DALLE2 vs Midjourney

DALL-E 2 vs Midjourney

I wanted to compare DALL-E 2 vs Midjourney, as Midjourney has recently enabled users to use --test and --testp to upgrade their images and make them much more realistic. Using --test and --testp (test being a more creative improvement and testp leaning towards more photorealistic improvement) and comparing to current DALL-E 2 outputs.

So I chose a variety of prompts to give DALL-E 2 and Midjourney, chose what I considered to be the best from each one and decided on a Winner (and therefore a loser!) from each comparison.

Toronto TTC Streetcars

"a photograph of a TTC streetcar on the street in downtown Toronto, foggy, winter, tall buildings, TTC streetcar, Canada, photography, street photography, city, 8K, realistic, detailed, real world"

DALLE2 "a photograph of a TTC streetcar on the street in downtown Toronto, foggy, winter, tall buildings, TTC streetcar, Canada, photography, street photography, city, 8K, realistic, detailed, real world"
DALL-E 2 streetcar
midjourney "a photograph of a TTC streetcar on the street in downtown Toronto, foggy, winter, tall buildings, TTC streetcar, Canada, photography, street photography, city, 8K, realistic, detailed, real world"
Midjourney Streetcar

The Dalle streetcar image is pretty, but lacks detail. It looks like a painting despite me asking for a photograph in the prompt. The Midjourney image is much more realistic.

Winner = Midjourney

Eyes

"a human eye, macro photography, 8K, realistic, highly detailed"

DAlle2 "a human eye, macro photography, 8K, realistic, highly detailed"
DALL-E 2 eye
midjourney "a human eye, macro photography, 8K, realistic, highly detailed"
Midjourney eye

Dalle's eye is more photo realistic and Midjourney's is more stylised and arty.

Winner = DALL-E 2

Cyberpunk

"a cyberpunk man standing with the skyline of a cyberpunk city in the background, 8K, octane render, unreal engine, volumetric lighting, highly detailed, cinematic, realistic"

DAlle2 "a cyberpunk man standing with the skyline of a cyberpunk city in the background, 8K, octane render, unreal engine, volumetric lighting, highly detailed, cinematic, realistic"
DALL-E 2 cyberpunk
midjourney "a cyberpunk man standing with the skyline of a cyberpunk city in the background, 8K, octane render, unreal engine, volumetric lighting, highly detailed, cinematic, realistic"
Midjourney Cyberpunk

Dalle's image is dreadful here.

Winner = Midjourney

Teddy Bears

"a studio photograph of a cute teddy bear, realistic"

dalle2 "a studio photograph of a cute teddy bear, realistic"
DALL-E 2 teddy bear
Midjourney "a studio photograph of a cute teddy bear, realistic" DALL-E 2 vs Midjourney
Midjourney teddy bear

Both are nice images, but Dalle's is more realisitc.

Winner = DALL-E 2

President Spongebob

"Official portrait photograph of United States President Spongebob Squarepants, sitting in the oval office, smiling at the camera and wearing a suit with blue tie, professional photography"

dalle 2 "Official portrait photograph of United States President Spongebob Squarepants, sitting in the oval office, smiling at the camera and wearing a suit with blue tie, professional photography"
DALL-E 2 President Spongebob
dalle 2 "A journey to the underworld"
Midjourney President Spongebob

Midjourney's image is nice, but completely misses the Spongebob component of the prompt.

Winner = DALL-E 2

Underworld

"A journey to the underworld"

DALL-E 2 underworld
midjourney "A journey to the underworld"
Midjourney underworld

Dalle's image is very literal (a man exploring a cave) while Midjourney's is realistic but still outworldly and intriguing.

