If we talk about artificial intelligence to create images based on a description, one of them best options available at the moment in Midjourney, one of the first IAs that reached the market and that, with the passing of the months, has grown in giant steps.
A few weeks ago version 5 of Midjourney was launched, a version that God, a jump in quality with respect to the previous version, one of the main novelties being the possibility of interacting with this AI in a totally natural way, as if we were talking with a person and not with a machine, without having to separate them with commas (although you can continue using them) the concepts we want to see reflected in the image we want to obtain.
At that moment, the person responsible for this platform affirmed that they were working on version 6, although for the time being, it was not a priority, but they wanted to optimize the functioning of the current version. However, as expected, the first leaks about what Midjourney has planned to implement in the next version of this IA to generate images have come to light.
New features of Midjourney V6
One of the main problems of all the Artificial Intelligences that allow us to generate images based on a description is the final resolution of this, a resolution that, by default, is set at 512 and 1024 in width and height, a limitation that prevents them from being used to a greater extent without using an AI or applications that allow us to increase the resolution losing the minimum quality possible.
When we interact through Discord to obtain images with Midjourney, the images we obtain are 1024×1024, a resolution that if true is superior to the native Stable Diffusion, it is not sufficient for many uses. Midjourney version 6, will increase the resolution of the images that generally has 2048×2048, double the current resolution that will allow you to get sharper images and much more possibilities of use.
The natural language compression that this platform introduced with version 5 assumed a substantial improvement when interacting with this IA, however, and, although it is not necessary to continue using commands to specify the characteristics of an image, this IA takes them into account at the time of generating the results.
With the release of version 6, Midjourney will be able to interpret from natural language to perfectionincluding even though it is informal, for that, it is time, for the end we will stop using commas to specify specific characteristics, which we want to have the image we are looking for.
Aunque parece no estar en hoja de ruta de Midjourney, muchos son los usuarios que desearían tener la oportunidade de generate models in 3D through a text, a function that will probably arrive in the near future, because the first thing is to improve what it already offers to add new functions when the current ones are completely polished.