|
Listen to this story
|
Midjourney today announced the release of version 5 of its commercial product. Midjourney v5 is said to be capable of producing photorealistic images at a quality level that some AI art enthusiasts have described as “creepy” and “too perfect”. The new version is available as an alpha test for Midjourney subscribers via Discord.
The developers of Midjourney v5 have made significant improvements in various areas, including aesthetics, quality, and details. The latest version boasts a wider stylistic range, understands user requests more accurately, and can produce higher-quality images instantly—thanks to twice as much resolution and a wider dynamic range.
The new version also excels in small details, such as the number of fingers on each hand and facial detailing. Developers have claimed to have made significant improvements in facial detailing, making it easier to create well-known figures from popular culture.
As per a Twitter user, “Midjourney’s AI can now do hands correctly. Be extra critical of any political imagery (especially photography) you see online that is trying to incite a reaction”.
Midjourney v5’s neural network uses default styles to create photorealistic output when no style is specified. This feature has grown in significance in recent years, with the mention of artists’ or movements’ names becoming increasingly important.
The new version also supports seamless textures, wider aspect ratios, and better image prompting, making it easier to produce a more diverse range of outputs. Midjourney hopes to encourage its community to test Midjourney v5 and explore the new features it has to offer.
The v5 was launched earlier for select users which, according to Midjourney, wasn’t performing well as the images were coming out blurry. However, Midjourney claims that after disabling it, the issues have been now fixed.