Winner = Midjourney

1980s apartment

"stylish apartment from 1980s Miami, furnishings in 80s style, pink and blue colour scheme, magazine photography, 1980s"

dalle 2 "stylish apartment from 1980s Miami, furnishings in 80s style, pink and blue colour scheme, magazine photography, 1980s"
DALL-E 2 80s apartment
midjourney "stylish apartment from 1980s Miami, furnishings in 80s style, pink and blue colour scheme, magazine photography, 1980s"
Midjourney 80s apartment

Both capture the atmosphere but Midjourney's is far more realistic.

Winner = Midjourney

Bibimbap

"food photography of bibimbap"

dalle 2 AI generated "food photography of bibimbap"
DALL-E 2 bibimbap
midjourney "food photography of bibimbap"
Midjourney bibimbap

Both are great but Dalle's image is more accurate in terms of the ingredients in an actual bowl of Bibimbap.

Winner = DALL-E 2

Brutalist

"A man walking through the spacious hall of a cavernous brutalist building, concrete, imposing, brutalist architecture, 8K, realistic, moody"

dalle 2 "A man walking through the spacious hall of a cavernous brutalist building, concrete, imposing, brutalist architecture, 8K, realistic, moody" DALL-E 2 vs Midjourney
DALL-E 2 Brutalist
midjourney "A man walking through the spacious hall of a cavernous brutalist building, concrete, imposing, brutalist architecture, 8K, realistic, moody"
Midjourney brutalist

Midjourney's is more realistic and atmospheric.

Winner = Midjourney

Woman smiling

"a close up portrait photograph of a young woman with red hair and freckles laughing and looking directly into the camera, portrait photography, studio lighting, detailed, 8K"

DAlle2 "a close up portrait photograph of a young woman with red hair and freckles laughing and looking directly into the camera, portrait photography, studio lighting, detailed, 8K"
DALL-E 2 smiling woman
midjourney "a close up portrait photograph of a young woman with red hair and freckles laughing and looking directly into the camera, portrait photography, studio lighting, detailed, 8K"
Midjourney smiling woman

This is a good prompt to show the differences between Midjourney and DALL-E 2. Dalle's image is more realistic, while Midjourney's is too but in different ways. Tough choice but going to go with Dalle purely because if you squint your eyes it looks more realistic.

Winner = DALL-E 2

Velociraptor eating pizza

"a Lomography photograph of a velociraptor eating pizza in New York City"

DALL-E 2 velociraptor eating pizza
"a Lomography photograph of a velociraptor eating pizza in New York City" mid journey
Midjourney velociraptor eating pizza

Dalle captures the lomography aspect of the prompt correctly, while not getting the velociraptor part correct. Instead it's clearly an iguana eating pizza. Midjourney gets it more right but lacks the lomography aspect. It looks more like an illustration than a photo.

Winner = DALL-E 2

Corgi

"a photo realistic photograph of a corgi, 8K"

a photo realistic photograph of a corgi, 8K dalle2
a photo realistic photograph of a corgi, 8K midjourney

In this case both are lovely images, but Midjourney's leans more towards a painting of a corgi than a photo of one.

Winner = DALL-E 2

Blackbird 

"Blackbird SR-71 flying above the clouds, photograph, film grain, 1970s, view above, high quality, realistic"

Blackbird SR-71 flying above the clouds, photograph, film grain, 1970s, view above, high quality, realistic dalle2
DALL-E 2 blackbird
Blackbird SR-71 flying above the clouds, photograph, film grain, 1970s, view above, high quality, realistic Midjourney
Midjourney blackbird

Midjourney is the clear winner here.

Winner = Midjourney

Fynbos

"a photograph of fynbos flowers in South Africa"

fynbos DALL-E 2 vs Midjourney
DALL-E 2 fynbos
fynbos DALL-E 2 vs Midjourney
Midjourney fynbos

"Fynbos", literally 'fine bush' in Afrikaans, is the name for the unique flora of the Western Cape in South Africa. Both DALL-E 2 and Midjourney images are great. But going to give the win to Midjourney as it's slightly more photo realistic.

