Chapters:
0:00 An idiot talks about their last video
2:03 List of artists covered in the video
2:30 The experiment prompt
3:00 Jack Kirby
3:21 Osamu Tezuka
3:30 R. Crumb
3:36 Charles Addams
3:47 Mike Mignola
3:54 Bill Sienkiewicz
4:02 Charles Schulz
4:10 Alex Ross
4:18 Shel Silverstein
4:26 Frank Miller
4:34 Charles Burns
4:42 Matt Groening
4:56 Dr. Seuss
5:04 Bill Watterson
5:12 Steve Ditko
5:20 Chuck Jones
5:28 Gary Larson
5:37 Will Eisner
5:48 George Herriman
5:56 Jamie Hewlett
6:05 LIST OF TOP 5 ARTIST COUNTDOWN
9:23 An idiot recaps and tells jokes
Midjourney is becoming hugely popular and my last video was met with quite mixed reactions. I thought it was an opportunity to explore some of my favorite comic and illustration artists and what Midjourney would create based on using the same prompt across 25 different artists.
For each of these artists I show an example of their actual work then a couple frames of what Midjourney decided their work looked like using the same prompt for everyone. It’s interesting, strange, and pretty cool to see who Midjourney can replicate and who may not have been included in their data set. Check out the artists below.
Just to be clear. I’m not advocating that everyone drop their pencils and move to AI only art. It’s a tool in the artist’s arsenal just like photoshop or vodka. Use it when appropriate but some discretion is advised….
Check out my first video on making an entire comic book in Midjourney and Inferkit here:
https://youtu.be/tjj6KsPSHZc
Awesome bro! thanks for your effort! yes, Mike Mignola deserved it!! his style worked for me! MJ caught the right style of him!!
Cool, do more like this?. Very interesting results
Good way to approach it. So how can we get an AI storyline?
More great information here, thanks for sharing your efforts! I've had interesting results using Walt Kelly as a prompt, it mostly works for cute animals, not sure if it's usable to make a comic but it usually has that old newsprint look. Any tips or ideas on how to get it to consistently spit out the same character in different situations? I think it will sort of work if you use a celebrity as the main character but that's not always desirable.
Yup, there's definitely no putting the genie back in the bottle, or putting the demons back in Pandora's box, however you wanna slice it. Looking through the community feeds on the MJ site, I was gobsmacked by the sheer breadth of creativity that's being showcased and enhanced by these new tools. The weekender types will soon tire of Christina Hendricks in tight dresses and Winona Ryder as Trinity in The Matrix but the true creators will keep finding more and more interesting ways of using the toolsets to produce amazing work that wouldn't likely have seen the light of day outside of this fusion between human and computer minds.
Dude… Thanks for this video, this will set me on the right path with midjourney again.
Ashley Wood is a fun one to play with.
I do love the perspective you bring to working with midjourney. Thanks so much for taking the time to help us see things from a different angle than many of the other enthusiasts who have come from photography and photoshop strengths.
Any tips for keeping characters (etc) consistent throughout a storyline?
As thing goes, if the comic book community doesn't accept it, they will be left behind. I'm sure a new community will emerge with many, many creators that are now able to make comics for their stories and they will gather their own followers. "The genie is out of the bottle". "Fool me once, shame on you. Fool me twice, shame on me" is the saying 🙂
I wonder how reddit would have reacted if you didn't tell them it was AI
Anyway to prompt it create more superhero poses?
Is there a place where we can have share all this comic book, in a common place? I mean, feels like everyone needs to have at least made one comic book in there lifetime having access to this techromancy.
Good work bro.
Isn’t this all based off of scraping other artist work without their permission or being compensated for it?
The tech is amazing but ethically I can’t support it.
The short term gains are great but this doesn’t encourage future artist to make new original work that will just be eventually scraped but someone making ai prompt art for a but or clout. It takes years of hard work and sacrifice to develop these skills.
I dunno.
It just doesn’t feel right.
Finding your videos really interesting. I would love it though if you could do a deep dive into comic life. A slow tutorial on how to use it would be great.
i like that you didn't apologize even though you love and respect this art. when i was in college our computers had 4 colors s when i took art classes i took them with what we called "advertising art" students. they were very good. everything was done by hand and what was most impressive was how clean their projects were.
barely ten years later and adobe has taken over.
design is made so much easier to do. it's going to happen. the issue that will be interesting is copyright. right now US law says all this stuff is in the public domain. if midjourney spits out something no one owns it despite what their tos says. it has already been ruled on.
