Good day everyone, welcome back to another instalment of my studio 2 development series. I do apologise, its been a little longer than normal between updates.
Today I want to talk about part of my skill-set that I am focussing on trying to improve and my research into how I can improve it, and that is, high poly sculpting. Also I want to talk about the 3D asset production pipeline and how I am currently implementing that into my workflow.
*Note for studio 2 facilitators – the following pertains to ANM220.LO07*
Typically, a contemporary 3D production pipeline would look something like this:
When I found this image, I was a little sceptical, looks very stylized but when I really looked at it it happened to be the best example of what the 3D production pipeline really is that I had come across.
My focus is on the modelling and texturing phases of the production pipeline. Currently I use a very circular workflow for achieving the best results from sculpting. I either build (in 3DS Max) or utilise a previous base mesh as a starting point and take it into Mudbox to sculpt/shape and otherwise make quick alterations. The altered mesh then gets retopologised (either in Mudbox with their dynamic retopology tools or manually in 3DS Max [Depending on my needs]) and resculped with more detail. This process gets repeated as many times as necessary to achieve the results I am after. I find this process very useful as it allows for a very rapid and as-you-go iterations on your designs helping to give a stronger design at the end of the process. Below is an example of this process:
As you can see the idea evolves rapidly as each iteration is completed and then modified. After this process is compete I then move on to create a working model from the high poly sculpt and back out a normal map, ambient occlusion map and material ID maps. I then take the model and the baked maps into Quixel Suite where PBR materials are produced and exported ready for the model’s implementation.
*Note for studio 2 facilitators – the following pertains to ANM220.LO05/6/9*
Sculpting is an area of 3D modelling that I find very interesting for a few reasons. Firstly, I find it to be a great way to very quickly produce the shape you are after in a model. The tools available to you in a 3D sculpting program make it very easy to take a base mesh and very quickly alter it to become what you want it to be, add details and get the shape. Secondly, it is, in my opinion, the best way to incorporate extra details into your model. Lastly, it’s an intuitive and artistic approach to modelling, it’s less process based then polygonal modelling in something like 3DS Max and more artistic, you can achieve shape and detail on a model just like you were painting it on (which just happens to be what you are doing).
From my research I have learned several things about how to better improve my sculpting skill-set and subsequently improve the overall quality of my work. Similar to when I was first learning the Quixel Suite and how it interacted with models and affected the workflow for creating usable art, sculpting also requires some changes to the way I produce my models. Sculpting programs utilise subdivision of surface geometry to quickly add more workable data to your model allowing you to add in more fine detail to your work. I have learned that the key to achieving the best results from sculpting models is to use a base mesh prepared specifically for the purpose of sculpting. This is to say a base mesh that has geometry that can effectively be subdivided without creating errors in the mesh or ineffective results. Quad geometry is the easiest for algorithms to effectively subdivide and thus maintaining a mesh of quads is an important part of preparing the mesh for sculpting. Additionally to this it is useful to put in edge loops around areas of the mesh that will be for different parts of the mesh.
In terms of the actual sculpting workflow, best practice is to do as much work as you can in the lowest subdivision levels you can. I struggle with this as I like the model to look ‘pretty’ and the deformations to be precise and so it is something I am working on as I can understand the benefits to why it is done this way. Working in lower subdivision levels on the mesh means you are effectively moving more geometry at once. If you move an undivided quad you have to move 1 surface, subdivide it once and you have to move 4, subdivide it twice and you have to move 16, and so on and so forth. Thus working at lower subdivision allows you to more efficiently alter your mesh and then move up to a new subdivision level when you can no longer achieve the results you are after.
The final key point that I unearthed through research that I want to talk about is anatomy. I don’t mean this in a specific way such as the musculature of a creature (well I do but I mean it more broadly) but I am talking about general knowledge of the subject matter. If you are sculpting an object it is important to know about that object, what it is made of, how it works and all of it’s intricacies. If your object just happens to be a human then this would be bone structure, muscle structure, fat repositories, the affects of age and movement and so on and so forth. The better you know your subject matter the better you will be able to create it. It will allow you to add in the subtleties that take good work and make it great and great work and make it exceptional.
I know it was a long one today but hopefully you found it interesting. Links to some of the research I have done and helpful information are below.
Anyway, as always,
Till next time,
James Day – 1002467
Digital Pipeline Research