Having fun and exploring the power of Photoshop Generative Fill
I gotta admit, being a designer feels a little magical lately. There is so much going on with AI that will drastically change the way we design.
The same goes for dealing with images these days. Not only can we generate whatever creative stuff our minds come up with. But with the introduction of Adobe Photoshop’s Generative Fill we can alter images now as well!
And let me tell you, is it definitely fun to play around with! No worries, it’s also useful.
I’ll take you along a little explorative journey I had last weekend and show you a bit of fun and a bit of useful stuff as well.
My main question for Generative Fill?
What can Adobe’s new Generative AI feature do for interface designers?
Is it only a lot of fun? Or can we make it meaningful and help us improve our work?
How good is it? And is it ready to for professionals usage?
Those are just some of the main questions I had about Adobe Photoshop’s new AI feature called Generative Fill.
To start things off and state the obvious: yes it’s amazing! Here’s a little preview from Adobe themselves, for who’s still unfamiliar with Generative Fill.
Starting with a bit of fun.
It all started with a bit of fun, like a lot of good things in the life of a designer.
Doing the opposite of what everyone else does
We’ve all seen the examples where you can let Photoshop’s Generative Fill, fill out blank area’s and as you could call it ‘complete’ your image.
But what if we turn that around? We get a picture. Delete the inner part. And ask photoshop to fill in the blanks. Why? For no good reason other than a good healthy dose of curiosity. Because, that’s where good things start with.
Here’s a few fun examples or what I like to call, reverse generative fill.
On to more fun
How powerful is it in altering images?
Let’s see of adobe’s new feature can help us turn awful stock images into something fun.
This is actually were it got interesting. I realised that changing little details like the color of the cap can make an image on-brand or not on-brand. Or matching with your UI or not a match with your UI. More on this later.
Time to get serious: what can Generative fill do for UI design professionals in their day to day job?
While these explorations above all seem like playing around. I needed to know if Adobe Photoshop’s Generative Fill feature could help me with the following.
This is where it gets interesting
- Can we change elements in images which get’s them more on-brand with your brand guidelines? Not just in terms of colours, but by adding and removing objects as well.
- Can we change colours to get images more on brand or let them match the UI a bit better.
- Can we remove or add elements, to get images more onbrand?
- Can we make stocky images less stocky with Generative Fill?
Getting images on brand by changing colours
Before we show some before-afters, do you know what some of my colleagues call designers? Picture-scrollers. Because we, designers, scroll. A lot. Looking for the right picture. The right accent colours, composition, light, the exact art direction we’re looking for.
Looking at the examples before, it kinda look like Photoshops AI feature can now help in this matter.
I think it’s pretty amazing—especially the first image above. The yellow sweater, the dog, and the changed painting on the wall make the image less stocky and more fun & energetic.
Actually using generative fill to… fill.
You know what else makes designers scroll for hours and hours? The wrong way an image is cropped. This especially sucks for backgrounds.
Example. I was looking for a bowl of fruit for a UI design. I found this image.
What I don’t like about this image is the lack of whitespace around the bowl.
I need space to put a title and other UI elements.
I used Photoshop to cut out the bowl, and ‘ask’ for several different backgrounds.
This enabled me to actually use the entire image as a background, rather than a cut-out bowl.
Or do both: use the entire image with the actual tablecloth it sits on, including the shadows, but also cut out the image to create some interesting depth.
Final words & conclusion
I was actually tempted to say it’s extremely useful for designers. Especially for the use-cases we spoke about above: changing a color, adding some details, removing some details, all to get images to look more on-brand or less stocky.
But… if we zoom in we notice elements look a little shitty often.
It doesn’t look like it nearly as far as Midjourney currently is. Especially with the recent release of Midjourney 5.2
But nevertheless: Adobe Photoshop’s Generative Fill is very valuable. Sometimes you get shitty results, but sometimes it looks like magic.
It requires a bit of fooling around in the software, but its definitely worth it.
A final use-case
We’ve establishes that designers, scroll. To find images. The perfect one. And sometimes, when you’ve found the perfect one, you need two more.
Introducing, Adobe Photoshop Generative Fill, enabling you to create more ice-cream images.
Can you tell which is real?