MidJourney does realism quite well. Let’s see if we can make a realistic portrait of a Japanese woman. In the first prompt, I am going to add the words beautiful, symmetrical face, and young. This will give the best results. It’s also best to use the word photo realistic. Let’s start with a prompt.

Prompt: Beautiful young Japanese woman, symmetrical face, photo-realistic

This is a great start, however, the word young has different meanings and some of these results are too young. I was thinking of a woman in her 20s. No joke, I just filled out a government form that described the word young as under 39. We need to be more specific, also lets at the 4:5 aspect ratio to the prompt. This will give a longer image.

Prompt: Beautiful Japanese woman in her 20s, symmetrical face, photo-realistic –ar 4:5

This was not what I was going for. You can see that MidJourney actually made it symmetrical with a line. Let’s remove that from the prompt. Let’s also add some more detail. We could add the subject in the country with a mountain behind or we can make it an urban setting. Let’s make it an urban setting.

Prompt: Beautiful Japanese woman in her 20s, in Tokyo, photo-realistic –ar 4:5

This is pretty good. However, let’s take a look at what MidJourney is doing automatically. First, it’s putting the woman in traditional clothing. Let’s change that to business clothing. Second, it’s adding a depth of field to the shot. I was going to add in this time, but it seems to be doing it automatically. If you look at the previous images, depth of field was not added in some variations. Let’s also add 32k to the prompt to get a more detailed result. Let’s also re-roll the result so we get 8 variations to look at.

Prompt: Beautiful Japanese woman in her 20s, business clothing, in Tokyo, depth of field, photo-realistic, 32k –ar 4:5

I think version 4 of the first batch is the best. However, version 2 is also good. Let’s make 4 variations and see the result.

I like version 2 the best. Let’s upscale it and see what happens.

I think this result is quite good. A lot of the material on this blog has not been photo-realism or urbanism so I wanted to change that.