PERFECT Consistency for Your Midjourney Character
Vložit
- čas přidán 5. 07. 2024
- 📙 Midjourney COURSE mastersofmidjourney.com Beginners to Advanced
🚀 FREE Midjourney Cheat Sheet tokenizedhq.com/freebies/mj-c...
🔗 FREE Supporting Material tokenizedhq.com/freebies/vide... Folder: 2023-07-14 - Crazy Simple Character Consistency
🔗 FREE Promptalot Extension promptalot.com/extension
Add Midjourney to Your Own Server • Add Midjourney Bot to ...
Insight FaceSwap Github github.com/deepinsight/insigh...
-
📰 AI Newsletter 👉 Your Inbox tokenizedhq.com/newsletters/ai
🐤 Follow me on Twitter / chrisheidorn
💼 Follow me on LinkedIn / christianheidorn
📸 Follow me on Instagram / christianheidorn
💬 Join the Tokenized AI Discord tokenizedhq.com/invite/discord
-
🎬 Character Design Series 🎬
Playlist: • Create CONSISTENT Char...
Part 1: Create Consistent Characters • Create CONSISTENT Char...
Part 2: Place Characters in Action Scenes • PLACE a Character in A...
Part 3: Apply a Consistent Character Style • Apply a CONSISTENT Cha...
Part 4: Create Multi-Character Scenes • Create MULTIPLE Charac...
Part 5: Give Your Character Emotions • Create FACIAL EXPRESSI...
Part 6: Infuse Your Characters with Themes • INFUSE Themes into Cha...
Part 7: Consistent Characters Still Work in V5 • CONSISTENT Characters ...
Part 8: Build Scenes with Consistent Characters • BUILD SCENES with Cons...
Part 9: Crazy Simple Character Consistency • PERFECT Consistency fo...
📺 Recommended Related Videos & Playlists 📺
Commands, Syntax & Parameters: • Commands, Syntax & Par...
Prompting Methods & Experiments: • Prompting Methods & Ex...
Character Design Series: • Character Design with ...
Storytelling: • Storytelling with Midj...
Watch Me Prompt: • Watch Me Prompt in Mid...
This Brand Does Not Exist: • This Brand Does Not Exist
Midjourney for Merch: • Make Money with Midjou...
-
I know this topic may not be new for many of my viewers but considering some of the questions and comments that I keep getting, I think I need to address this anyway.
A recurring criticism of many of my videos related to character design is that the characters usually aren't 100% consistent.
And yes, if you take a very close look at the faces, their features are not always perfect.
This is especially true for my video in which I demonstrated how you can use ChatGPT and Midjourney to create a simple comic book.
While I do think that many tend to exaggerate the need for absolute consistency, I understand that it seems to be a major source of frustration for them.
But here's what I don't understand: There are very simple methods out there that solve this problem and they are not secret.
So in today's video, I'm going to make sure that every single last one of you finally understands how easy it is to give your characters a consistent face.
Cause let's face it. That's what everyone keeps complaining about.
Hopefully, after today's video, we can finally put this silly complaint to rest and start focusing on all the great things that generative AI art has made possible.
Remember, just a year ago, none of you were able to create anything remotely like this.
Focus on the positive!
-
⏰ Timestamps ⏰
00:00 Perfect Character Consistency
02:11 How to Get Insight FaceSwap
03:29 Preparing the Assets
06:39 Swapping the Face
08:38 Swapping Multiple Faces
10:36 Swapping 1 of 2 Faces
13:29 Limitations of FaceSwap
#midjourney #midjourneyart #midjourneyv5.2 #outpainting #consistentcharacters #aiart - Věda a technologie
Swapping Multiple Faces
InsightFaceSwap supports swapping multiple faces in one image! To do this:
1. Register multiple source photos with `/saveid` using different names.
2. Use `/setid` to set default names separated by commas:
/setid name1,name2,name3
The bot will swap left to right based on the order you specify. You can also use `_` to indicate a face to skip swapping:
/setid me,_,you
Pinning this excellent comment 🙌🏻
I wasn't aware that Insight FaceSwap supports this already. Thank you for your helpful contribution. Much appreciated 🙂
I tried this option and it says it’s for subscribers only
They made it a paid feature recently.
I had no issue with Inswap putting my face of comic book characters. The colors and styles matched the art.
I didn't say it doesn't work. I said it depends on the style of your comic. It's less reliable for graphics novels that use line-art.
Wow! I've been waiting for this. You've really delivered the goods. Thank you!
My pleasure!
Thank you for this series. Thank you. I've been pulling out my hair since November trying to write a story, and you've finally brought some sanity back into my life. I think I'm going to sign up for your course.
I know the feeling. Usually it's a good idea to just compromise. The tech will get there faster than we think.
This is fantastic. Thank you for making this vid and sharing your knowledge.
I LOVE these different art tools videos! 💫🌟⭐🌟💫
Glad you like them!
