Limited Collectors’ Edition C-35 Shazam Cover

August 8, 2024|In General

I was absolutely the kid that watched the Shazam/Isis show on Saturday mornings in 1974. And DC was kind enough to capitalise on the success of the show with this volume of Fawcett reprints wrapped in a photo cover of Jackson Bostwick in his Captain Marvel costume.

I’ve been wanting to do some design variations with the photo for some years, but could never find a high-res copy of the original image. Getty Images has a few of Bostwick from other shoots, but nothing of this mountain-backed scene in southern California.

Thankfully, image-editing tools have continued to progress, making my design work possible now, and we examine where we came from and where we’re headed with these advances.

Colour and packaging by me.

Limited Collectors Edition C-35 published by DC Comics

As published.

Photo editing and manipulation by Scott Dutton

Photo extracted from the published cover, retouched and extended.

AN EVOLUTION IN PHOTO RETOUCHING AND MANIPULATION

Back in the 1990s when computers were less powerful and Photoshop was young and spritely, retouching and altering photos was a very labourious task.

Before Cintiq-style interactive monitors, you were lucky if you had a regular drawing tablet. Many of us didn’t have the budget for even that and we had to work with just the mouse. No pressure sensitivity, no feathering. And the brushes were basic too.

With those basic tools we cut things out by hand, zooming in as best we could to catch that edge just right. Depending on how complex your subject was, it could take you more than an hour to get it masked properly.

Extending images to make a size or create a bleed could be quite simple if it was flat colour, or you used the clone tool to fake some more image and then you’d blur or airbrush it to blend it in, or drop a gradient over it to fade to black or another colour.

Putting in a new background was often straightforward. Machines then were more prone to slowing down as you added layers, so economy in file structure helped, and saving multiple states as layers, then flattening your current work file before adding new layers saved processor cycles and RAM usage. Because most of us didn’t have top-of-the-line boxes maxxed out.

Moving forward into the current era, we have powerful CPUs, GPUs, whackloads of RAM and massive solid-state drives. Lag is no longer an issue.

Monitor tablets with high resolution bring us ever closer to what it is to retouch and paint in real life. Pressure sensitivity, all kinds of variable brushes and material simulation bring a complexity and naturalness to digital imagery now.

And the software has moved forward as well. What we currently call AI with machine learning driven by ultra-fast super computers is being integrated into image editing software. Sticking with Photoshop (though they are not unique here), the next generation of tools has been growing over the past few years.

Use the select subject command, make a mask, you’re done the major work in as quickly as you can hit the keys. Finessing the mask by hand – if needed – is a few minutes of touchups instead of a major task.

Similarly you can use the select sky command and replace it. The dull grey day is suddenly a more impactful polarised blue. Warm up the colour balance, make the image a higher contrast and you’re done.

The successor to the clone tool and brush tricks is the content-aware fill. Using the new AI framework it more intelligently extends shapes and patterns to yield results that need less manual work done to them.

The latest game changer are the generative tools. Generative fill, with or without giving it instructions will intelligently change details or extend an image. Is your subject off to one side but you need them to be centred to fit your design? In less than a minute the tool will return with options. While it can get confused sometimes, in the vast majority of images I have had good results.

THIS COVER

The photo above was extracted from the published cover. Select subject masked out Captain Marvel in a few seconds. A little bit of retouching to fix selection errors on the cape and I was done. That took care of the yellow border and the sky. The mountain background was left in. Elapsed time: 5 minutes.

Next the yellow border was cropped off around the mountains. 1 minute.

Using the generative fill, the mountains were extended on the left, right and bottom edges. 5 minutes.

To replace the sky I went to a couple of stock sites to find one that matched the original. The chosen one was sized and dropped into the background. 15 minutes because finding the right stock image is always a fucking mind-numbing task. Total time: 26 minutes.

AI – IT CAN HAVE A USEFUL PLACE IN OUR TOOLSET

I’ve gone through this bit of very-dull history about process to demonstrate that this is my career. I’ve been a practising commercial artist for over 30 years and I have a vested interest in seeing my profession not only not disappear but thrive. I welcome anything that can help me do the mundane stuff because it saves me not only production time, but also having to abandon an image and try to find a new one (the most time-consuming thing to do as noted above.) It means reduced compromise.

Photoshop and other Adobe apps have another AI tool – Generate Image – and that one I do not welcome. Write a prompt, create a whole image. That’s been all over the news with various apps. I’m not going to get outraged here. Just know that I believe that people who write, draw, paint, sculpt, photograph, compose and make other creative things are the ones we need doing creative work. Not some person who hasn’t learned a craft trying to call themselves a creator because they can make infinite mashups of catalogued material in a database. All in a bid to finally eliminate having to pay those people they resent for the entertainment and culture they exploit. The ultimate in soulless, narcissistic capitalism.

If and when we do get true AI I hope it will be used to control logistics, financial systems, weather prediction and other massively math-oriented problems that are too big for a human mind or group of human minds to manage effectively. I don’t want it to digitise the human experience.