Website Images & Content ©2024 by Robb Schmidt / All Rights Reserved

From the "X" Files


The Joy of Multiple Image Generations

Artistically speaking, AI art websites that offer you several versions from your prompt are actually far superior to those that don't! It's always better when you have choices to choose from.

Since I decided to take a bit of a hiatus from publishing my worX at the Deep Dream Generator, I thought it would be a great time for me to go eXploring other AI websites, and I'm glad I did!

I've discovered that many, (if not MOST), of the other AI sites offer artists a tool that DDG doesn't, and it's a VERY IMPORTANT TOOL when creating AI artwork.

When you type in your prompt at one of these other AI sites and generate an image, you don't just get ONE image from it, you get FOUR! If you've only created at DDG you might not think that this is important, but for those of you that have also gone wandering off to one of these sites, you already KNOW how wonderful this is!

Now that I've used this several times elsewhere, I'm actually really surprised that DDG doesn't do this also. Having several versions from your prompt, right from the get-go, lets you chose the one that best suits the image you might already have in your head, and then you can continue your generations from there.

Artistically speaking, it is AMAZING, and ALL AI art creating websites should offer this feature, including DDG!

I certainly haven't been to ALL the AI sites, but I can tell you that BING offers this, Midjourney offers this, and Playground offers this. Today's X-File will discuss my eXperiences with the latter. If you're not familiar with Playground AI, you can find it here: https://playground.com/

Of course it can feel a bit daunting when you first visit another AI site, as everything feels foreign. They all seem to do things in their own specific way. But if you have eXperience in creating with AI, (which I'm sure you do), it all comes to you pretty quickly.

In case you'd like to eXplore Playground yourself, I'll walk you through a brief eXplanation of the site's creation page.

One of the first things you'll notice at Playground is that there are 42 presets to chose from, including the "none" option.

A winter scene filled with trees and snow that morph into one another. A sort of combination of the two.

You'll find 42 different Styles/AI models to choose from at Playground AI

You'll find these following options along the right side of your screen, along where you enter your prompt, negative prompt, and other options. You'll notice at the bottom of my screenshot that I can make 1000 images per day. This might be because I now have a paid subscription at Playground, but I don't seem to recall hitting an image limit while I was creating there for free.

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

(The Playground options along the right side of your screen)

To begin with, I thought the option to make 1000 images per day was absolutely ridiculous! Who on earth would ever make so many images in just one day? But now that I've made a few creations at Playground, I totally understand the reasoning behind it.

It's what this article's about, having the option to generate multiple images from a single prompt, seeing the different pictures the AI creates from it, and then evolving the images that you find interesting.

I haven't come close to my 1000 image limit yet, but I do recall that one day I managed to get up to 250! I'll tell you how in a moment, but first I'll show you some of the options on the right side of the screen...

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

You have the options to choose the Model type, the image Dimensions, how closely you want the AI to follow your Prompt instructions, and the Quality & Details of your image.

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

You'll also find a Refinement option, a Seed option, a Diffusion Sampler, and most importantly (to me) an option of how many images you want the AI to generate from your prompt!

The number of images option is truly amazing, as when you use it you'll get 4 different versions from your prompt to chose from. Don't care for the first set of 4 the AI gave you? Just generate 4 more images and see if you like any of them then! You can just keep doing this over and over until the AI creates something that interests you.

As you see each set of 4 images, you can change your prompt as you like before generating the neXt four, or you can choose one of the 4 images and click on "make variations" and then the AI will give you 4 more versions of just that one picture, and so on.

You can really 'fine tune' how you want your image to look and feel this way, slowly evolving it with each generation of fours.

I'll show you what I mean.

Here is one of the set of four images that the AI gave me from my prompt.

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

A set of 4 images from the same prompt.

I had generated a few sets of 4 before coming to the above set, and nothing was interesting to me, but when I saw the four smiling girls in front of that psychedelic sun, well, the image sort of "spoke" to me and I wanted to play with it!

So I clicked the option to "create variations" of just that picture and the AI gave me four more images created from that specific version.

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

A set of 4 images from the same image.

Then I could just continue on my way, choosing my favorite image of the four options and making variations of it and changing my prompt, (or not), with each new set of four images the AI created for me.

I thought the girl in the white dress should have a big yellow "X" on it, so I changed my prompt, chose my favorite version, and made 4 more images to choose from.

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

Ohhhh, how eXceptional!

Then I just continued on my merry way again, tweaking my prompt and choosing the version I liked from each set of four.

After generating 76 images this way, I eventually ended up with the following finished work. It took a bit of time to create, but it was wonderful to have the option to create in this way, as I am happy with how it finally turned out.

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

"Sunshine Sisters"

I think every artist at DDG can probably attest to the frustration that is often involved in trying to create something they like while burning through all their energy points in the process.

This system at Playground.com is far less stressful and aggravating, and far more satisfying with the actual creation process itself. In fact, it's a FAR SUPERIOR WAY of creating AI artwork than what DDG offers.

You end up with something you like every time, and you don't have the stress of watching your points sailing away with nothing to show for them.

I have often heard people mention that they use Playground or Bing, (or other free AI accounts), to create their source images, then return to DDG and work off of them there.

This is actually quite ingenious, as you start with an image you like and then fine tune it at DDG, saving LOTS of energy points along the way!

I was curious about this, so I took my Sunshine Sisters over to DDG and did just that. Here was the outcome:

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

Playground

A table displays strange items in a nature setting with trees and mountains. Unrecognizable items and beings scattered about.

Deep Dream Generator

I used the DigitalDaVinci option and made 3 versions, at 30%, 40%, and 50%. The DDG image above was the 50% version, which I thought was the best of the three.

I think the DDG version gave me more realistic clouds and hair, but I lost other details I liked from the original. To be honest, I prefer the version I made at Playground, so I didn't bother publishing my DDG version publicly. But one can certainly see the advantages of making source images elsewhere and then bringing them back to DDG to make a 'final version'.

I think it would be awesome if DDG offered this option of multiple images generated at once, like so many other AI websites do, but I doubt very much that they ever will.

They WANT us to burn through all of our energy points so we'll buy more of them or upgrade our accounts to a more eXpensive option, and the site is designed around just that! It IS a business, after all.

But one can't help but wonder how other AI sites, which are also businesses, can offer such a wonderful option for their users, and DDG can't.

Multiple image generations is a tool that, once you've used it a few times, you'll find very difficult to live without!

"X"

Easter Sunday, 2024