Great video! For those who don't want to retype the original prompt: /imagine little gray wolf character, multiple poses and expressions, 100% white background, children's book illustration style, simple, cute, full color, 3d model, light and shadow, no shoes, gray hat, gray hooded dress, gray clothes, flat color, --no outline --s 50
Tutorial article: semicolon.dev/midjourney/make-consistent-characters (outlining the process of creating consistent characters in Midjourney for those who like to read instead)
this was great. unfortunately, i've done this multiple times with seed and image link and never get the same character twice. what am i doing wrong!! Sigh.
It's definitely very interesting to see the techniques, but this is not good enough to create a children's illustration book. Characters have to be the same throughout the entire story.
if u want much better ur experience I'll introducing AI CHAT DEEPAI the AI CHAT DEEPAI is so powerful and much better than OPENAI since start 2015-2018 and u can show ur picture and video and anything u want
It does actually I’ve done entire generators fine tuned on just prompts which is basically useless for normal LLM’s. But MJ has extremely strict weighting and biases for seeds and prompt tokens
@@AtomicPixels man i would like to have a convo about this i have been struggling so bad i could use a bit of advice i could pay you for some explaining
great tut, thanks so much. tbh though, seems the ai isnt really great. no angle control, no zoom, no pose correct, seems still extremely obtuse and sluggish. It's still learning thought, right?
That's true, but this was done in previous version of Midjourney. Also, it was never designed to make consistent characters. All of this is just techniques for increasing chance of generating similar characters.
i do follow your instruction but when I sent my prompt it failed Invalid link The content type of the provided link is text/html, instead of an accepted image type. I try the 2nd time with input 6 links --s 0 (input value) together with --seed and --cref url but the result not consistent. Something wrong?
It's only worth getting most expensive one for privacy mode. All your images can be seen publicly elsewhere on Midjourney. I used to have 3rd plan, but downgraded to $10, as I haven't been recently using it that much anymore. Might pick it up again next year for art references.
Your directions are so clear and I'm following step by step and I'm still not getting it to do what I want. Ugh! What am I doing wrong? I ask it to have the girl holding a baby, put in the URL for a generic baby, but then it gives me images of a girl holding books or dogs. Help!
Thank you at all times. Is it only possible to take pictures from Midjourney? Is there any way to make my photos various poses and facial expressions for example?
5:36 you'd save a lot of time and energy with Christian Heidorn's Chrome extenstion "promptalot" which means no more cut and paste. It's super helpful.
Actually, that's to load an image (either .jpeg or .png) as a character reference (--cref). To paste the image into the chat line, just right click on it, copy the link address, then paste it directly into the chat line. I hope that helps?
For some reason, when I include the "multiple poses and expressions" when asking Midjourney to illustrate identical twins, it doesn't give me the multiple poses and expressions. Any tips?
bro i have spent around 10 minutes just to get myself logged in ghost together but every time i try your web automatically gives me a display name which isn't valid at all other then that the display names given by me is not accepted As well
Thanks, question. What if you have already designed the cartoon character without AI but want to create more poses and different scenarios with the character using AI. How can this be done?
Process is the same. Just instead of generating the initial character images, use images of your own character for the image prompt. Add them as URLs into your prompt. Might have to create a few poses of your character first though not just one.
@@ghosttogether My character was actually made by Midjourney, but it took AGES to get it to do a hand-drawn children's book character and now it seems unable to reproduce it. I keep trying to get more poses and expressions, I even hired someone to help guide me and learn how to make better props but the damn thing can NOT recreate this one style. It really wants to make it look more like digital art. 😕
@@kendalbrenneman Yes, I have that same issue. I've tried seeds & importing files and MidJ just recreates complete new images, but I'm still searching.
Thank You for this great video. It's very clear and well made. I appreciate all the effort that went into making it for 'us learners' out here. Is there any way I can split the images to them copy/paste them without using Photoshop? I managed to separate the images but can't get the URL for the next stage. I'm trying the learn how to do illustrations for children's books - it's all very new for me. So any help is greatly appreciated. I'm trying to join Ghost Together on Midjourney.
Yes you can do the same with any image editor. On windows you might use something as simple as MS Paint. Same idea, open, draw marquee around the character and Ctrl+X. Then Ctrl+V into Discord.