Winner = Midjourney

Final Results of DALL-E 2 vs Midjourney

So in this epic contest the final results are

  • DALL-E = 7 wins
  • Midjourney = 7 wins

A tie! Who would've suspected?

These tools are both great and I would encourage you to check them out on your own!

Other Blog Posts


Comparing Craiyon, DALL-E 2, and Midjourney

Comparing Craiyon, DALL-E 2, and Midjourney

I thought I'd add photos from three AI image generators; Craiyon, DALL-E 2 and Midjourney.

How they describe themselves;

  • "Craiyon, formerly DALL·E mini, is an AI model that can draw images from any text prompt!"
  • "DALL·E 2 is a new AI system that can create realistic images and art from a description in natural language.
  • "Midjourney is an independent research lab exploring new mediums of thought and expanding the imaginative powers of the human species."

My relatively uninformed opinions of them;

  • Craiyon is good for funny and topical images, but images lack detail.
  • DALL-E 2 can create some almost photo-realistic images, but can lack artistic imagination.
  • Midjourney creates stunning images that lean towards fantasy art rather than the real world.

In this post I just wanted to share some examples of the images produced by these three.

A Cat looking up at a South African flag, digital art

A Cat looking up at a South African flag, digital art by craiyon
Craiyon
A Cat looking up at a South African flag, digital art dalle2
DALL-E 2
A Cat looking up at a South African flag, digital art created by midjourney
Midjourney

A Galaxy inside a Snowglobe, cinematic

A Galaxy inside a Snowglobe, cinematic by craiyon
Craiyon
A Galaxy inside a Snowglobe, cinematic dalle2
DALL-E 2
A Galaxy inside a Snowglobe, cinematic by midjourney
Midjourney

Two Astronauts Exploring a Tomb on Mars

Two Astronauts Exploring a Tomb on Mars craiyon
Craiyon
Two Astronauts Exploring a Tomb on Mars dalle2
DALL-E 2
Two Astronauts Exploring a Tomb on Mars midjourney
Midjourney

A Photograph of the Toronto skyline

A Photograph of the Toronto skyline by ai art craiyon
Craiyon
A Photograph of the Toronto skyline dalle2
DALL-E 2
A Photograph of the Toronto skyline by mid journey
Midjourney

A Hideous Monster staring you in the face

a hideous monster staring you in the face

by craiyon
Craiyon
a hideous monster staring you in the face

by dalle2
DALL-E 2
a hideous monster staring you in the face

by midjourney
Midjourney

A Beautiful, Cinematic Rainbow

A Beautiful, Cinematic Rainbow by craiyon
Craiyon
A Beautiful, Cinematic Rainbow
DALL-E 2
A Beautiful, Cinematic Rainbow from midjourney
Midjourney

A Window into Another World

A Window into Another World by craiyon
Craiyon
A Window into Another World images generated by dalle-2
DALL-E 2
A Window into Another World by midjourney
Midjourney

Two Astronauts explore the interior of a Colossal, Derelict Alien Spacecraft. Cinematic

Two Astronauts explore the interior of a Colossal, Derelict Alien Spacecraft. Cinematic created by craiyon
Craiyon
Two Astronauts explore the interior of a Colossal, Derelict Alien Spacecraft. Cinematic created by DALL-e2
DALL-E 2
Two Astronauts explore the interior of a Colossal, Derelict Alien Spacecraft. Cinematic by mid journey
Midjourney

Star Wars characters depicted in a medieval tapestry

Star Wars characters depicted in a medieval tapestry craiyon
Craiyon
Star Wars characters depicted in a medieval tapestry dalle-2
DALL-E 2
Star Wars characters depicted in a medieval tapestry midjourney
Midjourney

Alien Covenant

Alien Covenant depicted by craiyon ai generated art
Craiyon
Alien Covenant created by dalle2
DALL-E 2
Alien Covenant created by mid journey ai art generator
Midjourney