This was an awesome video. by accident I combined two different set styles: Hayao Miyazaki and HR Giger and the results weren't bad.
Indeed Ralph and Moebius produce excellent results.
Will give either (or both) a go, now that I have some time on my hand. 😉
I think with midjourney if you are going for a style it is due to the amount of content available to and how smart the AI is at this point. So the issue now is scrapping the internet and finding the best images that will give the AI what you need at a starting point.
I have been intrested in this AI picture making things for awhile. I am an artist who has been been making drawings, paintings for a very long time. The process you discribe from this video and the other comic making video reminds me alot of collage work. The process has more to do with hunting down images as you would go through countless old magazines to find useable images which you then would splice, paste up and the like to create the work. The computer kinda does alot of that type of leg work.
It's nice to see this topic covered from more of an exploritory artist perspective. I hope to see more. 😀
AI still sucks big time with outlines, that is clear. The best results are from painting, photography or renders, anything that doesn't rely on the outlines but mostly on color shapes. It picks well brush strokes, but not well defined lines. Well, thank god.
A tip, for free (so value it accordingly) if you change your prompt to end ", made by XXXX" it seems to do a much better job of understanding that you are looking for an art style, as opposed to including the artist in the image or redoing some of that artists work in another style.
I've been using prompts like yours to get consistent characters between panels. So a prompt may be "Cristiano Ronaldo walking a dog in a city, made by Jack Kirby".
What about Bilal?
Hello! I got your videos suggested and I love it. It was exactly what I was looking for. What is your experience with creating a comic book with the same character across multiple panels and pages? Do you think it could be done?
Awesome! Please keep making this type of videos, this is quality informations 👍 thanks a lot for these, you really motivated me to make a graphic novel with A.I, this topic is gold, keep on it ✌️
These are definitely five artists whose work you can use Midjourney to manipulate and then reformat into a comic book. Then there's the ethics of doing this, which you don't discuss or even seem to consider. All five artists are no longer living so they can't complain. But not all of the work is in the public domain. You're not selling anything as far as I know. Of course as soon as you even attempt to sell anything you've made this way, you could be guilty of about a million copyright violations. There's been a number of instances where Midjourney even copies some mutilated form of the artist's signature. That's a very important part of this exercise. You should probably mention that if someone's going to do this, they shouldn't try to sell it. This is fan art, until it is not. When it no longer becomes fan art is when you're making a profit from it. Then it becomes something potentially illegal.
So ethical question #2: what if these artists were living and found themselves on this list? I could tell you what I would think. If I found myself on this list, I'd think that someone is using software to attempt to replace what I do, and that would make me very unhappy. Not just someone copying my style, but someone using my work, my actual work, to make something resembling new work that I did not make. Could you see why, if you were that artist, how what might be a fun little experiment for you, might seem like an insult to the meaning and purpose of what they do? Because the software doesn't make art for the same reasons these artists make their art. It doesn't have their intention or motivation. It just regurgitates style. And art isn't just style.
While I think what you're doing is intriguing, I don't think you're doing it in a thoughtful or responsible way. You're not thinking about anything but the results you can get. There's a whole conversation to be had here that you don't seem that interested in. And I think that conversation is potentially a lot more interesting than these images you're throwing together in a day or a few hours.
I think your videos are really fantastic. Keep going. The advice is super helpful
I was able to make some amazing high quality art. Right now the art is being used for book covers and other things. It really works
Thanks, I really like your videos!
Great reference! Thanks so much! Very helpful.
I use MJ, Dalle2, and Stable Diffusion. Not only hoping to use them for my own graphic novels and picture books, but trying to figure out some interesting ways to incorporate them into my youtube animations as well. Exciting times!
David Finch …. you're welcome
Cool video. Watched your other graphic novel vid. Subscribing. Don’t about about the Reddit trolls. Keep at it.
I love your website! How did you make it? Can we get a video tutorial on that also?
How do you get a a consistent character image out of Midjourney?
I’m an artist and writer, and I’ve been closely following this ai creative explosion. I tried a few names myself and most of it came back as gobbledygook like you did. But it is interesting to know that once the algorithm has been fed enough samples, it can start to mimic an artists style. Strange days ahead for sure.
