Ai, short guide to writing prompts and generating amazing images
[ad_1]
The premise has now become a must and perhaps there is no reason but it is good to do it to avoid criticism. At the state of the art, generative artificial intelligence alone is not able to replace a designer, a graphic designer and an illustrator. In the state of the art, it is worth reiterating, and probably never. They are tools that speed up time when you have an idea and you want to see it implemented right away. or that can help you generate all those images that you don’t have or can’t find in your photographic databases. In short, they can be integrated into creative processes but under the supervision of a designer, graphic designer or human director.
That said, there is much to study and learn. If we consider Midjourney, which is perhaps the most surprising generative Ai tool to date, we need to study the “grammar” of the prompt. It means that descriptions in natural language that are translated into images require multidisciplinary skills and knowledge ranging from photography techniques to art history.
Real technical guides exist online and on Ai sites. For example on Midjourney but also with Dall-E2, Stable Diffusion and so on you can be precise (and happy) if you know what you want.
It starts from the prompt which is then the description in natural language of what you want to be brought to the ground. In Midjourney we start with /imagine. The sooner you describe the subject, the more details the better, then you move on to how you want to make this image (with what tools, with what style and with what level of creativity). With some software you can provide (upload) photographs as a reference to start from. It is good to pay attention that they are owned by us or with copyright-free licenses. It can be done but it is not recommended because the textual description is the most interesting anyway. Finally there are also parameters, it is not all and only qualitative. At the end of the prompt, numbers (parameters) must be entered to determine the image’s aspect ratio and size. In the case of Midjourney the “U” buttons are used to enlarge one of the images, while the “V” buttons allow you to generate variations. Here is a brief guide on the settings.
Once you have described the subject and the elements you are looking for, you need to take care of organizing them. You can decide the type of cameras you want to use, the lenses and the aperture to manage the light. You can choose the shooting position of the photo. You can then be inspired by artistic movements, narrative currents, photographic styles such as minimalism, grunge or manga style. You can also directly choose an author such as Pablo Picasso, Leonardo da Vinci or Gustav Klimpt. You can work on emotions and expressions and let the AI be creative to create a temporal context. If we enter the /settings command in the Midjourney command prompt, we will have access to the image settings. There is also a command called “chaos” where you can decide on a scale from o to 100 how much freedom of interpretation to leave to the machine. And then there’s the cost. Midjourney works with the GPU time system, i.e. the minutes it takes to generate an image. Basic Plan — $12.20/month or $117.12/year (VAT included) with 200 minutes of GPU time per month.
[ad_2]
Source link
