fourlens.ca

I continue to work on my ability to get Midjourney to render photo realistic images and am reading a daily digest of AI articles.  This morning was an article by a fellow known as "William" who recommended a series of interesting prompts, including a fisherman on a foggy day.  Thanks for the idea, William.

This month's photo assignment from the Grimsby Photo Group is called "Seeing Red".  With that in mind, I grabbed my iPhone 14 Pro Max and ventured off to a nearby car show.  Here are a couple of red vehicles that I found there.  

On my bucket list is a trip someday to the hot air balloon festival in Albuquerque, New Mexico.  I asked the Midjourney AI engine to render me this image of what it could look like.

  I thought I would do some travelling in today's blog post.  Here is an AI-generated view of the blue church domes in Santorini, Greece.  It is such a beautiful place!

Sitting around the firepit two evenings ago, good friend Jim was telling us a story about coming across a black bear drunk from eating fermented apples in an orchard.  While at the firepit I showed Jim how Midjourney AI may be able to produce an image of the scene.  This is what it rendered.

Today's post is a step back from the AI-generated images with a simple photograph I took on Sunday with my iPhone.  The leading lines take your eyes from the cart path around the curve and into the photo.  No AI in this one! 

Overall I like this image generated by Midjourney AI, especially the action and the lighting.  However, I'm not happy with the horse's apparently stubby legs.  It might be worth another try.

My Arizona neighbour and good friend, Jim, has a beautiful yellow 1971 Jaguar in his garage (likely the only one in the world whose engine was swapped out for a Shelby Cobra!).  I have been trying to coax the Midjourney AI engine to produce a photo-realistic image of Jim's car.  So far, I am not happy with what I have been able to coax out of Midjourney.  The images are too cartoonish, not very photo realistic.  Clearly, I need to improve my skills working with Midjourney.  

I was curious to see what Midjourney AI would render as an image of the city I was born in the year I was born.  The result is interesting but needed some aging so I put the Midjourney image through Distressed FX to give it an older look.  Here are the two images.  

I generated this tall image using Midjourney AI as a reminder of a safari in Africa many years ago.  I took so many giraffe photos with my old Minolta SRT 201 that the people I was on safari with started making fun of me.  "Look, over there.  There's a giraffe you havent photographed yet!"  Well, now with the help of Midjourney, I can generate all the giraffe images that I want!

As I have mentioned in a prior post in this blog, I love the ability to use the Midjourney AI engine to move my thoughts to a different point in time and space.  My wife would like to plan a trip to Scotland and she wants to see some castles, so I asked Midjourney to show me the inside of a Scottish castle sometime in the past.  The image is rather dark, but I guess this castle depends on outside light, candles and fireplaces.  

The Midjourney AI engine does an incredible job creating images of people.  That said it sometimes struggles with some features like faces or hands.  In this cowboy shot, I had to work with Midjourney to coax it into a more realistic left hand.  Initial images were not very good.  Even in this one the hand is somewhat longer than it should and the index finger seems a bit long.  Other than that it is a neat shot of a cowboy by an open fire.

This touching image, made by Midjourney AI, has been well-received on my Instagram.

I have been playing with Midjourney AI for weeks now and it has been a blast asking for images.  But my skills using AI need much more work.  I was reading this week about asking Midjourney's ability to  render double exposures and stained glass.  Wow!  

The trees have been a bit slower to turn colours this year.  So I asked Midjourney AI to render me a red maple leaf, and here it is.  By coincidence I was watching Toronto Maple Leafs home opener against the rival Montreal Canadiens while I was playing with Midjourney.  The Leafs won the game in a shootout.  (These Leafs don't usually fall until the first round of the NHL playoffs.)

In keeping with the minimalist images I posted yesterday, I decided to have a go with my iPhone.  This one had three challenges: 1. cutting the construction paper at the right angle to give a 3D illusion, 2. getting the egg to stay standing long enough to take the photo, 3. positioning a light source to get the shadow in the right spot.  Photo was taken in Apple RAW with iPhone 14 Pro Max and edited slightly in Luminar NEO to remove blemishes.

I've always been a fan of minimalist photos, but have never been particularly good at finding the right opportunities.  I thought I would see what Midjourney AI would generate if I asked for a show of some minimalist images.  These are two that I love.

My daughter, Laura has a successful photography business (www.ellehudsonphotography.ca) and she does amazing work with female models.  I wanted to show her the kind of images that Midjourney AI can generate and so asked her what she would like to see.  Laura asked for a fashion model in a red, flowing dress on a roof top with a skylight.  Both of us were impressed with the images rendered and the lighting on this one, in particular. 

Looking for something a little more colourful today, here is a close-up image of a ladybug, as rendered by the Midjourney AI engine.

I was interested to see what Midjourney AI would produce with my request for a silhouette and a reflection.  It is an interesting result.  I like that Midjourney added the human form and the setting sun in this image.

Here's a lighter post today from Midjourney AI.  This handsome pair of ducks features the fellow on the right in a plaid shirt and overalls.  Love the look!

I mentioned in yesterday's blog entry that it is sometimes fun to give the Midjourney AI tool very little to work on and see what it renders.  Though I had to do some more definition to get to this image, my original prompt had one word, "checkmate".  I was quite surprised to see a woman in medieval clothing by a chessboard in a dark room lit by candles!  I certainly didn't ask for all of this in my one word request, but did find it quite fascinating. 

Sometimes the most interesting AI images are ones where you provide little detail in the prompt.  This request to Midjourney was for 'acoustic guitar' and some of the results generated were quite lovely.  Here is an example.

There was a question at my last iPhone special interest group as to whether images generated by Midjourney AI could be edited like photos from our iPhones.  The answer is yes.  Here is a snowy owl created by Midjourney and I have added a neon glow using one of the new AI capabilities in Luminar Neo.

Sometimes I like to give Midjourney a single word or small phrase prompt to see what it will render.  In this case I asked Midjourney to produce me an image that showed "Joy" and this is what it came up with.  Note that I need no 'model release' to use this 'photo' as the person does not exist.  Also, if anyone (including me) tries the same prompt "Joy", Midjourney will produce a different image.  

Yesterday's blog entry showed how images generated by Midjourney can take you to a different place and time, most interesting if it reflects an important point in your life.  Here are two more images that I asked Midjourney to create for me.  They remind me of when I sat atop Mount Kilimanjaro watching the sunrise over the Serengeti and an Acacia tree like the ones I saw while on photo safari in East Africa.  Thanks for the memories, Midjourney!