“Coffee Break”, part one

This entry walks through the creation of this image, “Coffee Break”, which will be printed on fabric and transmogrified into an art quilt. As usual, click on images to enlarge if desired. (One of these days I’ll get around to fixing the site’s style sheets and layout. One of these days. Probably about the time I get my sewing machine repaired.)

 

As with much of my work these days, Coffee Break was created with a combination of 3D graphics and Photoshop. Here’s the layout in wireframe mode, showing the models in the scene.

 

Assets

If you work in 3D, you either have to make your own models or find them readymade. I did the latter in this case, using purchased models of a fairy, wings, hair, and the forest in which she’s relaxing.

The fairy model is the digital equivalent of a posable doll. I like to credit those who provide or create the assets I use. Alas, this model is merely credited as a “Daz Original”, so I have no idea who made it. It might have been created by a Bulgarian artist working in a shed, a corporate slave at Daz headquarters in Salt Lake City, or someone else entirely.

Regardless, I appreciate it. When I need to make a specific figure, I do, but otherwise the process is so time-consuming that it’s nice to be able to get figures “off the shelf”.

 

This fallen tree is part of a construction kit that includes trees, vines, and little chunks of terrain. The pieces can be arranged as one pleases to create a custom environment. It was created by Stonemason, aka Stefan Morrell, an artist whose models and scenes are much beloved. He’s an interesting person to read about in his own right:

 

Arranging the scene

I began by arranging the fallen jungle tree in a sunny field, then plunking my creepy bald fairy on top of it. The resulting image is mildly disturbing. It’s a good example of the adage “you have to start somewhere”.

One of the nice things about working in 3D is that we can rapidly try out different lighting, figure poses, models, and camera setups. We can keep the elements that have potential, such as the tree and the fairy, and change the things which aren’t working.

 

The second iteration of work. The fairy’s pose looks more natural; I’ve moved her left hand so that it’s draped across her thigh instead of stiffly hovering above the tree.

She now has hair and wings and her eyes no longer have a hideous staring quality. I’ve added some trees in the background, so we have a sense of place. The scene now feels more natural than having a fallen tree sitting out in the middle of a meadow.

However, we’re not done. The lighting is poor and the waterfall background behind the trees isn’t a good addition to the scene.

 

I’ve added a light source in the upper lefthand corner of the scene, beyond the edge of the picture so we can’t see it. The light gives the scene some contrast and dimension.

Her hair shows up better beneath the light, but the color is uninteresting.

A tree in the foreground gives the scene some depth. I’ve also turned on the depth of field setting for our virtual camera. Our fairy and the tree she’s sitting on are in focus. However, the trees in the foreground and background aren’t. This helps draw us into the scene.

The fairy now has a book, a fancy-looking volume bound in leather with gilt embellishment. This helps develop the story behind the picture a little. We aren’t just looking at a random bored fairy who’s sitting on a log; we’ve caught her in the midst of doing something.

The type of book tells us something about the fairy: she appreciates classics. Perhaps she’s even reading a volume of fairy tales. If she had a stack of magazines or a trashy-looking paperback with a studly bare-chested male elf on the cover, that would tell us something different about her.

 

The fairy’s hair is now red, which is a little more interesting than the blonde.

I’ve added a coffee cup, which adds to the story: we’ve interrupted a fairy who is taking a coffee break or simply trying to have a quiet moment or two to read. Perhaps she’s been trying to dig into her book for ages, but when she sits down for a break at home, her kid starts bellowing about the oh-so-stressful mission he just completed in Far Cry 5 for the ElfBox One. She feels like her head is about to split open and some not-very-dainty things are about to erupt from her delicate lips, so she’s flown off into the woods for some quiet time.

I debated with myself about what kind of cup to give her. The possibilities are limited only by what one can download or create. How about a coffee cup shaped like a cat’s head or a feminine china cup painted with roses? As with the book, each option tells us something about the fairy. However, I didn’t want the cup to become a distraction, so I opted for something simple and white.

