On Wednesday, Midreis announced version 5 of its commercial AI image synthesis service, which can produce photorealistic images at a quality level some AI art fans call scary And “too perfectMidjourney v5 is now available as an alpha test for customers who subscribe to the Midjourney service, which is available through Discord.
“MJ v5 currently feels like I’m finally getting glasses after ignoring low vision for a little too long,” said Julie Wieland, a graphic designer who often shares her Midjourney creations on Twitter. “Suddenly you’re seeing everything in 4K. It feels strangely overwhelming but also amazing.”
Wieland shared some of her Midjourney v5 generations with Ars Technica (in a gallery below and in the main image above), and they certainly show an advancement in image detail since Midjourney first arrived in March 2022. Version 3 debuted in August and version 4 debuted in November. Each iteration added more detail to the generated results, as our experiments show:
Midjourney works similarly to image synthesizers such as Stable Diffusion and DALL-E, in that it generates images from text descriptions called “prompts”, using an AI model trained on millions of works of human-made works of art. Recently, Midjourney was at the center of a copyright controversy related to a comic book that used earlier versions of the service.
After experimenting with v5 for a day, Wieland noticed improvements including “incredibly realistic” skin textures and facial features; more realistic or cinematic lighting; better reflections, glare and shadows; more expressive angles or overviews of a scene, and “eyes that are almost perfect and no longer waver.”
And of course the hands.
Just a heads up: Midjourney’s AI can now execute hands correctly. Be extra critical of any political images (especially photography) you see online that try to provoke a reaction. pic.twitter.com/ebEagrQAQq
— Del Walker (@TheCartelDel) March 16, 2023
Over the past year, the idea that AI art generators can’t render hands correctly has become something of a cultural trope. Midjourney v5 in particular can generate realistic human hands quite well. “Hands are mostly correct, with 5 fingers instead of 7-10 on one hand,” Wieland said.
In the service’s Discord release notes, Midjourney also noted that v5 now responds with a “much wider stylistic range” than version 4, while also being more responsive to prompts, generates less unwanted text, and offers 2x higher image resolution.
If there’s a visual downside to the Midjourney upgrade for AI art fans, it might come from images that can be so realistic and “perfect” that the precision of the model takes some of the excitement out of repeatedly generating AI images to find a suitable result. one might call a “slot machine effect”. Although a Twitter user was named Philipp Lenssen noted“If you have a specific image subject in mind, it’s still a bit like a lottery. But with higher odds than v4.”