BIG Midjourney Update - CONSISTENT CHARACTERS
HTML-код
- Опубликовано: 20 май 2024
- Midjourney Beginner Course ➡️ ftpai.thinkific.com/courses/m...
-
-
-
-
Timestamps:
0:00 Intro
0:20 How to use it
2:09 Set a Shortcut
4:54 Character Weight
5:45 Selfie??
6:29 Combine with Style Reference
8:30 Create Photoshoots
10:14 Using the Website
Releasing videos of updates is much appreciated. Thank you.
And I appreciate you taking the time to comment!
@@FutureTechPilot Hello, I have a question that has been confusing for a long time and I want to ask you. It uses real photos of pets to generate secondary anthropomorphic photos of pets. The characteristics of the pets cannot be consistent with the real photos of pets. For example, if I use a photo of a Dalmatian dog, the head of the generated anthropomorphic photo is always different from the real photo of the pet. Do you have any good solutions here? I hope you can reply to me when you are free.
@@Felixe227 Maybe 'Vary Region' could help you. Sounds like a frustrating problem for sure!
@@FutureTechPilot thank you
bro... those pictures you made with your face are DEFINITELY you 😂
Yeah. They look EXACTLY like you. Maybe it’s like when you hear a recording of your own voice… it always sounds off. But to everyone else it is 100% you.
yeah man. it was him lol...
Hajajajaja
😂 it's obviously not but I respect your opinion
It’s definitely him 😂
Finally i can make comics and visual novels with midjourney images!!! ❤
right there with you. This is amazing
The future is amazing!
How lol, the clothing will be different every generation. And the background also won't be consistent.
@@undergroundxp should be able to use the vary region for the clothing to help a bit. But yea you are right
@@undergroundxp I agree with you. Consistent faces are a good start for Midjourney, but I hope we will also see:
*1.* Consistent body types/physiques
*2.* Consistent clothes
*3.* Consistent backgrounds
*4.* Consistent objects (e.g., a unique-looking sword or a magical item for our character to interact with)
*5.* The ability to copy a pose from an image and put our consistent character in that pose
*6.* The ability to pose multiple consistent characters together in an image
*7.* The ability to create consistent characters of unique imaginary creatures that don’t actually exist (e.g., imagine a character like Pikachu that doesn’t look like anything in real life, or maybe an anthropomorphic character: It would be great if Midjourney could portray original characters of imaginary creatures in a consistent manner)
All of these things are important to anyone trying to create images for comics or graphic novels, so I really hope they will be implemented in Midjourney.
Amazing job, as always! You’re my number one MJ source of information. Thank you ❤
and I really appreciate the support!!
yessssss can't wait to try it :) thank you so so much for being so quick as the release of this update just came out. You're a champion :)
And thanks for taking the time to leave a comment!
Wow literally came to see if you posted a video on this and it says 4 minutes ago. I love your consistency with this, keep up the amazing work man!
😂 I'm running on 4 hours of sleep, so I really appreciate you taking the time to comment some nice words. Thanks!
@@FutureTechPilotAlways! Your the first and only channel I ever go to when it comes to Midjourney updates
Sooo helpful. Thank you - I miss all the office hours and really depend on you. You never fail!
I'm glad I could help, and I really appreciate you taking the time to comment!
Thanks for all the great information! I came across you a week ago when I was starting my mid Journey foray. I've learned a ton from your informative videos. Keep up the wonderful work.
I'm really glad to hear that, thanks a lot for the kind words!
@@FutureTechPilot hopefully you got my token super chat. I'm checking out your digital class as well.
yes, you're a legend!@@BradMorton
Love these "how-to" videos. Excellent. I'm using the new --cref to create a character for our online group mascot. Can't wait to put it in all kinds of situations but keep him consistent.
Isn't it so much fun? 😂
Excellent details, can't wait to play with this one. We've been waiting long enough!
Yah it's such a cool feature to have
thanks for the shortcut tip - wouldn't have thought of doing that and it makes it so easy!
Yes! I'm glad I could help
Thank You! Stunningly brilliant!
Cheers!
OMG I am super excited!! Thank you so much always for bringing us the latest info and taking the time to do so! Hope you have an awesome day!!!!👍
Right back at you!
You are so good and your very clear in your tutorials. Excellent!
Well thanks a lot for the kind words and the feedback!
I do not think people realise how much this will revolutionise comics. For the better, imho. Nolan, you're a legend.
The art world is going to be a fascinating place, eh?
I dunno. Everyone and their dog will do it now.
for real lmao@@modusvivendi1442
@@FutureTechPilot A fascinating place where its that much harder for artists to earn money with their passion. I mean, I like Midjourney, and it makes so many things so much easier, but still. I'm an author, and the idea that I can be replaced by a tool that just earns money for a big corporation... I dunno, man.
