13 Steps to Better Prompts for Generating Photo-Like Images Using Mid-Journey AI — Part 02

Kat S.
7 min readDec 30, 2022

--

Image by https://www.instagram.com/katooosz/

This is my base prompt created by following the above steps (included in the part 01):

full body from the front, gloomy and dreamy, [pale-skinned eccentric girl standing in a rain by Petter Hegre], [sad green eyes short black hair and a fringe, wearing cellophane raincoat], background: park, cinematic lighting, Gucci, 50mm lens, muted dark pastels, shallow depth of field

or

full body from the front, gloomy and dreamy, [pale-skinned eccentric girl standing in a rain by Petter Hegre], [sad green eyes short black hair and a fringe, wearing cellophane raincoat], background: park, cinematic lighting, Gucci, 50mm lens, muted dark pastels, shallow depth of field, [clean focus, HD, 16k, accurate, detailed, art photography]

  • This is a little Addition for ensuring quality: [clean focus, HD, 16k, accurate, detailed, art photography]
  • and Mid-Journey settings : — v 4 — ar 2:3 — chaos 90

Note: As you can see, not all images will feature a full body shot. It can be challenging to get a full body image, especially when asking for detailed features such as facial expressions. One trick to help achieve a full body image is to include specific details about the subject’s clothing or appearance, such as the type of shoes they are wearing. This can help the AI incorporate these details into the image.

full body from the front, gloomy and dreamy, [pale-skinned eccentric girl standing in a rain by Petter Hegre], [sad green eyes, short black hair and a fringe, wearing cellophane raincoat and leather loafers], background: park, cinematic lighting, Gucci, 50mm lens, muted dark pastels, shallow depth of field, [clean focus, HD, 16k, accurate, detailed, art photography]

It seems that AI knows rainy weather better than us humans! It appears to prefer wellies over leather loafers. If we want to reinforce a particular feature, we can add it to the front of the prompt enclosed with “!!” at the start and end. Or, we can just keep re-rolling the prompt until the desired result appears. It’s basically like playing the lottery, but with prompts instead of numbers. Good luck, and may the odds be ever in your favor!

When capturing photo-like images I believe it is worth learning about photography specific technical details related to lens, cameras, shooter speed and aperture.

For example, to photograph a whole body and capture details, you will want to use a lens with a wide aperture (low f-number) and a moderate to long focal length.

For the lens, a prime lens with a wide maximum aperture (f/2.8 or wider) would be a good choice. A lens with a moderate to long focal length, such as a 50mm, 85mm, or 135mm, would also be a good choice because it will allow you to capture a narrower field of view and create a shallow depth of field, which can help to isolate the subject and draw attention to their features and details.

Here are some specific examples of lenses that might work well for this type of photograph:

  • Canon EF 50mm f/1.2L USM lens
  • Nikon AF-S Nikkor 85mm f/1.4G lens
  • Sony FE 85mm f/1.4 GM lens
  • Fujifilm XF 90mm f/2 R LM WR lens

OK, let’s try:

full body from the front, gloomy and dreamy, [pale-skinned eccentric girl standing in a rain by Petter Hegre], [sad green eyes, short black hair and a fringe, wearing cellophane raincoat and leather loafers], background: park, cinematic lighting, Gucci, Nikon AF-S Nikkor 85mm f/1.4G lens, muted dark pastels, shallow depth of field, [clean focus, HD, 16k, accurate, detailed, art photography]

I changed the orientation of the image from square to — ar 2:3, in order to to give AI a little bit of help and add some room for legs ;). It’s still not exactly what we asked for, but there’s no denying that the images created are stunning!

Directors:

Movie still from Martin Scorsese movie
Sofia Coppola
Quentin Tarantino
Photographers: Cindy Sherman, Henri Cartier-Bresson, Helmut Newton

Changes above are definetly subtle in nature!

Diffrent outfits, colour pallets and backgrounds are more pronounced:

Clothing: Victorian dress designed by William Morris
colour palletes: neon greens and vibrant pastels, low purity pinks, vibrant reds and greens,
Clothing: School uniform

Background!

Backgrounds: Abandoned factory, green house, New York metro station

Note: the ‘’abandoned’’ factory was influencing the overall look of the subject, so I tested the same prompt adding to the subject ‘’rich’’ here are the results:

Rich girl, abandoned factory
full body from the front, gloomy and dreamy, [pale-skinned eccentric girl standing in a rain by Petter Hegre], [sad green eyes short black hair and a fringe, wearing cellophane raincoat], background: city at night and cars driving, dim city lights, Gucci, 50mm lens, muted dark pastels, long exposure
full body from the front, gloomy and eerie, [pale-skinned eccentric girl standing in a rain by Petter Hegre], [sad green eyes short black hair and a fringe, wearing glossy translucent cellophane raincoat, background: city at night and cars driving, dim city lights, Gucci, [Aperture: f/2.8 , Shutter speed: 1/125 second , ISO: 1600], muted dark pastels, long exposure
full body from the front, gloomy and eerie, [pale-skinned eccentric girl standing in a rain], movie still from Quentin Tarantino movie, [sad green eyes short black hair and a fringe, wearing glossy translucent cellophane raincoat and a lace dress by Alexander McQueen], background: city at night and cars driving, dim city lights , [Aperture: f/2.8 , Shutter speed: 1/125 second , ISO: 1600], muted dark pastels, long exposure

full body, [girl standing in a rain movie still from Quentin Tarantino movie], designed by Alexander McQueen, dim lights, [Aperture: f/2.8 , Shutter speed: 1/125 second , ISO: 1600], long exposure

here I removed most of the descriptors from the prompt, my prediction is that now having ‘Quentin Tarantino’ in prompt will have stronger influence on the image generated! Let’s check!

Like predicted! we are getting very strong main character vibes!

Quentin Tarantino’s films often feature strong, complex, and independent female characters who are not afraid to challenge traditional gender roles and expectations. These characters are often depicted as being tough, confident, and capable, and they often play a central role in the story.

full body, [girl standing in a rain movie still from Quentin Tarantino movie], dim lights, [Aperture: f/2.8 , Shutter speed: 1/125 second , ISO: 1600], long exposure

— I have further removed designed by Alexander McQueen.

girl standing in a rain movie still from Quentin Tarantino movie
girl standing in a rain movie still from Quentin Tarantino movie

I personally love Quentin Tarantino movies, however I feel we are loosing photography quality of the images here!

Thank you for reading! What you would like me to try next? For more visuals visit my https://www.instagram.com/katooosz/ where I share prompts ;)

In addition I started a little pinterest project where i share way more images https://www.pinterest.co.uk/kats00000/

--

--

Kat S.

I am a Senior UX architect with a passion for design, tech, & the future. I love exploring new trends, especially in AI.