Great overview and clarity as usual, I hope this extends to object swap, keeping a consistency with a unique object, e.g. a concept motorcycle that is part of the identity of a character….
Thanks for sharing! 🙏🏻
Any videos on how we can have greater control in Midjourney are appreciated. Character consistency is very important for projects like comics and graphic novels.
Midjourney is random. I don't see that changing all that much anytime soon.
Very nice! I can't believe this wasn't on my radar. Thank you. 👍
It's been around for a while, so it's not super new but I know that many people still aren't aware.
thank you for your upload and the knowledge here. very nice from your subscriber bro!
Thanks for the sub!
Another great video!! 🔥🔥🔥
Appreciate it!!
Great idea! Thanks!
This is REALLY useful as I've recently started working with an author,and although I've got a reasonable consistency with characters, it's not always perfect - THANKS !
Depending on the project you're working on, it's often the only way to get the results you want.
Hello from france, for real thank you very very much for your vid!!!! It's so clear and very good.
I dont even understand why you just have 1k like ...
ppl press the damn like button
Thanks for the tipps! I think i have to use my own server. Thats a good idea!
Wow, this is a great feature. Thanks.
You're welcome!
THANK YOU VERY MUCH! I'm very excited to try this!!
You tried it?
Yes, it works easily and well.
Have fun!
@@reycaribe Yes it's easy and free. It has trouble wit extreme mouth positions but is great with eyes and nose.
Useful for book covers and character designs.
Definitely!
Great tutorial thanks... 👍
You're welcome!
Yes, this is complex, but I have tried this and it works. Thanks for sharing,
Awesome 😍😍😍
In trying this out, one of the things I've noticed is that Midjourney's preference for "perfect" faces can make it harder to see the difference between an image that's been generated and one that's had a face swap with a different generated face. I was trying out a few, like "Nordic woman" that I picked up from CZcamsr, cyberjungle, and I found that when I swapped the faces with a specific Nordic woman's generated face, it felt a bit off. I then tried using my own face with a different scene I'd created, and the effect was much more convincing. I think this exhibits the kind of feedback loops that AIs can get into when reconsuming their own output or even output from other AIs. At least for creative outputs, this seems like creating the headshots should add distinguishing, "uglifying" features to produce more realistic effects.
As with everything, some faces will work better with certain scenes than others. That's part of the process.
hugging face has some face swap spaces as well. If you do not want to wait in the queue. They are not as good as insight but they will do the trick. Set your adversarial defense ratio to 10.
Great tip 🙌🏻
Thank you
You're welcome
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?
You can't change your own photos in Midjourney.
You're a legend.
No, you are! 🙂
I wonder if you did a face swap on a comic book style then referenced that image if after a few re-rolls would produce the correct comic face. I will have to try that. Nice job on this video
where is this video you are showing at the beginning ? chapter 8 is the grand canyon/ pan out / vary examples. Am i missing a video here?
I linked it in the description.
Hi! Very informative video . Are you using affinity photo 2 ?
No, but I'll be upgrading soon.
thanks for the tutorial, it's still hard to implement
Thanks for the visual tutorial. Tell me, if three different portraits were saved, does Inswapper always use the last one? Do I need to use the /swapid command to select a specific portrait?
It always uses the last one you added by default. But you can change it by using the SETID command.
@@TokenizedAI Christian, thank you for the answer
hey. that's great however this 'saveid' thing isn't working for me. I am upload any image after giving the id name. How to go about doing that
This is awesome! Thanks. "Picky" made me laugh. I would imagine that anyone choosing to watch these kinds of videos are, indeed, picky. ;-)
I've noticed a goal of mine is to guess the image you will choose on upscale. 😂
🤣 What's your hit rate?
@@TokenizedAIlol.. probably 90%?
Is it because I'm so predictable or because you largely approve of my choice? 😅
@@TokenizedAII would say the later. 😅
insanely crazy. I have to join your course, no other way. Is there a code I can use? BTW plausible video as the rest of your work. Thank you.
Looking forward to welcoming you! 🙂
I can't access the InsightFaceSwap Bot, as I don't have the Discord AUTH code, anyone knows what the code is or has a solution?
My question is, how to do variations with a new swoped in photo? Because under this, there are no "variations" buttons, and the photo even is not in my archive, so that i could work on with the picture id? Thank you
There are no variation buttons because it's not Midjourney. It's an entirely different tool. You do the face swap at the very end.
@@TokenizedAI ah ok, Thank you
I'm creating a set of cartoon characters for a web app I'm building - my issue is for MidJourney to (re)use the same visual style and aesthetic when generating each new character, ie. detective, dancer, monk. I have one character image that I absolutely love, but now I find it impossible to replicate that visual style when generating subsequent characters....any ideas?
Work on getting the style prompt right first, so that you can more or less reliably reproduce the style. Focus on more descriptive words. Once you have that done, then re-use the style prompt for different subjects and settings.
