Shocking over year later theres no control net, or outline identify ??? Leonardo, openart ai, fooocus, stable diff... all implemented these 12 months ago. Still gonna be jumping thru hoops in 2025 😢
Incredible video, tips and insights. I've been struggling with multiple consistent characters and have tried countless approaches. Your process mnakes more sense than any of the others. Can't wait to give it a try tonight. Bravo!
I am so glad you appreciate this one- there's a lot to cover in order to make it all click. I'm glad to hear you feel like it helped connect things for you. Good luck generating 😃
Yeah, you truly get a sense of control over the composition! It's kind of amazing how big a difference literally being able to crop things out of your generations make.
@@glibatree "back in the day" when I was a monthly sub, I had to pan left, then pan square, pan up, square, ad infinitum until I got the right proportions, lol. This looks so free-flowing and wonderful in comparison, and the masking looks much much better too.
That was such a wonderful video. I am by no means a MJ novice, and I work with it a lot, but you gave some new tips and creative ways to get good results I have never considered. Good job. Btw, I sometime use objects as CREF too to get similiar-ish objects in different images.
i really appreciate the knowledge you sharwed with us today. Very helpful and easy to follow. Also thanks for the GPT link. tried making my own but isn't as detailed and specific as yours.
Great work I tried the GPT link but only giving me five panels and the characters are inconsistent. Some of them even put in a female even though I asked for a male
Incredibly useful ❤thank you for sharing this. I subscribed to your channel. Will definitely try out all your methods and learn. However, could you please make a slightly detailed video on generating a character at different ages with precision. That bit went a little fast or maybe I'm just too slow 😅 Thanks in advance if you decide to make a tutorial on that bit. ✨ Also how to show a real life character at different ages? Any tips would be much appreciated.
Fantastic tips and vid! I have a client who needs me to recreate her for talking on stage realistic images of her life coaching but having issues finding rt program and steps for this based on a dozen of her headshots and photos I have for reference
Hope this ends up helping now, or for another future project! Midjourney might not be perfect if you're shooting for a true match- you might do better with DreamBooth or other fine-tuning approaches for a SD model. Good luck with your client, it sounds like a cool project 😃
Midjourney devs' official answer is no. They try to warp photo-real images of people to not match - but if you standardize your inputs and do what is talked about in this video regarding a prompt framework that works well - I find it pretty possible. Here's the video from when I did it: th-cam.com/video/KwAwWjCRaCY/w-d-xo.html
Thanks for your video brother--- handsome knowledge you have....but I think its better if consistency of surrounding and character should get introduced in the technology itself ....because at present it is a very hard working and trial error methodology we are left with ...and rest because most of the platform are paid so if we are allowed 100 images in our package only 20 fits the bill
That's a fair point, I like experimenting and seeing what's possible with what we have but- something like cref and bref (for background reference) could be really useful to a lot of people.
you are offering valuable information and from what I can tell are quite knowledgable but the music is competing with you due to poorly being mixed. I'm downloading your video and stripping the music away in order to better hear what you're saying.
Thanks for the feedback, truly. I'm still learning to get better at the music aspect of exciting. I think it can do a lot for the energy and pace of the video - but it's definitely a balancing act 😆
This is precisely what I was looking 💪🏼 thanks for the hard work.
I tried to put in all the answers to some of the hardest problems in Midjourney.. I think the video should really drill down into those essentials.
Shocking over year later theres no control net, or outline identify ??? Leonardo, openart ai, fooocus, stable diff... all implemented these 12 months ago.
Still gonna be jumping thru hoops in 2025 😢
Incredible video, tips and insights. I've been struggling with multiple consistent characters and have tried countless approaches. Your process mnakes more sense than any of the others. Can't wait to give it a try tonight. Bravo!
I am so glad you appreciate this one- there's a lot to cover in order to make it all click. I'm glad to hear you feel like it helped connect things for you.
Good luck generating 😃
Wow. Just wow. I appreciate your hard work and can’t thank you enough for the value you provided in this video. 👊🏻
Wow. I took a few month break from midjourney and it is crazy how much it looks like they improved the editor…
Yeah, you truly get a sense of control over the composition! It's kind of amazing how big a difference literally being able to crop things out of your generations make.
@@glibatree "back in the day" when I was a monthly sub, I had to pan left, then pan square, pan up, square, ad infinitum until I got the right proportions, lol. This looks so free-flowing and wonderful in comparison, and the masking looks much much better too.
That was such a wonderful video.
I am by no means a MJ novice, and I work with it a lot, but you gave some new tips and creative ways to get good results I have never considered.
Good job.
Btw, I sometime use objects as CREF too to get similiar-ish objects in different images.
❤
Such great work, Thanx Mang!
Much appreciated! Glad you enjoyed it
i really appreciate the knowledge you sharwed with us today. Very helpful and easy to follow. Also thanks for the GPT link. tried making my own but isn't as detailed and specific as yours.
Awesome to hear it's working for you! The prompt framework, is hugely undervalued I think - and it can really level up people's creations
Great work I tried the GPT link but only giving me five panels and the characters are inconsistent. Some of them even put in a female even though I asked for a male
Incredibly useful ❤thank you for sharing this. I subscribed to your channel.
Will definitely try out all your methods and learn. However, could you please make a slightly detailed video on generating a character at different ages with precision. That bit went a little fast or maybe I'm just too slow 😅 Thanks in advance if you decide to make a tutorial on that bit. ✨ Also how to show a real life character at different ages? Any tips would be much appreciated.
Glad to hear it's connecting for you, and that's definitely something I can add to the pipeline.
Impressive. Thanks for sharing your workflows!
Thanks man! Hope you're able to utilize some of it in your own workflows!
Fantastic tips and vid! I have a client who needs me to recreate her for talking on stage realistic images of her life coaching but having issues finding rt program and steps for this based on a dozen of her headshots and photos I have for reference
Hope this ends up helping now, or for another future project! Midjourney might not be perfect if you're shooting for a true match- you might do better with DreamBooth or other fine-tuning approaches for a SD model. Good luck with your client, it sounds like a cool project 😃
Can this be applied to self pictures using CREF?
Midjourney devs' official answer is no. They try to warp photo-real images of people to not match - but if you standardize your inputs and do what is talked about in this video regarding a prompt framework that works well - I find it pretty possible.
Here's the video from when I did it:
th-cam.com/video/KwAwWjCRaCY/w-d-xo.html
That video is simply awesome!
Thank you!
Thanks for watching, I'm glad you enjoyed!!
you are awsome, thanks, you got a new subscriber :)
This guy got more hacks than an 80s slasher film.
Thanks for your video brother--- handsome knowledge you have....but I think its better if consistency of surrounding and character should get introduced in the technology itself ....because at present it is a very hard working and trial error methodology we are left with ...and rest because most of the platform are paid so if we are allowed 100 images in our package only 20 fits the bill
That's a fair point, I like experimenting and seeing what's possible with what we have but- something like cref and bref (for background reference) could be really useful to a lot of people.
@@glibatree yes definitely
👋 hi
👋👋
great video
Thanks!
super
Thanks!
Dziękujemy.
You're very kind thank you!
Jesteś bardzo miły, dziękuję!
@@glibatree You make my job easier. I thank you!
you are offering valuable information and from what I can tell are quite knowledgable but the music is competing with you due to poorly being mixed. I'm downloading your video and stripping the music away in order to better hear what you're saying.
jajajajja
Thanks for the feedback, truly. I'm still learning to get better at the music aspect of exciting. I think it can do a lot for the energy and pace of the video - but it's definitely a balancing act 😆