Nice, promoting tools that would make millions of people lose their jobs. Very, very nice.
well all i can say is that by this way we can only create a short story but it is still awesome like anyone can create the story that they have in their mind for long
I'm a computer science programmer first and foremost, and a business major second, so with that said, I'm mainly concerned with the legal ramifications of AI generated art – if it's all proprietary, where did the training data come from, who owns those rights? In the long run I'd say that, just like how photography didn't replace lifelike drawings, and digital art didn't replace physical art, I can't see AI art replacing digital art. It's just another tool, to serve as inspiration or a starting point – to be touched up, modified into what you need.
Personally though I'm just concerned on a societal level on the fundamental architecture this AI method works on. It doesn't create anything original, it just regurgitates what it thinks we would like based on what already exists – and when you can control what exists and what doesn't (through what it gets trained on), you can lead to issues of overrepresentation or reinforcing of harmful stereotypes
I noticed that some people are very resistant to AI in art and other creative endeavors, and I mean practically intolerant of change. But, to be honest this is the world we live in now and most people use AI every single day without even thinking about it. I'm not sure if they are intimidated by these tools or dislike how accessible it is. There's so much gatekeeping that aims to restrict, quantify, and even diminish creativity. This is kind of funny considering that these tools are somewhat useless without intentional guidance from a creative person!
"There is no Spoon" – Watch Phil become a millionaire by next year… "Heed his word"
I’ve been thinking about incorporating Ralph Mcquarrie elements for all my AI generated art!!! I knew it was the Star Wars guy but I didn’t k ow his name, thank you!
My question is can you legally use this AI art and sell it? If anyone can use these prompts can't multiple people have the same art?
Absolutely fascinating, thank you!
Don't hold your breath for the artist community to accept this anytime soon, but you're right, the cat is out of the bag. I once dreamed of being a comic artist but long ago realized the competition was too intense. Now with AI art generation only going to get better and better, better to find other ways to make a living!
i really like this and the last video about this topic
keep up the awesome videos.!!!
I'd love a more dive into the publishing and printing routes of it too if you would
I have to say that Stable Diffusion does a way better job at "copying" styles than MidJourney and you can run it locally if you are rocking a GPU with tensor cores (the oldest would be the NVidia GPUs from the Turing architecture like the GTX 1660).
Btw, someone already said it, but if u put the artist prompt at the end "by xxxx" it will give u better results. Also adding characteristics of the art style may help.
Dude. I really second Winsor McCay. I've done some like from him, and it's really cool.
Great video again! I duplicated every one of your 25 prompts by an artist with Stable Diffusion. The only difference is I appended the artist's name at the end like ", by ralph mcqarrie". My settings were 768w x 512h, 50 steps, CF Guidance Scale 10, Sampler Method – k_euler_a, and did 4 with a random seed.
My results were far closer to the artist than your MJ in most cases. Just need to play around with settings and the prompt. It takes a while but is great when you get the desired result. AI art isn't just straight forward instant art. It takes a lot of work too. I know as Ive created over 10,000 AI art images this year already! lol
Love to know how to turn those AI graphics into t-shirt design!
Keep this stuff coming!
"like photoshop or vodka" – rotflmao! love it!
So is plagiarism allowed now, if the output comes from a machine? How can one feel proud of themselves while ripping off the superficial aspects of genius artists who spent their whole life achieving that magic? How is this different than forgery, besides that it requires less skill? What about copyright laws?
For Ralph Mc Quarrie, perhaps try this one "…+ painted by artgerm Alphonse mucha, Akihiko yoshida, sakimichan, krenz cushart,…"
Appreciate the information. I enjoyed the surreal journey of your first AI derived graphic novel. Now I am enjoying your informative videos on the AI assisted creative process. As another commenter mentioned, this process seems to be a lot like collage. Maybe this is where scrapbookers/junk journalers and comic creators collide …?
Great stuff and very useful. Thanks
Please, keep developing videos about mj
you are an artist now!
You can upload your own image to MidJourney and use that as reference.
How to upload an image into the chat (no larger than 8mb)
1. Press the + (plus) Button to get the option to upload image.
2. Right click on image and view source
3. Copy the sourcelink from the browser
4. Type the Imagine command and insert the address you copied from your image.
I havent been able to use this yet due to my trial period having come to its end but Im sure some of you will find it useful 🙂
Cheers and cheerio
I’d love tips and tricks for establishing consistency frame to frame, like sat for characters. I notice some MJ comics use celebrity likenesses so that they can have characters look consistent from page to page, but what about cars, objects, locations, etc.
Fascinating and informative. I also appreciate the diversity of genres and art styles you used to sample from, though I would love to have seen at least one female artist's work. How would Midjourney do with Alison Bechdel or Julie Doucet or Renée French or Tove Jansson or Rumiko Takahashi or Pia Guerra or Emil Ferris or Posey Simmonds or Jessica Abel or Aline Kominsky-Crumb or …?
