|
Preparation - I started off with a collection of hi-res photographs of the person I am going to model. This is important as I will be using camera projection method on this project. By having a good front, side, quarter and other views in raw format would be a great start. First of all, the raw format photographs would provide focal length info which will be useful for matching up the model with various camera views. Also, we could rely on the extra info on the raw images to remove shadow and highlight on the photographs for hi res texture projections. I would explain in details at the later stage on how I do this. So, it would be good if you happen to have these hi res photographs ready.
The entire process will use Maya for Modeling and Hair. Photoshop and Bodypaint for Texturing, ZBrush for sculpting and Renderman for Maya for final render. One little part that is different from my other projects is that, I wanted to achieve the final look without depending on comping the different render passes together. I must admit I did do a bit of color grading and adding a bit of film noise to better blend with the actual background image, but no render passes were used in the final image.
Modeling - To begin, I imported a generic head model into Maya. I have five photos of the person I am going to model in left, quarter left, front, quarter right and right views. Create five cameras and attached these photos as camera background images. It would be a good idea to make these photos in square size to avoid any accidental scaling. You will have to reshape the model, at the same time adjusting the cameras to match the head model. This process usually ended up with lots of frustration as you will most likely not being able to get perfect alignment. Well, one of the reasons is the actual human model might move or changes his expression slightly from one photo to another. So, you will have to decide from an artistic point of view to choose which will be the master view to follow. For me, I had a pretty good matching up on front and quarter views. Having bit of a trouble getting the ears right.
We will reuse the camera setup in Maya for texture projection inside Bodypaint. Before you begin adjusting cameras, please read my other tutorial on how to do camera setup in Maya so that it is transferable to Bodypaint. The link of the tutorial is at: http://www.sweekim.com/Tut_MatchingCamera1.html
Once you have got matching up done, the next step would be the actual modeling works. It really depends on the quality of the generic head model you have. You could reuse some of the area but for me, I prefer to start from scratch, using the generic head model only as a guide. I started off with modeling the eyes, nose, mouth and ears as separate pieces. I wanted to not worry so much of how I am going to join them together, but focusing more onto the flow of the facial muscle, getting the topology right. Once you are satisfied, you can begin joining the pieces together, at the same time carefully optimizing the model. As I will be having a rather close up on the final image. Be very faithful to the actual photograph references and you will notice human face is never symmetrical.
I used UVLayout pro for UV creation. The UV tool in Maya could also do the job rather easily especially using the later version of Maya. If you look at the way I layout my UV, I had nose and lips cut off as separate UV pieces. The reason for that is most of the time you won’t get enough space for details in this area, especially at the tip of the nose when you have to compromise yourself when trying to squeeze in everything into one UV piece.
Texturing - This step requires cleaning up of the photographs you will be projecting onto the head model. Basically, its a process of removing shadow and highlight. There are many ways you can achieve that. My practice would be using Adobe Bridge Camera Raw editor to quickly flatten the shadow and highlight as much as possible. Make a preset of the settings and apply that to the rest of the photographs. Load the photograph into Photoshop for further manual flattening. Remember, there wouldn’t be a way to flatten all area. You will need to regionally removing the shadow and highlight as much as you can.
Once you have finish cleaning up the photographs, you can begin loading the cameras into Bodypaint and start projection. I have written a tutorial on how projection works inside Bodypaint at: http://www.sweekim.com/Tut_ProjectionTexturingPg1.html
Well, if you have followed closely till this step you will notice that I am relying heavily on the photographs to provide me modeling and texturing information. By the way, I am using 4k resolution on all my textures as any higher resolution might be pushing my system too far and slowing me down.
Next, I will begin sculpting inside ZBrush using the color texture created as a base for using it as alpha for creating the skin displacement. It is always a good idea to add layers inside ZBrush for various part of sculpting, such as wrinkle layer, skin displacement layer and etc. I choose 16 bit TIF for my displacement map format. You can of course choose 32 bit TIF but that would mean more constraint in editing the image inside Photoshop. I wouldn’t go too much into details on how I did the sculpting inside ZBrush as I guess it wouldn’t be any different than most of the tutorials out there.
|