Being good at prompts and sifting through AI generated output has as much to do with being an artist as does programming a CNC Lathe with being a carpenter.
Wow, just tried this with your help. Finally I can visualize my characters in various scenes. I have aphantasia (when I close my eyes there is just darkness) so this is like getting a brain expansion, after all these years.
That's awesome to hear! Thanks for sharing that story
Awesome. This update is so valuable as it was a nightmare before. Thanks!
Cheers pal!
Great video! Thank you for the prefer option set tip!
Enjoy!
The style reference also took the camera angle, I believe you need a long shot style reference to get your alien invasion
Yah that's a good idea!
Amazing tutorial keeping in mind this feature is brand new. The other ones I saw are kinda fine but this one is expert level 💪🏻 Thank you!
Cheers pal!!
Thank you so much for this information, I've been having so much fun with this feature
I'm glad to hear that!
Awesome content. Straight to the point and answers all my questions. Liked and subscribed!
Cheers pal, thanks for the support!
Incredible. Loved your tutorial
I'm happy to hear that!
Exactly what I need. right on time.
Enjoy!!
Boom! Finally. Thank you. 😊
Cheers!
una vez mas tengo que agradecerte tus videos sobre. mejoras y novedades de MJ que me ayudan mucho a mejorar usando el programa ademas de divertirme... eres un profesor,, con un contenido en tus videos de 10,,, THANKS THANKS THANKS
Such valuable information! Thank you!!!
And thank you for the comment!
Thank you for the excellent video. Very thorough.
I'm happy I could help, thanks for the comment!
I didn't think I would like this as much as I'm liking now. It was never something I cared about even though it was one of the most requested features.
I don't think it is going very well with photographs yet, even midjourney photographs, but it is working very well with illustrations.
I wonder how popular it is at the moment. The midjourney team never created a cref-showcase channel in their discord. We're just using the sref-showcase channel and there aren't a lot of people there. Sometimes it feels lonely there.
That's a good point about the lack of a showcase channel, but I bet the team is really busy and maybe that was just something they forgot about haha - maybe you should mention it in one of the feedback channels!
FINALLY!!! This has been a huge problem for me because I could not create images for children's books, comic books, or any other project where I tell a story and include images. Believe me, I've watched so many videos with any hint of "AI consistent characters" in their tags or titles.
None of the strategies worked to my satisfaction.
You are THE MJ MAN Bro, keep up the great work!
Cheers pal! I really appreciate the support :)
great vid! thanks for your work!
and thanks for the comment!
Great teaching skills, thank you for not waffling..
And thanks a lot for the feedback, I'm glad my style of video works well for you!
Finally, this is like the one thing I needed
I know right?
I was waiting for cref to drop. Thanks for the heads up.
Hope you've been having fun!
very awesome update. thank you!
🤝
This is exactly why I subscribed and turned on notifications. You always bring us nuggets. Thanks again for working so hard to grow this community. Thrilled to be part of it!
That makes me happy to hear! Cheers pal :)
Your vids have helped me a lot, thanks.
I'm happy to hear that pal! Thanks for the comment
Thanks, great presentation !
and thanks for the support!
This video respects the viewers time. Thank you.
Cheers pal! Thanks for the comment
amazing update!
I completely agree!
AMAZING! Thank you!
And thank you for the comment!
Thank you for the video! 🙏🙏🙏
and thanks for the comment!
Goosebumps!
💯
Genius. Thank you!
Cheers!
woohoo! very fun to play with indeed!
I can't believe how far we've come!
you are the man, Nolan.
Cheers pal! I appreciate the support
Can't believe you don't have more likes here... maybe as we go along. Great stuff!
haha just trying my best! Thanks for the support
LOVE THIS CHANNEL
cheers pal!
SO Awesome!! 🎉
Feels like Christmas every time there is an update 😂
Just what I was looking for! 😁😁😁
Enjoy!
How was I not subscribed to your channel before now?! Error corrected. Great job.
haha well I appreciate the support!
Yes! Finally it's here! Great video, Nolan. Well, it's time to re-activate my subscription to MJ and dive in. BTW, isn't it possible to first upload some pictures of yourself, maybe with even a prompt and when you get a result that really looks like you, (there are always some good ones) you can then use that MJ picture for character prompts with your face. Might be a workaround to try out.
Maaaybe. I honestly think a face swapper software would be 10x easier, but maybe I'm missing something!
Thank you for releasing the info about the update! I was out of town yesterday and missed the announcement!
The Cref is great news but I am bummed that MJ did away with giving free hours for ranking images...
I just saw that too, big bummer. I'm sure they'll tell us more about that in the Office Hours today
I've just been trying out cref with terrible results!!! Can't wait to watch your video later to hopefully gain some insights on how to make it work.
and please let me know if you find any other tricks along the way!
