Sickdesigner

After the last few weeks when I focused more on the design update for the site , it’s good to get back in the tutorial game. And this time it’s a biggie: we’re going to give Venus de Milo her arms back. And we won’t be using any stock photos for it either.

The concept of this tutorial is my response to the overwhelmingly magic power of the new Content Aware Fill tool, which is part of the new Photoshop CS5. I’ve, since, seen quite a few designers/photographers becoming a bit worried that their photo manipulation work load will take a plunge, because Content Aware Fill is such an easy to use and yet uber-powerful tool.

These being said, please note that Venus de Milo has boobs. And they’re naked. I suspect no underage viewers will find trouble with their parents on account of reading this, but it’s better to be on the safe side. So, if your parents have a problem with a 2100 year old naked lady made of marble, shut down this page and go solve the Fibonacci sequence for the first 500 steps (that’ll keep’em busy!).

But, you see, Content Aware Fill doesn’t actually fill anything. It removes things. And that’s all it will ever do, because that’s what it was built for. And removing things isn’t a task impossible to transpose into an algorithm, Content Aware Fill is living proof of it. But what IS impossible is creating an algorithm for adding things. This tutorial aims not just to inform about creating stuff, but also to prove, to those who still feel as if they’re losing their trade, that photo manipulation will, for a very long time, remain a domain limited to the power and intelligence of the human mind.

The rest of you lot, scroll down and let’s re-member this chick!

I used this picture, courtesy of good ol’ Wikipedia.

01

Since, as I said, we won’t be using any stock photos for her reconstruction, a bit of planning is in order. I quickly jotted out a rough sketch of what I believe is a viable position of the hands. Granted, I’m not aware of the entire history behind the statue, so this could very well be way off in comparison with the original structure. So don’t start yelling at me that “it ain’t right”. Thanks.

02

So, here’s the gimmick of this tutorial: in a new layer, using colors from within the original statue, I sketched out her hands. This is very, very basic and, as you can clearly see, “draw within the lines” doesn’t mean much to me. For this particular project I used only one regular squinted 45 degree ellipse shape Calligraphic brush, which you’ll find in any run of the mill graphics software. I chose that brush because it gives the appearance of plaster on a wall, which is remarkably close to what Venus’ texture looks like.

03

Here’s where the fun begins: using the same brush but at 50% opacity, I sampled some more original colors (mostly random at this point) and refined the sharp edges of the last layer.

04

To achieve natural lighting, I would zoom out to about 15% & squint my eyes to see whether the arms look proper. Alternatively, you can flip the image horizontally.

And again, with smaller brush sizes and more zoomed in. Notice I still don’t care about drawing “within the lines”. At this point I feel lighting and coloring are far more important than shape.

05

Fact: old statues are full of cracks and dents. Using a dark brush (not black!) I sketched out a few imperfections. Try to make them harsher than the ones on the real statue. You’ll see why in the next step.

06

Whoa! Where’d the cracks go? A: ever hear of Subsurface Scattering? Until a few years back, I hadn’t heard of this term either. It comes from the world of 3D software and refers to an underlying layer that is only partially visible, thus the term Subsurface. Well, I applied that same idea to Venus. Because her natural cracks seem like smudges in many places, with lots of variable widths and coloring, drawing full pixels would’ve rendered a very harsh outcome.

07

Then, using more of the technique from step 3, I blended her new body parts more and more.

08

I got rid of the sketch layer and masked out the arms. Take that, drawing within lines!

09

I realized that while the new appendages were fine, lighting wise, their color variance was..well, a bit invariable. So, in a new layer, I used a dark brush to detail some highlight area and desaturated them a bit by setting this layer to Hue.

10

Obviously, the new parts cast/receive shadows. Well…mostly just the right arm casts a shadow on the body. The other minor shadowing has been taken care of already through the layering process.

11

And with a bit of lighting correction, a cool vignette and some Curves for contrast, we’ve got ourselves a whole lot of greek woman.

Well, there you have it, folks. It’s, obviously, not the best reconstruction work, but I’d like to see a tool in Photoshop that can come even close to doing this without any stock photos. No, really, I would, it would be fascinating. Until next time, may the sickness be strong with you, my little padaw…grass… …. friends.

by Radu Chelariu