We all know that MidJourney has improved with faces, but hands are still a problem. Moreover, sometimes we might want a specific element in our scene, and this is a great free trick to get it. I show a very simple process to bring the assets into Dalle-2 and generate the missing hands, fingers, or in this case, a helmet. You can use this trick to easily replace items missing on your MidJourney scenes and do it for free since Dalle-2 currently offers 10 credits a month for free.
#photoshop #photobashing #livestream #midjourney #AIArt #dalle2 #dalle
Join my Discord server, where we create ideas together!
https://discord.gg/vy8r35c
New! Become a member to get exclusive access to perks!
https://www.youtube.com/channel/UC9kC4zCxE-i-g4GnB3KhWpA/join
Enjoy this video? Consider buying me a coffee! https://ko-fi.com/sedetweiler
Twitter: https://twitter.com/sedetweiler
Instagram: https://www.instagram.com/sedetweiler/
Nice. Gonna start messing with midjourney today for the first time
OMG I never even thought of that, that is genius! 😮
if we pay 30$ do we have to pay again next month?
Pffft, the solution isn't available to us plebes yet.
I'm still on the waiting list for Dalle-2 but I gotta try this out once I have access. Great tip!
genius
There's a much easier way to do it. Just drag the small one into the larger one – ctrl select both images – go to EDIT and choose AUTO ALIGN LAYERS – then click OK. It does a great job of realigning the two photos.
I just signed up for MidJourney after seeing your videos. One of the most interesting tools I've seen yet. We are definitely entering a new age of art and media production.
It always seems to be the art critics that have issues with new tools and methods. They want things to stay the same and claim anything new is cheating. As if we are playing some game.
But the true artist has always been the inventor or explorer of expanding their abilities and finding new ways to express ideas. It's about the final result. Never about the process. A brush, digital or otherwise, is just a tool and always will be.
The camera was seen as a so-called "destroyer of art" by many when it became popular. And we know how false that was. Photography became its own art form.
From what I can tell, AI tools require a lot more setup and adjustment than any camera that I know of. And after all that, we still have 50% of the work to do by hand to complete our idea.
It's true that maybe some of our time is saved, but the more time we gain on the front, we can shift to the back and get a more accurate result to express our ideas. Which is only natural for any artist.
I suspect AI artists will split into their own category as photography eventually did. Both of which still require an artistic eye where the fundamentals remain the same.
Been on the dalle 2 waitlist for 2 months. :/
Been on the dall-e list for almost 3 months!!
Huh. I was planning on using dall e mostly for outpainting where I want more past an edge. This is a thing I hadn't really considered.
You have to pay for Dall-E too right? Or can you use the free version to just fix hands and stuff?
Great work Scott! What's that water droplets background you have? Is it a live wallpaper or something?
A Midjourney tutorial that requires dalle even though it isnt yet publicly available where im from
Are you seriously telling me that the only solution is to subscribe to another AI that isn’t available to the public yet just so I can make a portrait of Angelina Jolie in the style of Marie Antoinette, photorealistic, cinematic lighting, octane render, 8k —ar 9:16 —s 5000 … I think I’ve been on midjourney too long
With photobashing and midjourney u want actually have to have the extensive knowledge to paint or draw accurately yo do concept art. I could see concept artist using this in their work then painting in detail which I'd already done through photobashing. However things like storyboarding or visdev can not be created by a.i as it has to follow a story beat frame by frame which can not be automated by a computer.
have yo done a video about how to work with size? i think that's about the only thing im having problems with that and wording, but i can all ways look at other pic. for that
How are most of you making money with your AI + modified artworks? (or just AI art produced alone) Care to share?
Hey, u have any video about –q 2 –q 3 –s variations? If not please make some tutorial video about it, thx
Hey Scott, IMO there's a lack and the need for more guides for people trying to get great results out of Midjourney, so keep these coming. It can be tough to get what you want at first, with the tool being so open ended, so any advice you can share with newcomers is gold.
Thanks for this video. I keep missing your streams so hopefully one day I'll be there to see it
Thank you this technique of combining the results of the Ai outputs works surprisingly well.
Dude this was incredible, great work!
How long to wait for dalle 2 account i want to try this 🙁
Is there other app to do this? or just Dalle 2?
We gotta come up with a better term than "photobashing." It sounds like you are bashing on people's photos lol.
"Today we are reviewing your photos and tearing them apart. Welcome to photobashing!"
Anyone else finds this cringe?
Otherwise I really enjoy your videos.
Thank you for those videos, your channel has become kind of a newsletter on AI 😉
That panting feature of Dall-e is sick!
I've been wanting that exact feature for Midjourney for when you have a picture that is 80% there but not quite.
Scott, maybe going into best practices for MJ. You did discuss /fast and using up your "fast hours. I didn't realize I used 7.7 of my 15 fast hours screwing around upsizing + –q 2. I switched to relax. What happens when you use up your fast hours? I don't think you can Upsize to Max in /relax? So, your –q level also eats up your fast hours. I think I wasted a lot of my time in one session. Thanks. Question: so if you are /relax but you use –q 2 you are using fast hours? Thanks
Great video and that's exactly what I've been doing on my prompts!! Thank you for sharing this!
and then to become a member of Dalle-2, with the right opportunities that you show here, which allow you to do this, but in reality they write – you are still on the waiting list for Dalle-2. Was it difficult to write right away? not to watch a useless video.
On midjourney it's like the more you upscale the more the body parts start to look off. Eyes/mouth start to move out of position. It put me off midjourney.
Ye I noticed that MJ is kinda bad sometimes to understand the prompt too. I was not testing hands, but another area it was lacking is props, like "pirate themed props on a table" gave me a pirate ship in a sunset. Good looking image, absolutely nothing about the actual prompt.. I like Dall-e generations, but their pricing is horrible.
I think that offer is only open to either people from a certain country or to people who got an invite. I'm in Canada and it really only lets me apply for the DALL.E 2 waitlist
Great stuff man, will join the waitlist for DALL E!
Tencent restoration tool can help a bit too, especially with eyes and noses (provided the demonic deformities from mid journey are not so severe)
dalle2 still has a waiting list, for this reason mid journey is way better lol
Awesome tip! thank you!
I think considering how fast MJ generates, I'd be willing to put in a few minutes pasting a google searched helmet in Photoshop.
Cannot wait for MJ to implement that technique! Humans in general are a struggle and one of it's major issues.
This is great.
Personally, I don't think that I'd go through this much effort to get a Midjourney image the way I want it, but I love that you made this guide and I think if I was ever desperate enough this video will be a great reference to come back to. Definitely keep these coming!
but how to use that to fix hands tho? I'm still struggling
This was excellent mate
how can you do this without photoshop
Sucks that we're in the grift stage of ai art. Eventually a free ai art program will become just as good as the extortionate ones, but until then