Creating a shortcut for this is a really good idea, that seems obvious, but I didn’t think of it!
Please let me know if you find out any other good workflow tips!
This is amazing informatoin. Thank you friend
Cheers!
terrific, thank you so much
Cheers pal! Thanks for the comment
Ugh. I don’t want to pay Midjourney any money, but it’s so far ahead of Stable Diffusion that it isn’t even funny. And much easier to use too.
😂 it certainly is special! I can't recommend it enough
I don't want to pay for milk or bread, but I like to eat. 🙄
@@morpheus2573 🙄 I can grow my own food at home. Remember, open source exists.
Do it pay the money! I have not a single regretted cent spent. My wife and I have already been able to start multiple ventures with just Midjourney being the main engine of production!
@@angryox3102 All that AI fast food can't be good for you. 🍔
Thanks for all the great content you provide. Would the "zoom out" feature help to establish the "fighting off alien invasion" prompt you requested? Seems like the final result was more of a Close Up where a Medium or Long Shot may have yielded more of the particulars in the image you desired.
Hey that's a good idea!! Maybe it would work
Hey Nolan, this is great stuff.
Cheers pal! Enjoy
Hi, for uploaded images, I just discoverd the way to use it : put image reference, style reference and character reference to that same image, style raw, --s 0.
With a proper text prompt, this gives a good chance for MJ to internalize your photo 😀
😂 that's brilliant!
Thank you. You saved me the most valuable asset we have as humans now. Time.
haha I know what you mean and I'm glad I could help!
Definitely another essential video. I love how you break it down and go at a pace that we can understand and be able to pause and write down the instructions. You are a true teacher.
That makes me happy to hear! I really appreciate the feedback and I'm glad my style of teaching works well for you :)
The easiest way to incorporate other elements is change the construction of the sentence, at least in my experience: "An alien invasion being fought off by a woman" instead of "A woman fighting off an alien invasion" should ensure that the alien invasion is present in the image, though the woman might be less prominent.
Brilliant! Thanks a lot for pointing that out. Cool way of thinking
@@FutureTechPilotNo problem. Also, I have to say, cref seems to work a bit better for you than it does for me (so far), but I recognize that it's a new feature, so perhaps it will improve in the future. Thank you very much for the video. Cheers.
Fantastic! Does this work with other subjects/objects other than people?
Namely, can I get a car consistency across multiple images?
I don't thiiiink so! But maybe in the future
I love your videos. Thank you for all your work! This one was extremely helpful. I wonder if anyone knows of a way to make TWO characters consistent in one image?
I'll have a video about that on Monday I think!
Wow, this is such a big deal for me, I'm making a movie with midjourney and this is going to save such a huge amount of time since i used to re-roll promps to come up with a similar character
💯! Takes a way a ton of frustration
Woohoo! This is great. My characters looked a bit strange, but I will figure this out!
And I'm sure this feature will improve over itme!
Thanks!
Well thanks a lot!
Great videos, thanks a lot❤ And pls help, is there a way to make consistent characters from real photos other than train your own Lora? For example I need to make comics with real people as characters in various situations, poses, etc
No, not right now, unfortunately! But I can imagine that working sometime in the future
@@FutureTechPilot thanks 😊 and can you please suggest a method how to do a face swap in SD and achieve likeness?
@@tokerbabysuperfly I don't have any expertise in SD, sorry I'm not more help!
amazing bro happy day and enjoy time❤❤❤❤
Cheers!
This makes my day - I love you (platonically)
Right back at you, slick
Thanks great video
Cheers!
best update so far
💯
thanks for the tips
Happy to help!
GREAT VIDEO!!!!
Cheers!
Great job, Nolan. This is a great first step from MJ towards character facial consistency. But TBH, it isn't really a storytelling tool until we're able to pose a complete repeatable character into different environments. Let's hope MJ devs give this top priority instead of fizzing about with peripherals.
I wonder where we'll be in a year from now!
@@FutureTechPilot In one year you will have replaced yourself with your AI influencer version of yourself. You'll be smarter, better looking, able to mesmerise dinner party guests with your erudite conversation spanning a multiverse of topics in any language. Oprah will return just to interview you. Virtual Obama will award you the Medal of Freedom and Canada will give you the keys to Virtual Canada, the new AI version of the country that replaced the tired old bricks and mortar version. It's now more expensive to live in virtual AI countries than the real thing... which is so 5 minutes ago.
You did ask.
Hai nolan i love your content so much! And I’m thinking of joining your new classes, but i’ve been enrolled in several udemy class about midjourney. Is your class have refund policy?
I'm really new to this hosting website, but I *think* there is a refund policy - and it looks easy to use unless you use Paypal. (If you do use Paypal, I think I can still refund but it looks like a longer process)
Thank you for the video! Is there a way to make this work with vehicles? For example I'm trying to use this for a sci fi warship but can't get a consistent vehicle using --cref.
