got them done today. I was fairly worried about the hands as my previous attempts were all, how should I put it, crap. You can see what I mean from the set driven keys post early on. This tutorial however proved that I was in fact seriously over complicating things as I often tend to do, and actually it was a very easy process. I skipped a lot of the stuff to do with detailing such as knuckles and fingernails as my model needs to be simpler than that to suit the style. However it was extremely useful for the sake of getting the opposable thumbs right.

I also whacked in some rims to divide the clothing. Im not sure about the hat though, but I've saved it in stages in case Phil or anyone else has any better ideas. As you can see I did all this on the nice clean model before I attempted the bag, as I still need to work out a way of doing the bag that's not going to result in ridiculously unclean, complex, or triangular geometery.

For the eyes I used simple spheres squashed up and positioned, and the eyebrows are basically shaped cubes. These parts of the geometery are seperate to the head, though they will still be either combined or grouped with the main mesh. This should make blendshapes easier, and also saves complex extrusions that may mess up an otherwise smooth face. Similarly, the antenna are primitive shapes that are seperate from the mesh for the same reasons. (Extruding from the top of the head will do nothing but ruin the shape, create unnecissary triangles and generally involve complex work on the mesh.)

We are fairly behind now, so Phil and myself plan to take two days of straight work on this to get the last bits finished up. I.e hair, bag, hat, textures.

Bag troubles

The bag has been a complete nightmare. Not the bag itself, but the straps. The way they go around and under the shoulder is very difficult to model without leaving it impossible to place edge rings. I may need to take a different approach to simply extruding faces.

Facial Modelling

(The attempt that worked!!)

(Creases in the face resulting from a model that has rough topology.)

This has been a lot trickier than I anticipated. As I have never modelled a face properly before it's been a bit of a process. My first attempt was modelled using the Stop Staring tutorial, whereupon the face is constructed entirely from the mouth outward with polygons. The issue this presented is severe difficulties with keeping the surface smooth, as you can see from my test render. Softening edges, converting to subdivision and smoothing did not sort it out.

So I made a second model, this time constructing the mouth from nurb circles to ensure a smooth finish, using the curves to adjust the surface. Then I converted the surface to polygons and attatched it to a polygon spehere with appropriate faces removed. This made for a far smoother model, as you can see in the other screen grabs.

Texturing/Blendshapes: before or after skinning?

Having spoken to Alex Hulse, I have a clearer idea of what order I should really be doing things in. Both texturing and blendshapes should be done before skinning. With blendshaping the head, it's important that the head doesn't deform due to weight influences. I did a quick test and sure enough that's the way to do it. Here it is.

Thanks Alex!

Set Driven Keys

Heres a link to a good tutorial website.
http://www.learning-maya.com/24-0-character-setup-tutorials.html

It's there where I found this tutorial into set driven keys for hands.
http://www.final3.com/html/tutorials/drivenKeys.htm

This tutorial was definately useful, as it explains different attribute types that led to my revelation in the last blog. It also pointed out for me that I can in fact key more than one driven object at once, meaning that I can have just one attribute on a hand controller controlling the whole finger movement. This may not sound any different to what I had done before, but there is a big difference. Previously I had set driven keys in a chain reaction effect. I.e I keyed one joint to be effected by an attribute, which in turn drove another joint, which in turn drove the last. This meant that if I were to individually move any of the joints in the chain it would inevitably drive the others. The way the tutorial suggests however, means that none of the joints actually drive anything; only the attribute does. This means the animator would be actually be able to move individual finger joints without causing the rest of them to curl up.

The only downside to this however, is that I can't create any new attributes attatched to the same rotation used by the driven joints as Maya won't let me. Meaning I can no longer have attributes in the channel box for every bone. You can still adjust them manually, but it's not as animator freindly.

Either way, this is still progression. However, I will look into it further to see if there's a way around this. And I know there is. The tutorial talks about using a 'seperator' which it promises can help solve this issue. Theres a lot of MEL bogging down this tutorial though, so I will have to look at this tomorrow when I'm not knackered.

High-rez / Low-rez switch: im such an idiot

I worked it out. Im so stupid it was so obvious and under my nose the whole time. The switch between hi-rez and low-rez on Moom had nothing to do with any of this complicated control-rig / bound rig stuff from Hyper-Realistic Creature Creation. No, I've been barking up the wrong tree. In fact it wasn't a tree at all it was a lamp post. Thats how far off I was!!

The solutions very, very simple. All I ever had to do was create a new attribute and set the data type to booleans. This creates an attribute that has only two values; on and off. All I then had to do was use the connections editor to link the attribute to the divisions level found in the polysmoothface input for the all models geometery. Again, I'm very glad I worked this out now before I carried on with the mammoth task of trying to figure out all that MEL scripting.

You may need to click on the images to enlarge to see what I'm talking about. I used Lil' G from our last project as an example of how this will work for this one. A simple attribute attatched to a controller is a nice and easy way to make it more accessible for the animator.