I really like Yuji Shinkawa or Akita Toriyama as artist prompts. I dabbled with Todd McFarlane as well and got some interesting returns.
When AI gets better at copying certain styles of art, eventually there will be loads of lawsuits against AI artist… The art community already encourages to NOT buy AI art.
The best style match I got from midjourney was when making space ships in the style of Chris Foss.
What about Simon Bisley?
Thanks for doing sharing all this work. Very cool for someone like me, little knowledge of artists styles, but interested in learning more. This MJ stuff is proving to be a free way to learn.
Dig the subtle but engaging pokes of fun at your self too. Fun! Thanks dude.
You are Idea is great!!!Ai will not make good comics,without good intentions, and a clear vision, thanks for sharing your great idea with us!!!
Aha! I thought I saw a lot of Robert Crumb influence in the comic book video that you did before (?) this one. Great list of comic/ comic book artists.
Or maybe I'm seeing Wolverton in your comic book and thinking it's Crumb
If I recall correctly, Midjourney has something similar to Stable Diffusion img2img called "Remaster" (does it have it? I'm not sure. Can't remember). With that, you can add your sketches or photos as guides to generate images. I don't know if Midjourney has a tool to keep the consistency, like Textual Inversion, a way you can train AI to understand a character and reproduce it in any generated image later.
Many thanks for your time and effort. Any ideas on how to control perspective/camera angle? I'm thinking worms eye view, aerial view etc. etc. Also, love the –stop command however, once upscaled it loses the soft effect no matter if –uplight was used in the original command or chosen asn a main setting…
Ai storyteller,ai artist,ai musician,ai videomaker, ai coder…Seems like the whole human creative process is being trivialized and we are going to be left with only machine operators not artists
I like Kentaro Miura from Berzerk Rob Shields from Neon Cyber Ben Templesmith epic vampire stuff and Dike Ruan epic too.
Been sending my work to u on Instagram I wonder if you could do a video that's character design related for adding dialog between characters also thinking of some original art into the story thanks enjoying the journey think it's just the beginning of the possibilities my story is looking great
Also would appreciate a good recommendation for printing on demand preferably your book looks amazing
Great taste in comic artists, but I cant help seeing the irony of acknowledging an artists work as 'unique' and 'iconic' and then proceeding to try and get an ai to replicate it. Artists who I'm sure when they shared their art on line would never had thought or consented to it then being used by an ai.
I have a doubt, there is a lot of talk about the image rights of the images created by the ai, but what about the images that are used to feed them. For example, if I want drawings like Frank Millar, I had to feed the AI with Internet images of Frank Millar that may have copyright because they are living artists and their works are not in the public domain. this does not cause problems. apart from the ethics of having a frank millar zombie in ai.
…i might say try diversifying ur artist list a little and also using a grammatically correct sentence is always good as it does seem to understand plain english eg "a portrait of xxxx by xxxx in the style of xxxxxx, xxxxx, xxxxxx" these last 2 prompts are extra styles u miight like to add eg dramatic and/or 80s etc.. i dont work in midjourney but i have and it seems to work generally the same across… working on something similar and altho i dont wana give my whole formula away it seems like some pointers is fine at this point XD
You know what people love about comics? That comic artists made them. If you love comics and graphic novels and want to make them, do what the rest of us did and devote years/decades of your life honing the multidisciplinary craft that includes but isn’t limited to: storytelling, composition, line control, anatomy, perspective, camera work, design and layout skills, in addition to just plain old draftsmanship. This should be one of the only places in entertainment arts that should still be held up as an honorable place for under appreciated ARTISTS to thrive creatively and this kind of new-tech tourism is the artistic equivalent of invading another country because they’ve got resources you want. Shame on all of you.
I'm getting some good results with Frank Miller and Alan Moore when it comes to their original characters. Frank Miller is doing good with Daredevil, Batman, and Sin City. While Alan Moore is doing pretty good with Dr Manhattan. Just a few I thought I'd bring up.
For comics and graphic novel work at Midjourney, I've had much better options using black and white instead of color. Also, Geoff Darrow is another artist with a fairly unique style who has enough work for Midjourney (or at least it worked for me).
Can you tell me where you posted it on Reddit so I can post my work there as well? I want to see what kind of hate I'll get. It would make my day.
thank you!
Another excellent video Phil!