Cape Town, South Africa

Craiyon
DALL-E 2
Midjourney

The Gates to heaven opening and revealing the view of a Gigantic Wonderous City

ai generated images of The Gates to heaven opening and revealing the view of a Gigantic Wonderous City craiyon dallemini
Craiyon
The Gates to heaven opening and revealing the view of a Gigantic Wonderous City created with dalle2
DALL-E 2
The Gates to heaven opening and revealing the view of a Gigantic Wonderous City from midjourney
Midjourney

Other Blog Posts


multiple dalle2 generated images of a soviet robot falling in love

Multiple DALL-E 2 images created from the same prompt

For those that haven't tried out DALL-E 2, it's a new Artificial Intelligence (AI) system that can create realistic images and art from a description in natural language.

You can take a look at some previous posts on the topic

In those previous blog posts I tended to only show what I considered the best result from each of the prompts. What I wanted to show in this post is just how many unique images DALL-E 2 is able to conjure up from a single descriptive prompt.

In this case the prompt I used was:

A Soviet Propaganda poster of a Robot in Love

Below are the results of this same, exact prompt used again and again to generate new images in a few seconds. Each time you type in the descriptive text and click 'generate' four unique images are created. Each one beautiful and unique in its own way.

A Soviet Propaganda poster of a Robot in Love four examples by dalle2
A Soviet Propaganda poster of a Robot in Love four examples by dalle2
A Soviet Propaganda poster of a Robot in Love four examples by dalle2
four images of soviet union propaganada posters featuring a robot falling in love created by dall-e-2
Four AI generated images of a Soviet Propaganda poster of a Robot in Love four examples by dall-e 2 tool
Four AI generated images of a Soviet Propaganda poster of a Robot in Love four examples by dall-e 2 tool
Four AI generated images of a Soviet Propaganda poster of a Robot in Love four examples by dall-e 2 tool
Four AI generated images of a Soviet Propaganda poster of a Robot in Love four examples by dall-e 2 tool
four images of a robot in love from dalle2
four images of a robot in love from dalle2
four images of a robot in love from dalle2
four images of a robot in love from dalle2
four images of a robot in love from dalle2
four posters showing a robot in love created by dall-e2
four posters showing a robot in love created by dall-e2
images created by artificial intelligence tool dalle2 showing a robot in love
four images created by artificial intelligence tool dalle2 showing a robot in love
4 images created by artificial intelligence tool dalle2 showing a robot in love
four posters showing a robot in love created by dall-e 2

Hope you enjoyed reading about multiple DALL-E 2 images created from the same prompt.

Here are some more blog posts about DALL-E 2:

Full Size Versions of the Robots in Love

Here are four of my favourite images in full resolution.

a soviet propaganda poster of a robot in love made by dalle2
a soviet propaganda poster of a robot in love made by dalle2
a soviet propaganda poster of a robot in love made by dalle2
a soviet propaganda poster of a robot in love made by dalle2

Other Blog Posts


comparing different photo styles in dalle2

Comparing different photo styles in DALL-E 2

Wanted to show how DALL-E 2 will create different images based on the same subject matter, depending on what style you add to the prompt.

To recap, DALL-E 2 is a new online tool that allows users to generate images from text prompts in a few seconds. They can generate images of the same subject in many different photo styles, depending on what instructions are included in the prompt.

Below are four photos of each subject, each in a different style. Below each image in the caption is the text prompt that was used to create them.

A Samurai riding a horse on Mars

A Samurai riding a horse on Mars, Lomography created in DALLE 2
A Samurai riding a horse on Mars, Lomography
 A Samurai riding a horse on Mars, Wide Angle Lens dalle2
A Samurai riding a horse on Mars, Wide Angle Lens
 A Samurai riding a horse on Mars, Lego Bricks
A Samurai riding a horse on Mars, Lego Bricks
A Samurai riding a horse on Mars, Anime
A Samurai riding a horse on Mars, Anime

A photograph of a Cyborg exploring Tokyo at night

A photograph of a Cyborg exploring Tokyo at night, lomography created using DALLE2
A photograph of a Cyborg exploring Tokyo at night, lomography
A photograph of a Cyborg exploring Tokyo at night, 50mm
A photograph of a Cyborg exploring Tokyo at night, 50mm
A photograph of a Cyborg exploring Tokyo at night, wide angle lens dalle2
A photograph of a Cyborg exploring Tokyo at night, Wide Angle Lens
A photograph of a Cyborg exploring Tokyo at night, digital art
A photograph of a Cyborg exploring Tokyo at night, Digital Art

A Rhino Surfing in the Sea

A Rhino surfing in the sea, digital art
A Rhino surfing in the sea, digital art
A Rhino surfing in the sea, clay mation animation
A Rhino surfing in the sea, claymation
A Rhino surfing in the sea, lomography
A Rhino surfing in the sea, lomography
A Rhino surfing in the sea, award winning 4K photography
A Rhino surfing in the sea, award winning 4K photography

A Raccoon eating a smoked meat sandwich

A raccoon eating a smoked meat sandwich, UHD 8k
A raccoon eating a smoked meat sandwich, UHD 8k
A raccoon eating a smoked meat sandwich, claymation ai
A raccoon eating a smoked meat sandwich, claymation
A raccoon eating a smoked meat sandwich, made out of Lego bricks dalle2
A raccoon eating a smoked meat sandwich, made out of Lego bricks
A raccoon eating a smoked meat sandwich, isometric art Ai art
A raccoon eating a smoked meat sandwich, isometric art

A Tyrannosaurus chasing a TTC streetcar down the street

A Tyrannosaurus chasing a TTC streetcar down the street, lomography dale2
A Tyrannosaurus chasing a TTC streetcar down the street, lomography
A Tyrannosaurus chasing a TTC streetcar down the street, claymation generated by artificial intelligence
A Tyrannosaurus chasing a TTC streetcar down the street, claymation
A Tyrannosaurus chasing a TTC streetcar down the street, wide angle lens dalle2
A Tyrannosaurus chasing a TTC streetcar down the street, wide angle lens
A Tyrannosaurus chasing a TTC streetcar down the street, digital art dalle2
A Tyrannosaurus chasing a TTC streetcar down the street, digital art

Other Blog Posts


More Examples of AI-Generated DALL-E 2 Images

I wanted to share some more images I created using DALL-E 2. DALLE-2 is a new Artificial intelligence (AI) system that can create realistic images and art from a description in natural language. I tested out DALL-E 2 in a previous post, and since that time it has become available in Beta.

The following images were created in a few seconds using only text prompts.

velociraptors stalking through a corporate office, lomography dalle2 dalle 2

Velociraptors stalking through a corporate office, lomography

Two astronauts exploring the dark, cavernous interior of a huge derelict spacecraft, digital art dalle 2 AI

Two astronauts exploring the dark, cavernous interior of a huge derelict spacecraft, digital art

A still of Homer Simpson in the Blair Witch Project dalle 2

A still of Homer Simpson in the Blair Witch Project

A samurai riding a horse on mars, lomography dalle2 dalle 2

A Samurai riding a Horse on Mars, lomography.

A modern, sleek Cadillac drives along the Gardiner expressway with downtown Toronto in the background, with a lens flare, 50mm photography made with dalle2

A modern, sleek Cadillac drives along the Gardiner expressway with downtown Toronto in the background, with a lens flare, 50mm photography

A realistic photograph of a young woman with blue eyes and blonde hair made using dall-e 2

A realistic photograph of a young woman with blue eyes and blonde hair

A man standing in front of a stargate to another dimension DALLE-2

A man standing in front of a stargate to another dimension

Spongebob Squarepants in the Blair Witch Project photograph dalle2

Spongebob Squarepants in the Blair Witch Project

a velociraptor working at a hotdog stand, lomography dalle2

A velociraptor working at a hotdog stand, lomography

a man walking through the bustling streets of Kowloon at night, lit by many bright neon shop signs, 50mm lens

A man walking through the bustling streets of Kowloon at night, lit by many bright neon shop signs, 50mm lens

A BBQ that is alive, in the style of a pixar animated movie

A BBQ that is alive, in the style of a Pixar animated movie

A futuristic cyborg dance club, neon lights

A futuristic cyborg dance club, neon lights

The long-lost Star Wars 1990 Japanese Anime created by dalle 2

The long-lost Star Wars 1990 Japanese Anime

A hamburger in the shape of a Rubik’s cube, professional food photography made with dalle 2

A hamburger in the shape of a Rubik’s cube, professional food photography

A Synthwave Hedgehog, Blade Runner Cyberpunk created using artificial intelligence design tools

A Synthwave Hedgehog, Blade Runner Cyberpunk

An astronaut encountering an alien life form on a distant planet, photography dall-e2

An astronaut encountering an alien life form on a distant planet, photography

a dinosaur exploring cape town, photography

A Dinosaur exploring Cape Town, photography

a man falling in love with his computer, digital art dalle2

A Man falling in Love with his Computer, digital art

A photograph of a cyborg exploring Tokyo at night, lomography dalle2

A photograph of a cyborg exploring Tokyo at night, lomography

dracula walking down the street of new york city in the 1920s, black and white photography dalle 2

Dracula walking down the street of New York City in the 1920s, black and white photography

DALL·E synthwave airplane

Synthwave aeroplane

A man wanders through the rainy streets of Tokyo, with bright neon signs, 50mm created with dalle-2

A man wanders through the rainy streets of Tokyo, with bright neon signs, 50mm

a space shuttle flying above cape town, digital art dalle 2 ai generated image

A Space Shuttle flying above Cape Town, digital art


Testing out DALL-E 2

I was recently granted access to DALL-E 2, the impressive new AI system that can generate images from text prompts.

I don't have anything technical to say about it, I just wanted to share some of the incredible images it has produced. If you'd like to know more about it here are some links;

Anytime you enter a text prompt and click generate you are presented with six versions of your request, see below for a close up of a pizza.

Here is another example of the six different images Dalle-2 generates for each request you submit.

You can specify the style of the image, or it will make an assumption about which style might be more appropriate. For the hippos above, I did not specify that they should be a cartoon, but Dall-e made assumed that a cartoon style would be the best approach. Similarly for a man walking in the street in Khayelitsha, South Africa it chose a very photo realistic approach.

I've chosen some of the nicest images generated so far to include below. Enjoy!

Images I generated from text prompts using Dall-E 2

A watercolor painting of Toronto generated by Dall-E 2
A watercolor painting of Toronto
The space Shuttle floating in the sky above Cape Town, South Africa, polaroid generated by Dall-E 2
The space Shuttle floating in the sky above Cape Town, South Africa, polaroid
An aircraft carrier on the lake, in front of the Toronto skyline, digital art generated by Dalle-2
An aircraft carrier on the lake, in front of the Toronto skyline, digital art
A raccoon fortune teller, digital art generated by Dall-e-2
A raccoon fortune teller, digital art
An astronaut walking on mars, watercolour painting generated by dalle 2.0
An astronaut walking on Mars, watercolour painting
An alien spaceship hovering over Table Mountain digital art generated by Dalle 2
An alien spaceship hovering over Table Mountain digital art
raccoon with the toronto skyline in background disney animation generated by Dalle2
A raccoon with the Toronto skyline in background, Disney animation
A Matisse painting of a blue jay and a cardinal
A spaceship hovering above Toronto skyline, digital art
A possum, a skunk, a raccoon and a blue jay having a picnic together, album cover art
The statue of Liberty fighting with Godzilla with the New York skyline in the background

See more examples of images creating using DALL-E 2