We already know that AI art generators are not all created equal. Many of the free AI art generators produce strange blurry results that look like an accidental image shot on one of the first camera phones, while more powerful tools can create convincing photorealistic images and recreate artistic styles. Now a creator using the tools has done a head-to-head comparison of three of the best AI art generators, and the results are fascinating.
The experiment shows very different results from each tool, providing insight into how each interprets the requests. If we didn’t know better, we’d almost think that every AI art generator has its own personality and style like a human artist. What is clear is that each tool has its strengths and weaknesses, and sometimes a distinct look, based on its algorithms and the database of images it’s fed into, and that’s something that artists may be able to take advantage (see our guide on how to use DALL-E 2 if you want some tips).
DALL-E 2 vs Midjourney vs StableDiffusion mega thread: photography, illustration, painters, abstracts, these image synths are like instruments – it’s amazing we’ll get so many of them, each with a unique “sound” 🤯 rules: same request, 1:1 aspect ratio, no living artists pic.twitter.com/47syy7uPJJAugust 20, 2022
based in Berlin Fabian Stelzer (opens in new tab), who describes himself on Twitter as a ‘fast intern’ working on three AI-based projects, conducted the image comparison experiment using Midjourney’s text-to-image AI art generators, DALL-E 2 and Stable Diffusion. He entered the same requirements into each tool and used a 1:1 aspect ratio for the resulting images.
With requests ranging from “low game wealth, Cthulhu monster, 2000 video game, isometric view” to “1990s clip of a crazy laughing fax machine, Windows 3.1, MS-DOS, early art of computer clips,” the results that Stelzer shared on his Twitter link let us compare how the tools handle different types of requests.
image 1 e 6
Midjourney’s creations often feel very dark – almost apocalyptic. After all, this is the tool that was used to create the “last selfie on Earth” images that have been making the rounds recently (see our roundup of the weirdest AI art). We think this AI art generator definitely needs some tweaking, but it also seems to often produce the most natural results when it comes to art styles, especially with textural details. Every artifact looks natural, while in DALL-E 2 artifacts often look like visible digital defects.
DALL-E 2 has a tendency to throw in random made up words, but it seems to be the best tool for creating photorealistic images and handling facial expressions. Meanwhile, Stable Diffusion often seems to produce the cleanest results – Stelzer notes that it can also create incredible photos, but that you need to be careful not to “overpower” the scene. It is also good at recreating the artistic styles of particular artists.
Stelzer says he sees these and other AI art generators like musical instruments, each with its own range and timbre. The Midjourney is like an analogue Moog – beautiful but with a limited range, while the DALL–E 2 has a huge range but a more distinctly digital output “These image synths are like instruments – it’s amazing. We’ll we get so many of them, each with a unique ‘sound,'” says Stelzer, adding, “You want to play’n’prompt DALL-E / Midjourney / StableDiffusion individually for their strengths.”
image 1 e 5
The three tools also have other differences, beyond the initial results, in how they handle requests. DALL-E 2 offers a powerful painting feature that allows you to edit part of an image, while Midjourney has a large and active user community for support and inspiration.
Stelzer believes AI art generators will revolutionize creative work in ways we haven’t seen since the advent of photography—”what photography was to painting, image syntheses are to photographs,” he says, and predicts that soon anyone will be able to create. movie-like content by typing it. He’s using AI to create a 70s community-narrated sci-fi film SALT (opens in new tab) and the experimental game on Twitter @battleprompts (opens in new tab)in which players summon monsters through requests – is a fascinating look at the creative possibilities that can emerge through AI-generated art and AI-generated videos as well.