Crossing the uncanny valley WIP

Become a member of the CGSociety

Connect, Share, and Learn with our Large Growing CG Art Community. It's Free!

Thread Tools Search this Thread Display Modes
  06 June 2012
Thumbs up

the topology it's good , with skin pores details and wrinkles , i think it's will be impressive.
keep it up...
  06 June 2012
so working on the shader, id like to share a trick i've picked up from DICE that was used in battlefield 3 for faking sub surface scattering, but it can be very useful as a map inside the SSS2 shader.

here goes. you'll need to reverse normals on your model, so its inside out and than bake an occlusion map. simple as that. in the baked map, the most occluded areas will be the tightest ones - these are the areas where the tissue is thinnest (considering that your model is inside out) like ears, nostrils,eyelids etc.

all that is needed afterwards is to invert the map in photoshop and its done.

much better solution than having uniform sss all over the model, also much easier and accurate to setup than painting such map by hand. its great for controlling the amount of the reddish color bleed effect and focusing it into the areas where its needed (ears, nostrils,eyelids or fingers-in case you use it for a full body model)

*the map above is not yet applied to this render

Last edited by kybel : 07 July 2012 at 10:20 PM.
  06 June 2012
Hi Lukas, could you explain your process about camera matching in pftrack a little more, please?

It doesn't seem like that step is necessary with the great results you've been getting from Photofly and Maya, but it's certainly a very interesting looking use for image modeling.
  06 June 2012
photo-shoot successfully done. at last!

gear used :
canon 5D camera with 85mm prime lens and circular polarizer attached ; 3 lights with large soft boxes ; reflective plane for catching some illumination under the chin.

results after some retouches in cameraRAW and photoshop

Last edited by kybel : 07 July 2012 at 10:19 PM.
  06 June 2012
first projection mapping tests under way:

Last edited by kybel : 07 July 2012 at 10:18 PM.
  06 June 2012
Progress looks good, cant wait to see more!

Nick Marshall
Head of Environments / Generalists
Double Negative :: Vancouver
  07 July 2012
a word or two to the texture mapping process

i arranged all my photos and cameras in pf track usin the motion capture node
exported the scene to maya where i will be projecting my textures. i wanted to use zbrush spotlight at first, but since it is poly paint based, its just plain useless to me, cause it wont give me enough resolution. so i ended up using maya which gives me the freedom to export as huge res maps as i want. (still not decided if 8K or 6K)

the process of perfectly matching the photo with the model is as easy as in zbrush. only im not deforming the image, but the model itself

the quick and easy texture baking option found in vray for maya is also a huge plus

Last edited by kybel : 07 July 2012 at 10:18 PM.
  07 July 2012
Hi Lukas, all your images are dead. Imgur seems like they dropped the and replaced it with You're going to have to go through and update your links sadly :/
  07 July 2012
hmm, lets see.
and now ?
  07 July 2012
They're back up.

Did you polarize the light sources? Also, did you calibrate your camera's color response curve, as well as the blacks/midtones? That's the only way that you can be sure that you have accurate colors in your image.

Also, would you mind expanding on how you're matching the model to the images? You say you're using the motion capture node, however I haven't been able to emulate your results.

Last edited by NextDesign : 07 July 2012 at 12:09 AM.
  07 July 2012
nope. its pretty hard to get a hold of such large polarizing films,

well in pftrack i used the trackers on his face to calculate camera positions (with the motion capture node). then i export the scene with cameras and locators, imported the model and positioned it as best i could.

when it comes to projection mapping, i just pushed and pulled some areas of the geometry so the model and image is aligned perfectly. i did it separately for each camera. the model ended up messed up of course, but i cared only for its uv space, not xyz space.

texture map is about 80% finished. i spend quite some time on the ear texture this evening.
  07 July 2012
Really like what you're doing here, very impressed with the modeling accuracy

Can you explain a bit more on your SSS2 shader and maybe share a setup? and also explain how that inverted map fits into it? Did you plug it into the backscatter channel?
  07 July 2012

i will post some shader details when i come up with something more final. so far im just testing stuff and experimenting with values...

the inverted map should be very useful especially in the scatter color slot
  07 July 2012
Very interesting worklfow, I will try this too.
My Portfolio
  07 July 2012
I cant seem to create that inversed occlusion map. It seems when I inverse the normals the resulting occlusion is a white neck fading to an almost black head?? Any idea what could be going wrong?
reply share thread

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Society of Digital Artists

Powered by vBulletin
Copyright 2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump

All times are GMT. The time now is 06:30 PM.

Powered by vBulletin
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.