As is often the case with this kind of work, once you fix a few problems and rough in the major parts of a scene, other issues become obvious. At this stage I noticed that her back wing was at an angle, which makes it hard to see. That got added to the list of things to fix for the next iteration.

 

Here’s the final scene in wireframe mode. As my husband puts it “No, it wasn’t faked in Photoshop. It really is a scene put together in 3D.”

 

The entire scene as viewed from the top. We can’t see the fairy, but she’s at about the middle of the picture, hiding beneath the trees.

 

Here’s the scene viewed from top with the trees turned off and some color turned on.

The essential components of the scene:

  • Camera circled in pink. This behaves much like a real-life camera, with settings including f-stop, focal length, and depth of field.
  • Fairy, book, and props circled in blue
  • Light source circled in yellow. 3D programs offer many options for lighting scenes, such spotlights and point lights (similar to a light bulb). I usually find them annoying to position and adjust, which is no doubt a character flaw on my part. Instead, I often plop a plane in the scene and make it glow as though it’s a light.

 

Our fairy scene with a very minor adjustment to back wings, so they’re more visible.

I rendered this at 6300 x 5400 pixels, 150 dpi. That’s large enough to fill a 42 x 36” piece of fabric. So I’m done now, right? I spent hours putting this scene together and I let it render for 20 hours. Surely it’s ready to send off to the printer?

Ha ha ha ha ha. No. In some ways, the fun has only begun.

 

Post production

Look for problems

Maybe other people don’t have trouble with their renders. I always do. Always. Always. Even when I inspect the scene over and over and over again before rendering, I find problems. Sometimes they’re bad enough that I make adjustments in the 3D program and do another render. Other times I address them in Photoshop.

Here’s a typical problem, stuff poking through other stuff. When I zoom in on the picture at 100% and inspect it, I see moss and a leaf poking through the book. That might not be noticeable when the entire image is the size of a postcard, but when it’s blown up large enough to go on a wall, it’s a problem.

Ideally, one would find that kind of thing before blowing a day on rendering. However, if not, one can sometimes fix it after the fact in a program such as Photoshop.

 

Brighten up the scene

The picture looked a little dim for my taste, so I brightened it up. To do that, I opened the picture in Photoshop, duplicated the layer, and set the blending mode of the duplicate to Color Dodge at an opacity of 50%.

One can try different blending modes and opacities. It’s a good experiment; sometimes there are happy or at least interesting surprises.

 

Add a background

Right now the background behind the trees is black. That’s okay, but I wanted just a little visual interest so the scene would feel more realistic. However, I didn’t want the background to be so cluttered that it would fight with the foreground for attention.

I began with a large version of this picture, which is somebody’s imagining of a jungle. (It’s part of the Heart of the Jungle background set.)

 

The jungle picture after a gaussian blur in Photoshop

 

I could get the jungle background in my picture by adding it to my 3D scene, starting a new render, and waiting another day for the computer to finish that set of calculations.

However, I’m lazy. It’s faster to mask out everything in the foreground of my picture and drop the background in with Photoshop. Using a mask such as the one above is a nice way to isolate elements and make edits after the fact. It’s often faster than rendering the scene again.

Daz users: see my entry “How to make masks in Daz Studio/Iray”.

 

Here’s the scene with the jungle background dropped in. It’s a subtle change but it adds a little life.

 

Adding steam from coffee

Our next task is to give the coffee some steam. Again, one can use props and do this directly in one’s 3D program, but I prefer to do it in post.

The steam is a graphic from Textures.com, a wonderful resource for those who do 3D or other sorts of graphic work. I’d post a photo of it, but I’m concerned about violating their terms of service. Here’s the link.

 

I’ve pasted the smoke/steam on a separate layer. Its blending mode is screen, which makes the black background of the smoke disappear without further effort.

