Large Scale Central

AI in the Hobby

Hi Cliff,

So as someone who has been doing 3d and general development for a living for nearing 30 years now, and have been directly affected by AI (some clients are now using AI for things they would have come to me for, and others are flirting with the idea ), I am not really a fan of AI.

In many cases, the AI gives less than stellar results, and there have been clients who come to me with things AI botched in some way to fix it.

I must say though, that character mesh extrapolated from the 2d sketch is impressive.

As far as the rigging, while you can certainly add an armature to that mesh, it is not an ideal starting point, we generally start with T-Pose when building a character mesh, for several reasons, just one of which being the weight mapping, which in T-Pose, can be mirrored, so basically, half the work, and uniform across the lateral axis.

Also, in TPose, the limbs are well separated from each other, and the torso, in a uniform manner, so auto-weighting will do a reasonable first pass without too many artifacts to clean, and those that are there can again be mirrored, so half the cleanup.

With this pose, you may encounter issues to get a good rig weighted, such as at the thighs, which are rotated differently, the arms, etc. Sure, it can be done, and maybe easily enough if you aren’t a perfectionist (as I am), but to do it well will take a lot of work I expect.

Either way, the resulting print of this character looks very nice, I look forward to seeing it with paint (and glad to see there still remains room for some human craftsmanship in this project).

Also, this is an interesting thread, on a topic which affects me as a media developer, and also as a music composer. I have stayed out of the discussion though, until this point, because I am rather biased on this topic, being someone directly affected negatively by it, on multiple fronts.

Regards,
Dave