That depends, on what you mean by "get into" a 3D model. You can use this as 2D texture. But you still need to create a 3D mesh. There are probably better AI tools out there for that.
but thats still just one frame you got over there , you cant make animations from it. You got 1 sprite and all 4 look different you cant make animation
That isn't how Midjourney works. This is why this video is of the technique, that helps increase chances of generating a similar image, rather than a magic trick to actually generate the same character. If it actualkly worked that way, there would be no need to make videos on the subject.
nah this is not going to work too well you're better off just making the designs yourself! People will be able to tell it's not as authentic as it should look but You might get lucky people are gullible!
![[#2024MAMA] BIGBANG (빅뱅) - 뱅뱅뱅 (BANG BANG BANG) + FANTASTIC BABY | Mnet 241123 방송](http://i.ytimg.com/vi/EQsYfiPR9uM/mqdefault.jpg)
This video was full of false promise and half-baked techniques with hit-and-miss results.
so was your birth
@@ghosttogether BOOM
Great video!
For those who don't want to retype the original prompt:
/imagine little gray wolf character, multiple poses and expressions, 100% white background, children's book illustration style, simple, cute, full color, 3d model, light and shadow, no shoes, gray hat, gray hooded dress, gray clothes, flat color, --no outline --s 50
Thank you!!! I was hoping to find a comment like this LOL!
I noticed in the running clip it kept the character look the same but changed the fur color. how do we prevent it from changing that?
Tutorial article: semicolon.dev/midjourney/make-consistent-characters
(outlining the process of creating consistent characters in Midjourney for those who like to read instead)
this was great. unfortunately, i've done this multiple times with seed and image link and never get the same character twice. what am i doing wrong!! Sigh.
He didn't either, they're different colours
It's definitely very interesting to see the techniques, but this is not good enough to create a children's illustration book. Characters have to be the same throughout the entire story.
if u want much better ur experience I'll introducing AI CHAT DEEPAI the AI CHAT DEEPAI is so powerful and much better than OPENAI since start 2015-2018 and u can show ur picture and video and anything u want
What did you end up selecting ?
It does actually I’ve done entire generators fine tuned on just prompts which is basically useless for normal LLM’s. But MJ has extremely strict weighting and biases for seeds and prompt tokens
@@AtomicPixels man i would like to have a convo about this i have been struggling so bad i could use a bit of advice i could pay you for some explaining
These works are most precious and quite advanced! Without which MidJourney is near useless for practical animation works
You are explaining extremely good!!!!!!! Fantastic job!
Subscribed!🥰
if you pay attention to the details it not the same character?
Thank you very much! It's finally working!🤗
The whole point of the video is that the characters are consistent but they literally look different in every frame even in your example...
great tut, thanks so much. tbh though, seems the ai isnt really great. no angle control, no zoom, no pose correct, seems still extremely obtuse and sluggish. It's still learning thought, right?
Great tutorial! Thank you for the tips!
nice.. but the wolf you chose at first and the running wolf does not look alike at all nor are they wearing the original outfit :(
Great stuff. The characters though look different from the original, the details on the clothes are different.
That's true, but this was done in previous version of Midjourney. Also, it was never designed to make consistent characters. All of this is just techniques for increasing chance of generating similar characters.
@@ghosttogether Thanks for the tutorial but that isn't what your title says... 👀
i do follow your instruction but when I sent my prompt it failed
Invalid link
The content type of the provided link is text/html, instead of an accepted image type.
I try the 2nd time with input 6 links --s 0 (input value) together with --seed and --cref url but the result not consistent. Something wrong?
thanks. do you use the $10 plan or the $30 one?
It's only worth getting most expensive one for privacy mode. All your images can be seen publicly elsewhere on Midjourney. I used to have 3rd plan, but downgraded to $10, as I haven't been recently using it that much anymore. Might pick it up again next year for art references.
how would you make a mother or father grey wolf but in the same style as your selected seed?
Such a dreamer 😄
That was great man
This is good 👍. I teach Midjourney and it always makes me happy to see demo videos with good information so well done.
Thanks :) Glad it's helping someone out there. Just sharing what I'm learning myself.
how does it feel to be pathetic at actually making art?
Your directions are so clear and I'm following step by step and I'm still not getting it to do what I want. Ugh! What am I doing wrong? I ask it to have the girl holding a baby, put in the URL for a generic baby, but then it gives me images of a girl holding books or dogs. Help!
Thank you at all times. Is it only possible to take pictures from Midjourney? Is there any way to make my photos various poses and facial expressions for example?
cool stuff thank you so much
the link to the community isnt working
Sign up at GhostTogether.com/midjourney to join my Midjourney community
5:36 you'd save a lot of time and energy with Christian Heidorn's Chrome extenstion "promptalot" which means no more cut and paste. It's super helpful.