Thanx to Moom for the inspiration, but I actually worked this out while breezing through a tutorial on using set driven keys for a hand setup. I will post the link for this in the relevant post.

Control rig / Bound rig

After having a read through the advanced rigging section in Hyper-Realistic Creature Creation, it seems as though creating a low-rez control rig set to drive the high-rez bound rig is going to involve a serious amount of work relating to complex MEL scripts. I'm not sure if I have time to figure all this out, so while I'm holding this in mind for something I'd ultimately like to be able to do I'm lowering the priority of doing it for this project. I want to do something really good, and to that end simpler is usually better. Plus, the reason this technique is often used in industry standard rigs is simply that industry standard models often have a very high poly-count. (Particularly with hyper-realistic models, which is what this book is about.) Therefore this technique is perhaps innapropriate for the nice simple cartoon models that we have in mind for our project.

However, in trying to find a simpler way to do the low-rez to high-rez switch, I came across a golden nugget of wisdom in Mastering Maya 7. Refer to the problem mentioned in this link to a rig from last terms project:

http://gringoarts.blogspot.com/2008/11/bouncy.html

In particular the bit that says,
"Basically what I still need to do is sort out the orientation of the wrist joints, as at the moment they do not correspond to the rotation of the hand controllers. You can still rotate the hands, but its not very user freindly. "
Well I never did get that problem solved, as sorting out the local rotation axis of an individual joint is very fiddly, and even then it's very difficult to get it to be exactly where it should. However, the technique explained in Mastering Maya 7 will prevent this problem from occuring. The technique is to use EP curves to blueprint where the skeleton will go before actually building it, with the edit points being where the joints should go. This means when building the skeleton you can simply snap the joints straight to where they should be, saving the process of re-adjusting the positioning afterwards. (This is relevant, as it is at this stage where normally the joint orientations are knocked out of line.)

Handy.

Refer to the image for a visual example of what I mean. The two joints on the left have just been placed. As you can see their orientation is in the correct position (x pointing towards the child joint). The joint on the right however has been moved. When you move a joint the orientation remains where it was when it was placed. This is why it is out of line.

(You might have noticed that on the post for the old blog I was going to attempt the low-rez/high-rez switch then as well. In case you were wondering, that particular tutorial had no information on it at all, which is why it never happened.)

Moom

I thought I'd include moom here as part of my research, mainly because I respect this model and rig immensly. I'd also say it's a more relevant example of how I should be modelling than the examples in Stop Staring as it's a far more simple cartoon style. The fundamentals are still all here though; A widely diverse facial control panel, and a face modelled with circular topology surrounding the areas that deform the most. While this is the rough model, Moom has a control that enables the user to switch between a low rez and high rez model, which as I have previously mentioned, I want to do for my own. The head is disconnected from the body, which is important for doing blendshapes, otherwise I will end up with loads of entire models hidden in the scene.

Set Driven Keys

Here's a quick test of how set driven keys could potentially be useful in rigging. Watch this playblast and understand that the whole movement came from one attribute in the channel box.
(Ignore the crappy hand model, it was only a quick botch-job for the sake of a demo.)

While this has a lot of potential, unfortunately it means the animator can no longer have individual control over any of the joints in that finger without driving the movement of the next in the chain. This hopefully illustrates the importance of finding a way around this.

Reading Material

I figured I may as well post a link to the books I am using in case this blog inspires anyone to find them for themselves. I would reccomend them.

Stop Staring:
http://www.amazon.co.uk/Stop-Staring-Facial-Modeling-Animation/dp/0471789208/ref=sr_1_1?ie=UTF8&s=books&qid=1232302336&sr=1-1

Hyper-Realistic Creature Creation:
http://www.amazon.co.uk/Maya-Techniques-Hyper-real-Creature-Texturing/dp/1897177046/ref=pd_sim_b_2

Mastering Maya 7
http://www.amazon.co.uk/s/ref=nb_ss_b?url=search-alias%3Dstripbooks&field-keywords=mastering+maya+7

Blendshapes and modelling

Heres a fantastic quote from Stop Staring (second edition) page 103 in reference to blendshapes:

"The face is always expressing somehow. The point is this: The model you're starting with isn't your character. These shapes you're about to build are. Treat every shape like it's the only shape anyone will ever see because combinations of the shapes you build here really are all anyone will ever see."

Just thought that was inspirational and true.


Heres a quick list of the sections from the book that I will be using to help make sure I do this right:

Page 76-89- facial modelling.
(Teeth and tongue are covered after this extract, but im not sure if the models going to need to be this complicated or not. Will discuss it with Phil tomorrow.) This extract uses the polygon toolset, but if I am using a different surface material it can be easilly translated into the toolset for that. If I go with Phils idea of subdivision surfaces then it's the same toolset anyway.

Chapter 6.
Very useful section with practical advice on the best way to create the fundamental blendshapes used in a diverse mouth setup, including information on avoiding common pitfalls. (Primarily the dangers of additive shapes.) This section also shows the importance of a properly modelled face, and goes on to explain blendshape weights.

