Awesome, Julie! Love how effective this technique is, and how fast you walk through it! Well done and useful. Additonal tip: Try out giving your characters long fantasy names, that already have some traits in the name. You will be suprised how good MJ can recreate consistent characters with just like 20-40 letter tokens. At least faces, together with your technique its very easy to also get cloth and accessoir consistency! Just what I was looking for. Thank you. 🙏
This was SO incredibly helpful! I honestly hadn't thought yet of connecting Midjourney and Photoshop Generative Fill but now I have so many things I want to try 😄
Awesome tips! Thank you! I appreciate that you are so minimalist in your approach to tutorials. Better watch it, people are going to clone your voice for ASMR!
great, clear guidance here. But I cannot find information anywhere on how to use Midjourney to place an AI model within a scene I have captured in a real photograph! Any clues - tips? Thx!!
Fantastic Video - I was tasked by a relative with taking a real child's face and bringing it into Midjourney in various poses, costumes, and environments. I can see how this would work with a character is based on one done in Midjourney but if the character's face needed to be based on a real person can you tell me what your approach would be?
you could try working with a reference image (image link url in prompt), sometimes it gets pretty close to it, but it's trial and error! otherwise there's an option of using the Insight FaceSwap bot, check out this video: ruclips.net/video/TAlQJYyTiNw/видео.htmlfeature=shared
hi! I love a lot of this technique though I'm finding it has trouble with really unique faces or what some might call 'ugly' or 'average' is there a way around this?
you can play around with random names you make up that aren't related to any known personas/actors, also adding words like "imperfect skin", "asymmetric face" etc - sad we'd have to do that but it often works. and don't use words like "model"
Hi Julie. Thanks for the video. I also use PSD beta but not the way you did here but I will from now on. Question: What about generating with PSD beta and then importing back into MidJourney. Is their a way to keep the image consistent say with image weight or stylize to zero or something?
there isn't a way (to my knowledge) to feed it back to midjourney and keep all the same aspects of the image. however you can work with image links in the prompt - that's how most consistent character videos actually do it. in my experience, the faces turn pretty blurry/plastic-y, hence why I found my own solution to that problem back in march and it served me well ever since 😄
Can you do this with multiple subjects or group of people? For example I want to generate a group of soldiers marching in the forest in midjourney. Do you think it will work?
Interesting but in the future videos I´d like to see those prompts better and maybe a little bit more explained too. For example the Image split into 2 seems to be important. What is the vertical line in the prompt? or is it there. Sometimes it is and sometimes it´s not. This is too rushed and again those prompts should be better seen by viewer. But othervise very good. Just needs a lot of pause-ing by the viewer. Make these longer. I think we don´t mind.
thanks for the feedback - fair enough, working on my computer everything seems to be big enough and even on my phone it was pretty legible, but good input! Yes, the image split is important since that's the reason why you get two poses of the "same character" - where midjourney draws that 'line' is up to the tool, there's no way to prompt that it needs to be in the middle, but it usually just is. The prompt varies just in terms of subject, which is always up to the creative, so there isn't much more to say or the need to make the video even longer 😄🙏🏼
short and sweet, thanks for sharing Julie!
ofc 😌🙏🏼
Awesome, Julie! Love how effective this technique is, and how fast you walk through it! Well done and useful.
Additonal tip: Try out giving your characters long fantasy names, that already have some traits in the name. You will be suprised how good MJ can recreate consistent characters with just like 20-40 letter tokens. At least faces, together with your technique its very easy to also get cloth and accessoir consistency! Just what I was looking for. Thank you. 🙏
yes, I sometimes use that technique or use made up names for "in the style of" 😌 good additional tip!
Thanks for the tutorial, Julie!
Thanks Greg 🥹🙏🏼
This was SO incredibly helpful! I honestly hadn't thought yet of connecting Midjourney and Photoshop Generative Fill but now I have so many things I want to try 😄
they are a great power duo! :D
Sweet, a new approach on character consistency in midjourney, thank you for sharing 👍
Nice stuff Julie. I was not aware of this split image technique. Thank you!
thanks, I shared it a while back in march on twitter already but it's nice to have it in video form as well 😌🙏🏼
I really love the narration it's very soothing and thank you for the tutorial it was quite insightful
Thanks so much-this is super helpful!
Fantastic technique. Thank you for sharing!
