Forum: Poser - OFFICIAL


Subject: Modelling a face from photo?

Totoro3D opened this issue on Dec 27, 2004 ยท 18 posts


FyreSpiryt posted Mon, 27 December 2004 at 10:14 AM

For me, it depends if I'm going for a close match, or using someone as inspiration. For inspiration, I'll set up a image in fourths with front, side, 3/4 view, and a blank spot to put the character while I'm working. Then I'll load up that image as a background, move the camera so that the character is in the blank spot I left, and start playing with morphs, eyeballing a match.

If I'm going for a close match, I make two square graphics, one of the front view and one of the side. I also usually make a background image with a 3/4 view in the upper left corner to check against. In Poser, I load up two square props, rotate one 90 degrees, and apply my front view to the front-facing square and my side view to the other. Then I scale and move the rotated square around until the features on the side view are the same size and in the same location as the front view. When I've got them matched, I parent the rotated square to the front square. Load the character, switch to the front camera, set the character in wireframe, and move the squares up to the head. I then adjust the scale of the front square so that the head width is approximately the same as the characters. When I've got that, I move the square so that the top of the characters head is level with the top of the head on my square. (Head length, nose and mouth height, eye location, most of that can be adjusted with morphs, but there aren't a lot of options for head width or cranium size/shape.) By scaling the front square, the side square scales along with it (because it's parented). I switch to the right or left camera, get close to the head, and move the square set until it's in a good position to morph from. (Usually the back of the picture head even with the back of the characters head. Again, usually there aren't many morphs to affect the cranium size or shape, but the face can be moved in or out.)

Once I'm set up there, I start morphing, switching the character between a solid preview and wireframe and trying to match the mesh to my reference pics. I stay using the front and side cameras at first. Switch between them frequently; often an adjustment in one view will alter the other. Every once in a while I switch to the face camera and compare a 3/4 view to the one I put in the background. Every once in a while the front and side views will look very close, and the 3/4 won't look anything like the person.

I personally like the gray-scale my reference images so I'm not distracted by the color variations, but that's just my personal preference.