I thought I developed my own brunette woman character for my thumbnails on YT, but then another guy in my same sub-genre came out with the exact same brunette later that same day. So I think MJ5.2 has a limited set of faces/characters, and if you choose the prettier ones, chances are thousands of other people are choosing that same character.
Choose the "ugly" one's
Just upload two brunette movies stars and combine them... or combine one ugly and one pretty lol.
😆lol
13:35 I hope the FaceSwap will become better at comic book styles. This is a problem with many AI tools. They do well with realism, but suck at comic book styles.
You can fix them with some post-processing using the Roop extension in Stable Diffusion. But it's a bit of a hassle.
Hey, can you please show us how to work with midjourney automation software
What do you mean by "automation software"?
@@TokenizedAI Sorry, I was not clear, I saw some websites helping users do bulk prompts and upload automaticaly all upscaled images in a folder, and there was an artical about automaticaly downloading images into folder. I am using midjourney for 6 months now and I have a process, so I would like to become also faster in my work.Any ideas?
Can you please make a video on how to setup and use your PROMPTALOT
Soon. It's not quite ready yet. I'll share a video on the extension very soon. But the website itself still needs some work.
@@TokenizedAI we all thank you very much
hi Christian
The saveid option does not work for me!! This is while I am using version 5.2!
Then ask their support for assistance. That's what their Discord is for.
This has nothing to do with your version of Midjourney. There's absolutely no connection between the two.
@@TokenizedAI thank's for you
🎯 Key Takeaways for quick navigation:
00:00 🤖 *Se puede usar Inside Face Swap para reemplazar caras en imágenes y mejorar la consistencia del personaje.*
02:20 🛡️ *Es necesario configurar un servidor de Discord privado y agregar el bot de Inside Face Swap antes de poder usarlo. *
03:41 💼 *Se debe guardar la identidad de la cara que se desea usar para reemplazar con un nombre y la imagen correspondiente.*
08:05 🎨 *Inside Face Swap funciona bien con ilustraciones realistas, pero no tanto con estilos de dibujo de líneas o de cómics.*
10:43 🖌️ *Se puede usar edición de imágenes manual para reemplazar solo ciertas caras en una imagen y no todas.*
Made with HARPA AI
Useful for the most part but I want to do everything with Midjourney, don’t want another tool to pay for, just being real.
I'm going to be very frank on this one. Your expectations are unrealistic and they will remain unrealistic. In the end, no matter what you do, you're always going to need several tools. Just like you'll need an image editor at some point.
PS: The tool is free. So I don't see what the issue is 🤷🏻♂️
Love your channel! Is there a way to use the new Midjourney 5 tools, like zoom and pan, on images you've created with previous versions of the app, like Midjourney 4? (If so, that would make a great vide.)
I've tried to reply a couple of times with the answer but it keeps getting deleted. Not sure what the issue is. Maybe YT doesn't like using forward slashes?
1. Find the original image (quad) on your profile pg.
2. Copy the jb identifier.
3. Return to Discrd.
4. 'fwrd-slsh' Show
Does this method work for Pixar style illustrations?
Probably not
It is a great tool but I agree it doesn’t work well and in my case doesn’t detect faces which have greyscale tones or cartoon characters. I desperately need a solution for it.
Yeah, there's always a caveat 😅
I would love to see this work with manga styles, as well as other comic book styles.
Not working for my Pixar style either
I just tried that but it’s been an hour already and the swap is not done yet
Then something is wrong
Your stripe URL is down.
Seems to be working on my end 🤔
Probably, as its popularity grew, many new users flocked to it. Until yesterday, I could use it, and it allowed 50 face swaps per day, and the queue never exceeded 600. However, today, it shows this message: "failed, there are more than 1000 commands in the queue, please re-submit later"
There are other face swapping options out there. But they require you to install Stable Diffusion.
The bot isn't showing up in my list of users ... Why is that?
I don't know. Please ask Insight FaceSwap's support in their Discord.
@@TokenizedAI Hey, thanks for your response! I actually came across the solution online. Here's what to do for anyone having the same problem as I did: Click "Edit Channel" on your server. Then go to "Permissions". Click "Add Members or Roles" then under "Roles" select "InsightFace." When you go back to your channel, you'll see the bot displayed as Online 🙂
Keeps telling me face detection failed 😢
Does anyone know how to do this in Leonardo?
No, but why don't you just do it in Discord? Takes 2 minutes.
@@TokenizedAI you mean Mid Journey? Im too lazy to learn it, plus it's not free AFAIK.
Thirtees? lmao
Typos happen 🤷🏻♂️
To fast, too confusing. Your voice could be easier to understand.
I am truly grateful for all the knowledge you bring for all of us. Is there anyway I could reach you privately ? ( I tried to connect via LinkedIn already 😄 ) I am working for a non-profit organization in Vietnam. Thank you !
Say @TokenizedAI can I get a code for the course please