Lets say that i have an image with some characters. How can i use that to make more of them in same style?
Hey this is useful but it does not send me seed numbers after i submit envelope emoji
You have to go open the DM that Midjourney sent you after you press the "envelope" emoji.
Great Tutorial, for me it helps great
hi. thanks :)
💯💯💯
I don't understand how you can paste an image inthe chat line. I cannot do it . when I try nothing happens
I'm learning too, but I believe you need to click on the plus button right before the Message/ prompt, then upload file
Actually, that's to load an image (either .jpeg or .png) as a character reference (--cref).
To paste the image into the chat line, just right click on it, copy the link address, then paste it directly into the chat line. I hope that helps?
How u do multiple chatacter
Sora next?
For some reason, when I include the "multiple poses and expressions" when asking Midjourney to illustrate identical twins, it doesn't give me the multiple poses and expressions. Any tips?
slap it in the ai-cheeks
That's amazing 👏
It seems SD with control net do a good job than mid journey
I thought this video was in reference to creating consistent additional characters 😢
bro i have spent around 10 minutes just to get myself logged in ghost together but every time i try your web automatically gives me a display name which isn't valid at all other then that the display names given by me is not accepted As well
What's your username. I'll check if your account was correctly verified.
Thanks, question. What if you have already designed the cartoon character without AI but want to create more poses and different scenarios with the character using AI. How can this be done?
I have the same question.
Process is the same. Just instead of generating the initial character images, use images of your own character for the image prompt. Add them as URLs into your prompt. Might have to create a few poses of your character first though not just one.
@@ghosttogether My character was actually made by Midjourney, but it took AGES to get it to do a hand-drawn children's book character and now it seems unable to reproduce it. I keep trying to get more poses and expressions, I even hired someone to help guide me and learn how to make better props but the damn thing can NOT recreate this one style. It really wants to make it look more like digital art. 😕
@@ghosttogether thanks!
@@kendalbrenneman Yes, I have that same issue. I've tried seeds & importing files and MidJ just recreates complete new images, but I'm still searching.
What If I want something more mature n adult example like witcher game
Not in Midjourney. It'll stop you from generating.
Thank You for this great video. It's very clear and well made. I appreciate all the effort that went into making it for 'us learners' out here. Is there any way I can split the images to them copy/paste them without using Photoshop? I managed to separate the images but can't get the URL for the next stage. I'm trying the learn how to do illustrations for children's books - it's all very new for me. So any help is greatly appreciated. I'm trying to join Ghost Together on Midjourney.
Yes you can do the same with any image editor. On windows you might use something as simple as MS Paint. Same idea, open, draw marquee around the character and Ctrl+X. Then Ctrl+V into Discord.
@ghosttogether Thank You for your reply, I'll see if I can find a basic image editor on my Mac. 👍🙏
On a Mac, check out Pixelmator :) It's a small one payment app, but it's well made, like Photoshop.
Can you make the images you get into a 3d model?
That depends, on what you mean by "get into" a 3D model. You can use this as 2D texture. But you still need to create a 3D mesh. There are probably better AI tools out there for that.
iam looking for same pose......
it's really interesting but isn't midjoyrney paid?
It is
this one very good tutorial but im not a storyteller so no need to join ghost whatever site
lol
this is cool but what about AI generated sprite sheets, can it be done using midjourney or other?
They should implement a "render on alpha channel" options.
It doesn't respect the "100% white background" prompt ever... :/
How would you accomplish this with a photorealistic image of a person that is not a cartoon character?
but thats still just one frame you got over there , you cant make animations from it. You got 1 sprite and all 4 look different you cant make animation
Is this free?
Consistent-ish*
I hate to be the bad guy but your wolf images at 6:10 are NOT consistent. Look at the face for example... orange to peach to grey fur.
That isn't how Midjourney works. This is why this video is of the technique, that helps increase chances of generating a similar image, rather than a magic trick to actually generate the same character. If it actualkly worked that way, there would be no need to make videos on the subject.
@@ghosttogether my apologies. I understand now. Your tutorial is very helpful! Thanks for doing it.
What the fuck kind of difficulties are there?? Why can't he always create the fucking one we want?
It is not a consistent caracter, has not the same colors and face form
make this in a free tool
art is for artists, using midjourney is skill issue
this is far from consistant. sorry.
waw zero talent !
nah this is not going to work too well you're better off just making the designs yourself! People will be able to tell it's not as authentic as it should look but You might get lucky people are gullible!
Looks dumb
Replace your eyes.