Question about workflow - paint to 3D model?

Hi Blenders,

I created a new small team ( by background I’m AI/game world software developer so 3D stuff is known for me, but not in all details ), and not all subtleties of 3D designer workflow are known to me, but still I’m most experienced and have to resolve most of problems in the team.

so the question about workflow.

In our team someone creates characters as painted images.
But as regular perspective paints, not 2D scketches ready to be used in frontal and side projections for model creation ( the problem is that for example it is difficult to develop front nose when nose is very small and pointed - so effectively in 2D frontal projection it is just a point but it is useless for 3D designer and painter refuses to create something 2D projected ).

how in other teams the problem is resolved?
either painter creates sketches for all 2D projections or somehow there are tools to help designers to get from ‘impression’ paint to 3D model ( using perspective texture correction tools), or just the skills of designer should be enough to get from image to model,adding missing details which painter failed to provide?

so how this step is resolved?
P.S. maybe such or similar tools

of some use in this case?

I believe the main problem here is that your painter refuses to do 2d projections. They can be very very helpful when translating the concept into 3d. The 2d projections are usually not 100% correct (they are in the end hand painted), so the 3d modeller has to take this into account and adjust the model so it gives the same “feel” as the perspective paint, even if not exactly like 2d projections. With very skilled 3d artists, you could get away with just perspective paint. But this takes a very good eye for form and depth and also a lot of experience. And even then 2d projections can be helpful.

Here are some examples by David Revoy (the concept artist for Sintel movie). With such concepts you could immediatelly go into the 3d program and start making the model.:

You could do a temporary workaround for some things. For example if you’re modelling humans, take a 2d projection of a generic human to create the base model and then add details like on your concept. But I’d strongly advise you to get the concept artist to start making 2d projections. If lack of knowledge is the issue, you could order the Chaos&Evolutions DVD from the blender e-shop. Here’s a preview video (it also deals a bit with making 2d projections)

Also, there is no need for specialized tools, the issue is elsewhere.

really useful hints, thanks

now I see more clearly how things should go.