Chapter 7.
Practical tutorial on modelling and rigging the eyes/brows. (Looking at Phils designs the eyes are going to be far simpler than the setup from the book, but I will still need to go through this for the eyebrow setup.)

Chapter 12
A far more detailed explanation of making an industry standard control panel as the facial interface.


The only nugget of info Stop Staring doesn't include is when to actually do the blendshapes. I.e. before or after skinning the body rig, as I still don't know whether it's even important or not. I havn't been able to find an answer anywhere, so I'm presuming it doesn't matter, but I don't want to assume anything. I have emailed Alex Hulse this question, so I should get my answer before I have to start doing it.


By the way Phil, as soon as I finish finding out what I need to know from this book, I will lend it to you. Theres a really good section on facial animation and acting that I think you ought to read.

Basic facial control

Following a very basic tutorial from Stop Staring (second edition) by Jason Osipa, I have learnt how to do facial controls that are reflective of an industry standard setup. By this I mean a setup that uses objects in the actual scene to control the blendshapes, rather than using the blendshape window. This is highly beneficial for the animator, as it is far more user friendly. Check out this quick playblast, you'll see what I mean.

This tutorial has also given me insight into the possibilities of this teqnique. It makes use of the limitations found in the attribute editor, as well as creating and linking expressions to multiple blendshapes. I expect this can also be achieved through channel control and the connections editor, but knowing about expressions has led me to wonder whether or not this can be used to solve my concerns about using set driven keys. Also using the limitations from the attribute editor is particularly useful as you can set it so that you can move an object one way along an axis, but not the other. It's also in this editor where if you do allow it to move the other way, then when linking this to the blendshape it's essentially telling Maya to do the opposite of whatever the blendshape does. This means I only need to create one shape for the wide pose and when moving the slider the other way it will work out the narrow by itself.

It's also important to note here that this particular tutorial from Stop Staring has already removed a preconception of mine about the way the mouth opens and closes. Previously I had assumed that it would just open universally, but of course it does not. Only the lower jaw moves. Obvious now that I think about it, but still I am glad I realised now rather than later.

Needless to say this is going to be extremely useful, as I can use this same technique to create diverse controls for the eyes, brows, and a more expressive mouth.

Research

Modelling reserach:
I will need to look primarily into facial modelling, and how curves can be used as an easy control device for blendshapes. I also need to determine wether it would be worth using nurbs to account for the low poly-count limit while keeping the surfaces looking smooth, and wether the use of nurbs would effect the painting weights or uv mapping.

Rigging research:
I need to find out if theres a way of using set driven keys, i.e. for fingers, without losing control over the individual joints. I would also like to investigate the possiblities of building a basic low-rez control rig for animation that is convertable to a high-rez polished rig for rendering.

Texturing research:
I need to find examples of claymation, as the brief states they would like to see designs based around traditional claymation. This should help me decide on how to texture effectively to the needs of the comission.


This should all be done by Monday.

Starting out

Me and Phil had a good discussion, and we now have a clear idea of where to go with this. Here's the schedule we decided upon. To see what Phil's doing, click on the link to his blog and it should be in one of the early posts.

week one (this weekend) : research.
week two: First character modelled and textured.
week three: First character rigged.
week four: Second character modelled and textured.
week five: Second character rigged.
week six: (accounts for problems / delays)
week seven: set building /props.
week eight: lighting, quality testing, finalizing.
week nine: rendering / editing.

We have arrainged this schedule so that Phil can come up with the sequence design before I start rigging so that I know what the characters need to be capable of. We are to be working on developing the design almost simultaniously, though Phil needs to one step ahead of me on the design front so that I have something to work from.

Here we go again!

Ok so here we go, we have a new brief and it's a good'n: Design, model, rig, texture and animate a choice of characters worked up from a list of Zoo go-ers and residents. The maximum resolution for each character is 100 000 polys, and it's for a possible advert for London Zoo. I'm particularly excited about this one as it finally gives me a chance to focus my energies on my chosen specialism, which is character setup. We had a choice of three briefs here, but the character design one was instantly the most obvious choice for myself, as I will get to model and rig. I am working with Phil Brooks again, considering we worked well together in the last project, and we both want to do the same brief, AND he wishes to specialise in the areas that I do not. It's truly a match made in heaven. We are called Philing Vacant Dave.

I will be modelling, rigging and texturing. I have chosen these specialisms largly because rigging is an area I feel I have a particular flare for and would like to progress with it further. I also want to develop my rigging skills because it seems to be an area that no-one wants to deal with- this has already earned demand for my knowledge within the group, and I would like to think it could make me more employable to employers. Modelling is also an area I particularly enjoy, and I didn't get to do any character modelling or texturing in the last project so I would like to use this as an oppurtunity to properly have a go. I also feel that doing the modelling and the rigging will deepen my understanding of the importance of modelling to rig, which will be vital knowledge whichever field I ultimately end up working in. Also, texturing is an area that for me was under-developed in the last project, considering our simple cartoon style. While it will be similarly simple for this project, I still recognize it as an area of weakness that I would like to remedy through this practice.