That's an interesting question! I haven't tried that, but I imagine it being a little difficult... I'm like 99% sure this feature was built with people in mind, but maybe in the future it will work on more objects
He Nolan great work man... If I buy the course does it refresh every new update ? Or buy again if new version comes out tnx!!
No you'll never need to buy it again, I'll keep updating the course - but that's why the price will go up in the future!
@@FutureTechPilot I love it.. ! Tnx..
I understand that you will after a while though! Tnx
This tho - 6:20 !!! Thanks for the laughs. Great tute!
😂
Thanks man, good pacing. Maybe in future videos when you read out command prompts, consider adding super imposed large text of what you're typing as you're typing it. Anyways good job, this has been a problem for me for so long.
okay cool, good idea! Sometimes I'm up against the clock but if I ever have time for some extra editing, I'll put that in. Thanks for the heads up
Great video. Thanks.
Any chance you could put a bit more detail about the online Course contents on your enrolment page? It doesn’t really say much.
And what does ‘First 25% and First 50% mean?
I’d like to buy your course but would like to make sure it doesn’t just repeat stuff I already know.
Thank you for the feedback!! I went an added much more to the page - hope that helps!
You can include --cref in the value for your shortcut.
Thank you for the video.
Yah for sure! that will save you some more time as well haha
This will be great for image to video generation to make cool little stories
The possibilities are mind blowing!
Great tips
Cheers pal!
Amazing. Now let's wait for video, I think it will be amazing.
I believe images can be pulled directly into Sora and transformed into video (including with a prompt! So if your MJ image shows your character with a gun she can be shown angrily firing that gun in your Sora-generated video). Combining this MJ feature with Sora should be a total mind-bender...
It'll be wild I'm sure!
I've been trying to figure out how to do this on the Alpha website. Thanks Nolan!
Cheers pal! Glad I could help
@@FutureTechPilotIs it possible to use with "vary region"?
@@marcdevinci893 that's a great question! I'm actually not sure lol I'll have to experiment later
Fantastic:)
💯!
Amazing video! Midjourney noob here. What does the S 200 stand for?
The stylize value - a range between 0 and 1000, the default is 100. High numbers = more beautiful images, but might ignore parts of your prompt
my gods, finally
I loved this, Nolan. But what about environment? What if I want my character to keep appearing in the same kitchen? Is there a cheat where you can have 2 consistent characters?
Cool question, but definitely not possible right now! Photoshop might solve that problem for you though ... and no, no real answer for 2 characters yet, but I'll show you something you can try in my next video - hopefully on Monday
Hi man, I love your content. Just wantted to point out that a new feature has been released, and that is /describe for V6. It's very powerful and highly descriptive. I advice you to do a tutorial exclusively on /describe because it could teach us how to build prompts in a better coherent way. And it seems that now there's barely any "in the style of" in the prompts generated by /describe.
That's a good point! I looked at /describe a little, and I didn't notice enough of a change to make a video on it. But I'll spend some more time with it and see what I can come up with!
@FutureTechPilot I've noticed a big difference on the way they build prompts. It's now working specifically for V6 as I noticed, and the description is way more detailed.
IIIINSAAAANEEEEEE!!! It´s finally there!!! YEEEAAAAZZZZZZZZ!!!! =D
Can't wait to see what people make with this
Hi, Nolan, I’ve had success with using portraits of real people as both an image prompt and a character reference to generate *new* character reference images that are 100% midjourney generated. Then, you use these as your new character reference image
Hey that's a good idea!
Hey, on your course are you going to give away a list of prompts so we can easily copy and paste? Will you be updating the prompts into categories so it’s as easy as drag and drop? If so I’m completely sold !
I forgot to mention this in the video, but the course comes with every prompt pack I've created so far. Granted, the prompts were made in Version 5, but they still work well in the newer version - and yah you can copy and paste them from the PDF directly into Discord or the Website
Excellent instructional video. Thank you! Worth checking out also prettyface ai for consistent characters. Less options compared to MidJourney though
Thanks for the tip!
question on your course: is it possible to ask you questions? Like office hours? Or is it strictly the videos? Thanks. I'm using MJ to tell stories with images and I am struggling bit--less because I watch your videos religiously. ;-)
It's just a bunch of pre recorded videos on every feature! You can obviously reach me through email if you need some extra help, but no live office hours for the course. Maybe I'll think about including something like that my next course!
As far as using your own face, do you think using a face swapper on a Midjourney image and using that as a character reference is a decent workaround for the time being until they upgrade cref?
Oh yah. for sure. any face swapper can do the trick
hi...thanks for everything, I always follow you. All good news but the thing that doesn't improve is the deformation of human bodies which is too often present, don't you think?
Oh yah, it's definitely not perfect 😂