Excellent. Thanks so much
Nice work! Photoshop really provides the iceing on the cake!
Thank you very much. Looking forward to your next video.
Brilliant.
Awesome tips! Thank you! I appreciate that you are so minimalist in your approach to tutorials. Better watch it, people are going to clone your voice for ASMR!
thank you :)
the voice is AI 😂🙏🏼 it's elevenlabs!
@@JulieWDesign OMG. That's hilarious. Awesome. First time I was fooled!
And that ID pic is AI, too. You're really a 60 year old Russian male.
great, clear guidance here. But I cannot find information anywhere on how to use Midjourney to place an AI model within a scene I have captured in a real photograph! Any clues - tips?
Thx!!
Will try this - thank you!!
Do this technique can be use in playgroundai? Since i only use that 😮
try it out ;) I haven't tested it, only in MJ
I don't know why but I don't get the split view
Amazing! 🙌🏻⚡️👏🏻💯
Wow! This is a good trick. It worked for me but after lot of patience. Sharing and tagging you on Instagram. Thanks a lot! This is super helpful.
told ya, patience is key but it will serve you well 😬
Fantastic Video - I was tasked by a relative with taking a real child's face and bringing it into Midjourney in various poses, costumes, and environments. I can see how this would work with a character is based on one done in Midjourney but if the character's face needed to be based on a real person can you tell me what your approach would be?
you could try working with a reference image (image link url in prompt), sometimes it gets pretty close to it, but it's trial and error! otherwise there's an option of using the Insight FaceSwap bot, check out this video: ruclips.net/video/TAlQJYyTiNw/видео.htmlfeature=shared
hi! I love a lot of this technique though I'm finding it has trouble with really unique faces or what some might call 'ugly' or 'average' is there a way around this?
you can play around with random names you make up that aren't related to any known personas/actors, also adding words like "imperfect skin", "asymmetric face" etc - sad we'd have to do that but it often works. and don't use words like "model"
Hi Julie. Thanks for the video. I also use PSD beta but not the way you did here but I will from now on. Question: What about generating with PSD beta and then importing back into MidJourney. Is their a way to keep the image consistent say with image weight or stylize to zero or something?
there isn't a way (to my knowledge) to feed it back to midjourney and keep all the same aspects of the image.
however you can work with image links in the prompt - that's how most consistent character videos actually do it. in my experience, the faces turn pretty blurry/plastic-y, hence why I found my own solution to that problem back in march and it served me well ever since 😄
Can you do this with multiple subjects or group of people? For example I want to generate a group of soldiers marching in the forest in midjourney. Do you think it will work?
try it out :D
I found that multiple people are very hit or miss, but I never tried out your subject/theme, so it could maybe work
Thanks!!!!
Hi, Can I know when your Midjourney Ebook release? How long it take to publish. 🙂
🎉
🫶🏼
Mage space is pretty easy to keep same character like my avatar I made 100s of this character
Can u use ur own picture?
yeah !ta shi yi ge hao banfa! hahaha
You're under arrest for using light themed Discord
I can't work in dark mode on desktop 😂🤌🏼 mobile, different story!
Next step would be how to change clothing and acessories
easy: vary region aka inpainting in midjourney :)
Interesting but in the future videos I´d like to see those prompts better and maybe a little bit more explained too. For example the Image split into 2 seems to be important. What is the vertical line in the prompt? or is it there. Sometimes it is and sometimes it´s not. This is too rushed and again those prompts should be better seen by viewer.
But othervise very good. Just needs a lot of pause-ing by the viewer. Make these longer. I think we don´t mind.
thanks for the feedback - fair enough, working on my computer everything seems to be big enough and even on my phone it was pretty legible, but good input!
Yes, the image split is important since that's the reason why you get two poses of the "same character" - where midjourney draws that 'line' is up to the tool, there's no way to prompt that it needs to be in the middle, but it usually just is.
The prompt varies just in terms of subject, which is always up to the creative, so there isn't much more to say or the need to make the video even longer 😄🙏🏼
Cool! An ass dat deng richteg Stëmm!?
nee, elevenlabs deskeier :D
@@JulieWDesign haat ech mer geduecht mee kléngt super realistesch!
I'm sorry to tell that this tip seems to have become less stable with MJ 6.0
It still works pretty well, but you gotta adjust a few things on the structure of the prompt - I'll do an update as soon as v6 is out of Alpha
Are You American?
nope