Here is some images from the past weeks.
The large t-rex concept drawing is by Remi Strand.
The two images from our first testrender shot is made by Adrian Fjellberg, model by Remi Strand and rigg by Sebastian Antonsen.
The small chubby character is made by me (Einar Martinsen)
The robot with multiple arms washing a window is made by me (Einar Martinsen)
The clay render of the main character floating in the air is made by me (Einar Martinsen)
The lazy hungeback character at the end is made by me (Einar Martinsen)

I do not have any animation tests from Willam Christoffersen, or any material/composition testing by Xia Yefu. But check out their website, something may be there.

The subject is something that many people may want to know, I sure did before I figured out the pipeline.

Creating your 3d mesh ready for Rigging without texture, so you can adjust texture even after the animation is done!


The thing is to create a 3d model ready for rigging and animation. For weeks now I've been trying to achieve to right pipeline to make the rigger happy, along with myself as I would texture the model. I've enjoyed working with Sebastian on these models, and I'm thankful that he is a strict rigger. All the critique I've received in the early stage of making the characters ready for rigging has opened my eyes on how to make a more solid model in a technical aspect. It's a great way to work in the same studio with other artist that are going to use your model for rigging, texture, shading etc. You really need to think about what you are doing, keeping things clean, naming parts and material and keeping your files organized. Again, I cant describe how much I've learned about making a studio product.

OK, some step by step notes for making a model ready for rigging with the ability to adjust texture later on:

Smart polycount.
make sure things are standing in 0 XYZ axis.
make sure things are lined up correct in all views.
Unwrap everything, and UV map things that does not need a unwrap.
Make a multisubobject material.
Assign material to material ID's.
Name everything so other people see what you are doing.
NO turbosmooths or smooths for that matter.
talk to your rigger about attaching meshes.
Collapse all bend, symmetry, UVmap, Unwrap etc into a editable mesh/poly.
Attach mesh and make sure the different parts still have their material.
Make sure every mesh has a scale of 100,100,100 in all axis. Or apply a Xform reset and collapse it into the edit mesh/poly.
Do not group meshes.
Control check the model and send it of to the rigger.

Now you should have all your UV- cordinates and materials correctly placed even if one mesh is attached to 10 different meshes
Some difficulties may come up while rigging. Fix it and keep in mind what to do for the next model.

Now you are ready to make a new model and return to a previous one later to texture, sculpt normalmaps etc.
Time saved for the production!

(the image should explain a lot of what i just talked about)

The importance of planing ahead and save time.

Here I'm going to talk about painting a real looking 3d render without doing any zbrush/texture ect. By using techniques in photoshop that I've learned during my 3d experience you can create final "look development" of the character by rendering a flat model and paint on it using smart layer techniques, images and painting with a wacom tablet. The reason why I do this is to save a lot of time and get fast feedback on what looks best. For these two examples I've used a clay render and painted ideas and texture directly on the model. Or exported a mesh to a external render program where you can change textures and lighting in realtime fast. This technique is often used by concept development artist in the professional industry. The client is often someone who does not have technical experience in CG, and needs to see how a character would look even before you make it correct in a 3d application.

In this update I'm going to talk about pro's and con's of drawing concepts traditionally and digital.

In my personal experience its always best to start of a concept on paper. You get a lot of freedom, you can easily draw fast directly on what you see instead of drawing "blind" with a standard wacom tablet. The reason why I like to start a concept on paper is because: I have easy access to multiple papers. I can rotate and move the paper all over the place. It's mobile, I can hang paper on walls to quickly see what does work, and what does not. I can draw near to people and cooperate. To say it simple: A good amount of freedom in many ways.

While drawing digital with a wacom tablet, you lose some freedom. One thing is for sure, its easier to make things look nice. Easy to erase and change stuff. Easy to make silhouette's of whatever you draw to read the design faster. The thing that I don't like with a tablet is the way you draw "blind". You get a fake feel to where you have your pencil tip and so forth. Of course you get used to it after a while.

The other option that brings these two fine techniques together is a wacom CINTIQ! Something I've been drooling on for quite some time. Good news is that I will hopefully be a proud owner in a month or so. Cant wait =)

But for the time being, I often draw on paper, scan the drawings and paint further with a wacom tablet to make the concept look extra nice, or to create easy changes. A wacom is also awsome to "pre" paint a 3d render before moving onto texture/zbrush.

Here is some images that refers to the stuff I just talked about.

Since last update we have moved more into making models and drawing final concepts for the robots. We have also finished the first draft of the animatic. And the story looks solid and good :) I'm really looking forward to see some test in real life footage.

The past week we have been working on concepts, modeling and texture pipeline. Software that is preferred for modeling-texture-rigging at the moment is: 3ds max 2009, zbrush, xnormal and maybe bodypaint. The rest of the crew have been composting, testing materials and matching cg into real images.

I've been working a lot on modeling and figuring out the best pipeline for getting problem free models with enable texture adjustment after attaching mesh. This is for optimality towards rigging. All i can say is that I've learned about 7 new programs and functions, and the current pipeline works very well! The pipeline is a bit complex, with a lot of jumping back and forth between programs. But the results turns out to be good and for every time I do the pipeline, it goes faster.

Since last time I've finished two models ready for rigging. Normal maps from zbrush, to xnormal to max has been used alot. Here is some sample images

Since last week we have been drawing storyboards, making more concepts for characters and started the animatic. Things are moving good along with the production plan. Sebastian has been working on keeping the crew working hard, editing production plan and rigging. Remi has been testing different software for worklow while drawing consepts and storyboard. Adrian and Xia has been working on composting in nuke with renders from vray. And William has been doing a good job with animating a model i made in the start of the project, he is doing so to get the feel and practice for the loads of animation to come. Here is some images of the past week: