Chapters:
0:00 Intro
0:32 Finding an Image
2:04 Building a Prompt
4:10 Refining the Prompt
6:14 Remastering
7:07 Exporting
Follow along my workflow for creating fan art of superheroes in midjourney.
We go through some useful tools along the way
Here’s to Queen Elizabeth the 2nd
My prompt
Close-up of Queen Elizabeth II as a superhero, wonder woman, portrait, red armor clothes, intricate details, epic scene, Photography, Cinematic Lighting, Volumetric Lighting, ethereal light, extremely detailed, volumetric rays, a poster by Everett Warner, cgsociety, feminist art, dc comics, –ar 2:3 –testp –iw 0.85
Image to text tool
https://replicate.com/methexis-inc/img2prompt
Upscaler
letsenhance.io
If you found this useful, check out my courses on sale now!
How to make and sell ai art
https://www.udemy.com/course/make-and-sell-ai-art/?referralCode=C67A8D247B2578D4E762
Midjourney Mastery
https://www.udemy.com/course/midjourney-mastery/?referralCode=92BFBB305B81A1C7D1A0
#aiart #midjourney
What's the name of the site where you upscale the size of the image?
Hi Samson, I work with Midjourney 8 days and sometimes I get quite deformed results even the prompts are very clear and accurate, even when using image to text first. Also the eyes are an issue while using –testp. Do you have some tips?
P.S. I subscribed to your Udemy course. 😃
I started making a render 3 hours before the announcement, where can I show you? Love your content 👌🏻
You do always interesting videos with a good sens of humour. Thanks for all that 👍
Great video! This is an element that people tend to forget when they talk about A.I art, the amount of decision making and processing to get a final workable image. I've read on promptcraft people spending many many hours of their life failing to get what they are after. Those looking at this new tech and getting into it only see a 'one click masterpiece' generator just like people thought you pressed a button in a DAW and an electronic musical symphony came out of the speakers.
Skills are skills and with any new medium those with talent will utilise this as a tool rather than a be all and end all.
She looks very similar to the queen of the United Kingdom? It's it the same person?
How to get full body ? What is the best prompt?
Does MJ's upscaler work on external images (whether AI or man-made)? Or is it only available as an option on the command line, while you're generating a new image with MJ?
Can I use my photos to make these arts?
I'm early into watching your video and notice you pick "copy prompt". I would suggest you use "copy command" instead. It includes the prompts but also the other settings like ar, image weights, or if they used v3/test/testp or a specific seed. You'll need those if you are trying to go for a specific look that you found.
Hey! Thanks for this video about prompt crafting.
Great course on selling AI Art! Do you provide private mentorship, I am serious and want to get positive results!
Please, let me know
You go way too fast and miss out too many steps for this to be any kind of useful tutorial.
I bought your course in udemy. I like it good for begginers like me. will there be upcoming updates in the future?.
Thanks Samson for this video – super helpful! I have a question. If I land on an image that I really like in terms of style/design, how can I use that image to generate other images using different prompts with the same look and feel? I've tried the image prompt, but it seems to not quite work how I want. Also, what's the best way to add text to an image that's been created (link image prompt, use the original text prompt and then add more text?). Thanks
Your are one of the most helpful and informative AI art content creators on youtube. Thankyou and looking forward to more!
Amazing!
Great tutorial. You are one of the best on this topic.
As an american from México it is fascinating to see how art of the queen varies according to the level of knowledge you have about her or the country you live in. Here some people are doing art about how she helped pillage and slave thousands creating so much suffering for profit. In other countries they see her as a heroin or things like that regardless of her past action that brought so much misery to many generations.
So is midjourney allowing others to sell their works as well as Dalle-2? I only know of Dalle-2 allowing commercial use.
Thanks for the sharing of the content I would rather not use the queen. Unless I am doing horror art too much history of evil from that family to celebrate.
Helpful Thanks
Thank you sir!
absolutely cool
I'm getting awesome results on staggeringly beautiful women, perfect features on all of them.
My best advice I can give anyone is to let MidJourney know when it is has done good, give it a like if you get a result and then it responds to your prompts better, and keep rerolling until you get what you want.
Bro can i make one word , one image in single frame kindly reply
From across the pond, I raise a cuppa tea to you, and the Queen departed.
Would there be a difference if it said Armour instead of Armor?
Does rerolling cost you "goes" because as a free user I thought you are limited.
An AI that turns images into text that we can use another AI to turn back into images. The future is weird.
How can we do this in stable diffusion…?
First of all the midjourney stuff was interesting. Second of all, good to hear you do not support the monarchy 7:59 and you put that summary of your view quite eloquently. But 0:29 carried a nation??? The nation carried her mate. She pays no taxes and their family has a 40 billion empire of land . If I never had to pay taxes my kids would be millionaires when I died. Their children would have tens of millions and their great grand kids would be coming up on 50 million quite easily. #taxthemonarchy
Amazing what you can achieve with a bit of practice. My first attempts were dreadful.
I can't seem to get anyone who isn't already famous to work, does it also take into account famous people? For instance, mentioning that the subject is Queen Elisabeth II will tell the algorithm to use its existing knowledge of the queen, however if I have a friend's photograph I'm not getting anything that looks like them as if my reference photo wasn't even used at all.
Love your guides, thanks so much for the Image to text tool and the Letsenhance link, really appreciate it. Big salute from Canada 😀
I don’t get how added image has helped you, since you used “Queen Elisabeth ll” in prompt – so it’s not gonna work the same with a random woman portrait
Great tutorial, well explained, of a lady who carried a nation (through genocide, stealing land, pillaging and burning countries all around the world)
Great as always, but i think you missed explaining why at 4:33 you were using a different image prompt, no doubt from an earlier version. Id be interested to know explanation for this step.
playing around with it and never knew certain tNice tutorialngs. I thought you could only use it with a midi keyboard.
1:23 in you dragged a screenshot of Wonder Woman into “this” What is “this”? Meaning what platform did you use to generate the useful key words?
I am an artist painter currently transitioning to digital and Ai.
Thanks for sharing
WTF is he talking about with the brackets? @3:40
Found this tangent really annoying
Art has become the realm of coders.
Check out my ai art courses
Midjourney Mastery
https://www.udemy.com/course/midjourney-mastery/?referralCode=92BFBB305B81A1C7D1A0
DALL•E Mastery
https://www.udemy.com/course/dalle-mastery/?referralCode=8596FB368155A88041F3
Create and Sell AI Art
https://www.udemy.com/course/make-and-sell-ai-art/?referralCode=C67A8D247B2578D4E762
Great video! What did you use to get the key words from the WW poster?
perhaps the bloodiest hands in history. Thanks for the midjourney tips… but good riddance to her. bleh.
Great Vid, I've never used letsenhance before, I use Photoshop to resize, I think the quality should be better on PS unless theres something I'm missing about letsenhance
How often do you hear Pauly Shore?
What does the smiley system do when revising? Does it weigh further revisions of the prompts?
Also, I've noticed that sometimes rerolling makes humans look more realistic, and sometimes variations make a big difference- what's the most effective form of improving a result you like?
Cheers for the vid mate!
20 years later .. the matrix is comming.
ty daddy
Amazing artwork. Thank you for sharing your tips and experience!
bro so this is an add!?
had to leave your video because i couldnt her anything you were saying. try speaking louder or getting a new microphone. thumbs down.
This is great. You made Elizabeth a superhero by using Princess Dianna (of Themischyra.)
Thanks! Btw whats the name of the bg track pkaying? 🙂
What type of test document is that ? can you share the details please? 1:41
hey 🙂 i just wanted to let you know I absolutely love your content, and particularly this video and your superb rendition of The Queen 👑 i discovered it one day when it popped on my feed… since then i have watched it [and rewatched it] many times haha lol… love it so much, i actually featured it in one of my videos. With the proper source quote and link to it, of course hope that's okay with you, thanks again for such a great video! take care
How did you do the screen shot part to drag it into the image to text page? I watched it 10 times and can't figure out how you did it. Thanks for any help with this.
Does it not work to use images you put in as image prompts together?
With stable diffusion / dreambooth you can train your AI model with 10 – 30 pictures of someone (yourself) and then you give it a name so the prompt recognises this and uses your face when generating. Is this possible in Midjourney? Can't find it…. Thanks!!
Hi, can i ask u about your course? ive had a few hit n misses in the past?.
Im a really good talented artist, very strong detail etc. Im 50 (ok boomer!) now and not gel-ing with this new format, getting really ok results, think i need a bit of a kick start.
Any good pointers to why i should spend money on your course?
Cheers 🙂
very nice! can feature how to use your own photos and process them locally on windows? I've only tried stability diffusion locally…what option do we have for this, any tips pls?
what program do you use for screenshot? Thank you for sharing your tips and experience.
U don’t say the name of the first websight
This is good but can we do our own pictures in same manner ?
Holy crap shut up