Making a 3d picture with your digi painting.

 

File preparation

Needs: You’ll need Photoshop (or something else :p) and a nice little free soft called StereoPhotoMaker (http://stereo.jpn.org/eng/stphmkr/ ).

To begin with, I will detail the process when having a multi layered psd. It will be a lot more hard if you only have one layer to work with. In this case, you have to select part you want to move and try to avoid the selection to be too obvious by using flur etc…

To prepare your file, just make sure everything is organized. I actually don’t do it cause I’m a messy mess, and I just remember the correct place of each of my sometimes 50 layers lol. But it will be more easy if you make sure each layer in the layer window of PS is in the order it appears in your picture. That means no layer A above layer B in the window, when Layer A appears farther than layer B in your picture.

 

 

Let’s take this picture as example, with its proper organized layers.

 

Let’s make a merged copy of all that that we’ll put on top of all layers for later (make it invisible of course), that we call PictureL (for left). This picture will be your left eye, and you’ll load it first in StereoPhotoMaker.

Now, the fun begins. We will create the right eye, that will decide the whole 3D of your picture, as you wish it to be.

 

Making the 3D

For that, some simple explanations.

Everything in your picture will move, horizontally, to the left or to the right. That will create the 3D effect alone. Don’t move anything vertically. Don’t move things manually!!! Use control+t (free transformation) and change the x number at the top of the ps screen (with the triangle clicked to begin the count at zero)!!


 The depth is basically 3 categories.

_The things you don’t move (meaning it’s the same position, overlapping the L picture) will appear at the glass of your screen position, like if it was flat and glued to the screen. (It can be good for text for exemple, but you’ll avoid it if you want the whole scene to look 3D)

_The things you’ll moveto the left, with a negative X in the ps stat, will seem to be in front of the screen. Those ones will be pretty killing for your eyes if you do it to much. I generally don’t use more than -20 for it. (Important notice: all the values I’m using are for a 1920x1200 picture! The scale of the gap would have to be different for a smaller image, just adapt the values. When I say -20 is my maximum –x value, it would be -2 for a 10 time smaller image)

_Thirdly, the things you move to the right will look farther in the back.

After that its pretty simple if the psd is organized well. And you just have to make sure you have the proper distances, but I will help you (even if you can play with different scales and try things)

I use those general rules to help myself and it works quite well for my pictures.

_As I said, everything you want to appear really in front, not appearing entirely on screen (where you don’t see the bottom of it) is -20 maximum (the front trees for example here).

_The character is at 20, cause it’s not too far away and yet still in the depth.

_Everything in between those two are then between -20 and 20. You’ll have to estimate how far they are from the 20, or how near. For exemple, the bush here would be at 10, mi distance.

_For the ground, we’ll see after cause there is a trick.

_For the background, it’s up to you to decide if you want something really far or not, based on the scale we used. It’s always a matter of proportion based on the value we used for the front scene.

_There,  the trees don’t seem that far, so let’s say if the distance from the character to the screen is 20 (always for me), the trees behind will begin at 30, then the one after that will be at 50, and finally the last line at 80.

You’ve understood it’s all in the art of estimating, based on the scale of your picture. My videogame wall generally goes from -20 in front to 100 in the back, but for larger scale background, like my starfox picture, I went to 200.

 

Special cases

_When you move a layer and there is no information to fill the new position. For example, if I take the front trees here.

There is a gap when moving it (in the wrong direction for the sake of the exemple). When facing this problem, well.. just fill the gap homogeneously :).

_Ground:

For the ground you’ll use a special cool trick that will give more 3d feel to your picture. You will deform it. Using the skew tool in ps. Beware to NOT change the vertical size of the layer, never ever, or it will make this (http://3dporch.com/2pdy.wiggle look at the tree, even if it’s not higly noticeable in 3d view. The wiggle thing makes every problem appear really clearly.) That means only use the top and bottom narrows to move it, never the corners points!

The skew tool looks like that. Here it’s a little tricky to not make something weird. As your brilliant mind already guessed, you will here simultaneously create many 3D positions for the layer. But how to make that with the proper distances will you think ?, Well there is a simple rule for that. Use the markers ‘other layers) and the scale values.

Here, remember, as we have a fix mark it’s easy. The character is at 20. The ground goes from the trees to a front value we’ll have to decide. We said the trees are at 30. The more easy for that is to move the ground before you skew it. Move it to +30, then skew it so it match the proper positions of the layers on top of it. That means, look at the bush and where is it, relatively to the ground. Well, skew the ground until it matches that position and everything will be all right. Make sure you do that after you already moved the layers on top of the ground, or it will be useless.

_With this skew skill, you can make anything appear as a 3D perspective, if you make sure you respect the right distances by looking at the numbers you use.

_You can also use the free transform skill to make that on vertical object. If an object as a position, and you take it’s right side to move it to its right (expanding it), it will make the object appears as it was going deeper on its right side. Now, if you use the tool to make its right side go to its left side (shrinking it), it’ll appear as if it was going toward us.

 

Finalizing

When finished with the layers placing, just export the first merged Left eye, then this new right one. Open StereoPhoto Maker and choose “open right and left images”, then load the pictures (in the right order, left and right, but you can switch after that if it’s the wrong order). You can now verify everything is right by choosing the anaglyph view (with the glasses) if you have one, or the wiggle view. When happy (there are always some corrections to do in PS), save as a mpo and put it on the web!

With all those rules and skills, you can find your own way to do that I guess!

Thx for your time.

Follow me on Twitter if you don’t do it yet :) http://twitter.com/Orioto