 

Next I stretch it, rotate it, and erase the parts I don’t want in the picture. Voila. “Steam”. Only you and I know that it’s actually incense smoke which someone was kind enough to photograph and share.

 

Here’s the scene with steam added. It’s a small touch but it adds a little life.

 

Adding light rays

I want some light rays. They should come from the upper lefthand side of the screen, mimicking our light.

There are plenty of tutorials on making light rays in Photoshop. Here’s a good one.

 

I begin by making a new layer in my Photoshop document, setting my color palette to black and white, then rendering clouds (Filter -> Render -> Clouds). I then go to the menu bar and select Image -> Adjustment -> Threshold and hit “okay.” (All of this is in the tutorial I linked to above, btw.)

If I’m feeling wild, I’ll select a portion of the graphic with my marquee tool and resize it/expand it so it fits the screen. Different black-and-white patterns yield different results, so it’s good to experiment with them.

 

Finally, I take the layer with the black-and-white graphic and go to Filter -> Blur -> Radial Blur. I put the center of the blur near where the light source should appear, set the blur method to zoom, and fiddle around with the amount until I’m happy. There isn’t a preview mode on this blur, so sometimes one has to CMD-Z or CNTRL-Z, go back to the radial blur filter, and try a new blur amount.

This sort of looks like light rays. Sort of. But now I have a new problem – how to get the light rays in the picture without obscuring the items in the foreground.

 

Yep. It’s mask time again. This time I’ve made a mask which isolates the fairy, her props, and the foreground tree. I apply that as a layer mask on the light ray layer and set the layer blending mode to soft light. After I play around with the opacity, I end up with this:

 

Adjust brightness balance

It feels like the scene is most of the way there. However, the tree at left is bright and distracting.

I’ve muted the foreground tree out with a layer mask. But – oh, great – now the tree behind the fairy looks too bright. Shall I make another mask for that?

 

No. I’m sick of fiddling around with masks. It’s brute force time. I grabbed the burn tool and darkened the tree by hand.

 

Adjust eyes

Her eyes seem a little dark. She’s a fairy. Even if she’s having a rotten day and just wants to drink her coffee, she should have sparkly eyes.

 

Here are her eyes after using a dodge tool on the midtones. Now she looks depressed but with, you know, sparkly eyes.

 

Darkening of graphic during printing process

My goal is to print the coffee break scene on fabric. However, from sad experience, I know that when I send out images to print on fabric, they often come back dark. Here’s an example from a different project.

 

Check out that boot. That’s the boot in the file I sent to the printer, with subtle highlights and surface texture.

 

Here’s what I got back on fabric, an amorphous blob of black. Guess what? The fabric printer doesn’t do subtle highlights, at least not if they’re on the dark end of the color range.

 

Similarly, here we have a pile of rocks. They are distinct, individual rocks, happy in their grayness and individuality.

 

Aaaaand here’s what we have after printing on fabric. Another black blob. I can use thread to fix some of this, but it’s vexing.

No, it isn’t unusual for prints on paper to look different from the images one sees on screen. However, this is the most pronounced case of that I’ve ever seen. Perhaps its related to the fact that we’re printing on fabric, not glossy paper stock.

Regardless, anything at about RGB 70/70/70 or less will be black. Thus, the next step is to go through the scene with an eyedropper tool and the info box displayed. If I don’t want something to turn black, I need to adjust it so it’s brighter than 70/70/70.

 

Here’s a shrunk-down version of the file sent to the fabric printer. It’s 6300 x 5400 pixels, 150 dpi, which is large enough to print on a one yard length of “Cotton Poplin Ultra”.

 

Here’s the printed fabric. See any difference? I sure as heck do. As usual, the darkest tones in the design are even darker more after printing on fabric. In particular, the light rays I so carefully, lovingly put in are all but invisible. I’ll see if I can pep them up through my thread choices, but it’s annoying. I specifically corrected for that. However, I didn’t correct enough. Let the designer beware!

Next stop: quilting.

Leave a Reply