There has been a lot of moaning and groaning of late from photographers and podcasters on YouTube about how frightened they are of artificial intelligence, how terrible it is, and how it’s going to be the death of photography. I find this quite amusing if not ridiculous. Can you imagine a husband saying to his wife, “Honey, let’s make an AI picture of us standing in front of the Eiffel Tower and now we don’t have to go to France and take our selfies there.” Trust me: Neighbors would hear her yelling at him from at least a half mile away!
AI is not photography. It can be combined with photography and manipulated in Photoshop, but it is different than having a great experience – like standing in front of the Eiffel Tower – and taking pictures. In terms of making images, however, AI surpasses what a camera can do by light years. It is the most creative visual tool in the history of art, and if the artist in you loves making images, this is a tool you should not overlook.
Getting started
The best software to use for creating AI images at this time is Midjourney. You don’t download the software or buy it; you use it online through a subscription plan. You can start for just $10/month, or if you buy an annual subscription, the price drops to $8/month. This entitles you to 3.3 hours per month of fast rendering time for your images. The next highest tier is $24/month, and this equates to 15 hours of fast rendering per month. You can upgrade or downgrade at any time.
Begin by going to the website, discord.com. Here you sign up for a free account, and this is the gateway to not only Midjourney but many other sites as well including hundreds devoted to gaming. To subscribe to a specific plan, go to midjourney.com/account/. MJ used to offer 25 free rendered images, but because of the sheer number of new users, they stopped doing that.
Once you enter the Midjourney site, in the left hand column you’ll see a few ‘newbie’ rooms. This is where you create AI imagery. Go to any of the Newbie rooms to begin creating images.
The Prompts
Artificial intelligence images are based on prompts, i.e., words you use in a description field. When you enter a newbie room in the Midjourney site, at the bottom of the page there is an empty field waiting for your instructions. All prompt sequences must begin with this: /imagine
When you type this in lower case and with no spaces, hit the enter or return key and then you’ll see the word ‘prompt’. At this point, you describe what you want. For example:
A beautiful tall blonde bride in a long white gown with an exquisite train on the steps of a white magical cathedral, her back is to the camera, cinematic lighting, photo realistic.
Or, A young girl with red hair holding a fox kitten whose hair color matches the girl’s hair, standing in a meadow with a forest background, backlighting, photo realistic.
Or, The Egyptian gods Isis and Osiris in a romantic embrace, 3D, cinematic lighting, hyper detail, saturated colors, both gods have metallic skin.
Or, A charging white rhino, kicking up dust, powerful, terrifying action, filling the frame, sunset lighting, photo realistic.
Or, A female cyborg galloping on a chrome horse, she is wearing a cape that is flying in the wind, the horse’s mane is flying, highly reflective, highly mechanical, legs of the horse extended, very detailed.
You can see that you can be as detailed as you want. In fact, there are no limits to the number of words or the number of prompts. Midjourney renders the images very quickly – usually less than 60 seconds – and you are given four variations of each request. You may like all of the variations, none of them, or only one or two. Beneath the set of the four images, you’ll see U1, U2, U3, and U4. The ‘U’ stands for upscale, which means you can enlarge the thumbnail images that MJ has rendered. When you choose this option for one or more of the variations, you can then drag the enlarged, upscaled image to your desktop.
The resolution of the upscale images is only 3.45 megabytes. To address this issue, I process all of my rendered images in Topaz Gigapixel AI and enlarge them 6 times, or 6x. I then run them through Topaz Sharpen AI, and the final output is quite good. You never have to use a noise reducing program because AI images have no noise.
When you post images on Midjourney, all the online users see what you’ve posted and you see what they post. This gives you a lot of ideas regarding 1) what can be created, and 2) the kinds of prompts other people are using.
The sky is the limit with AI imagery. Whatever you can imagine in your mind can be created, and a lot more. Being a photographer helps you be creative with AI, and at some point you will want to import your images into Midjourney to embellish them in ways that will captivate the artist in you.