MidJourney Levels Up With This Major Upgrade
ฝัง
- เผยแพร่เมื่อ 14 ก.ค. 2024
- In this video I discuss some of the cool new abilities we have with the MidJourney 5.2 update. I also include tutorials on how to get multiple characters into a single scene, how to create a consistent character in different environments, as well as how to do the "infinite zoom" effect.
Discover More From Me:
🛠️ Explore hundreds of AI Tools: futuretools.io/
📰 Weekly Newsletter: www.futuretools.io/newsletter
😊 Discord Community: futuretools.io/discord
🐤 Follow me on Twitter: / mreflow
🐺 My personal blog: mattwolfe.com/
🌯 Buy me a burrito: ko-fi.com/mattwolfe
Resources From Today's Video:
www.midjourney.com/
/ 1672030749250277376
/ 1672025845613961216
/ 1672254705647452160
/ 1672315823631769611
/ 1672055696953573376
/ 1672234966124761090
/ 1672720456024043521
/ 1672593043252207616
/ 1672470475522584576
runwayml.com/
Outro music generated by Mubert mubert.com/render
Some links in this description are likely affiliate links. I do not talk about products based on whether or not they have affiliate links, however, if a product has an affiliate link, I usually sign-up for it. It never sways my opinion or impacts how I talk about a product.
#AINews #AITools #generativeart
0:00 Intro
0:49 Some of My Favorite MidJourney 5.2 Art
2:29 The Hulk Plot Twist
3:03 Cool "Infinite Zooms"
3:58 Crazy AI Animation from The Dor Brothers
4:20 AI Shapeshifting Effect
4:59 Tutorial - Multiple Characters in a Scene
7:14 Tutorial - Consistent Character Morph Effect
11:07 Tutorial - The Infinite Zoom Effect
16:15 My Thoughts and Future Tools - วิทยาศาสตร์และเทคโนโลยี
Still hoping it'll someday be easier to have the exact same character in different poses, situations and images.
It will be.
My guess is (because I've seen somebody talk about a paper about it) by the end of next year, we will start to get the ability to choose the camera position and choose what it is targeting. The objects in the picture will not be 3D objects, but their context will be remembered, which gives you a sort of fake 3D object scene.
At that point things like MidJourney will actually be in the movie making stage.
Technically it's all possible. StableDiffusion + ControlNet has shown it is now possible to do this. It's only a matter of time until Midjourney integrate this into their suite
5.2 is exceptional and is what 5.0 should have been. It’s the proper evolution of 4.0. All of my images now look bad and I need to make a whole load of new ones 😅
We should ask for refunds for previous images created with their primitive versions :P
Thanks for including me in this video 1:48 great selection of work form twitter
Absolutely fantastic video! The tutorials are clear, concise, and incredibly helpful for understanding the new features in the MidJourney 5.2 update. Keep up the great work, your content is truly enriching the AI and art communities!
Once I saw what this zoom thing can do for use in painting references, I am stoked to create entire worlds for my subjects. When I figured out you can also change JUST the zoom prompt, I was completely blown away. It’s wild.
I was trying (with no success) to figure out the infinite zoom thing. So glad you took the effort to explain it. Thanks!
This was amazing! Thank you for explaining these new MJ features! Looking forward to to receiving the newsletters!
Keep the updates coming! I appreciate you.
Amazing video no one else demoed the ways to use the new features in cool ways. They show the basic use case. Thanks so much
What a time to be alive, glad to take this trip with you Matt, thank you! 🙏
This is awesome! thank you for sharing these features about MidJourney 5.2!
Thank you so much Matt always look forward to ur videos and tutorials so.insojring and.very helpful straight forward and clear ❤
Thanks for the character morph tutorial. I used it to create something for my school with a consistent teenager's face morphing through a series of occupations. I just changed the prompt in custom zoom to "an architect" or "a doctor". Turned out nicely thanks to your tutorial.
Thank you so much for these tutorials and the easy explanations.
Excellent Tutorial Matt. Well done and very engaging. Love the new features. Off to try them for myself.
I'm waiting for them to fix the text generation issue. Then it will be truly useful haha
Yeah, awesomelooking infographics from the getgo will be another (small but substential) gamechanger IMO.
There are rumors that the 6.0 version will be able to generate coherent text on images.
There’s already models that do this somewhat consistently but nowhere near mjs quality sadly
It's not a issue tho midjourney was never made for text it's not a broken feature it's just not something it does maybe in the future there's talk of more understanding for text but I wouldn't hold my breath there's more interest stuff being worked on
I didn't know about custom zoom. LOVE IT! Thanks Matt!
Thank you so much Matt. It's absolutely amazing and I really happy that you share such valuable tricks. And you give clear manual 😊🙏
The best tutorials you can find, always with Matt.
You rock dude 😎
YES!!! I've been waiting for this vid to drop! YOU SIR ARE A SUPER STAR ❤❤❤❤
Great content as always, Matt! 👍
Cool stuff. Thanx for showing it off!
Your channel is pure 🔥🔥🔥 Learning a lot from you. Gollowing you since 100k. Keep going!
Fantastic indeed! Thank you for this tutorial, this infinite zoom effect looks very awesome and usable!
Excellent video! thanks a lot Matt!
Loving your content - learning so much - thank you ))
Great video. Thanks for showing some cool ways to create all of this.
awesome! thx for sharing these techniques!
Thanks for the video! Great info!
Thank you so much matt!!!
Gracias Matt! Estaba esperando al dia de hoy para poder ver tu nuevo video. Sos un genio!
Brilliant! Thank you!
Great video thank you so much!!
There, infinite love for your tutorials and work! 💜
Matt thank you for taking the time to show us all this amazing stuff!!!
Any new video on this channel makes my day. I love Midjourney and have generated thousands of images but I gave it up because I just have no reason to use it. Wish I did because it's fantastic at what it does and getting better every month.
This video was soooooo goooood! THANK YOU FOR THIS! ❤️💜💙
Brilliant Matt - thanks!
Thanks a lot for sharing!
❤ love your videos, thanks for your help
Simply awesome. Editing features expectations
Very cool images! I´ll be zooming in on this! 👍👍Creating stories is what life is all about! 📜 Back to the flip book trick almost
Nice work Brother! Love the passion and appreciate you keeping me up-to-date with the torrent of technology.
Thank you for all your effort
Great great great !! Thank you so much for your videos ! :)
Great videos and info!!
Thank you. This is very interesting.
incredible thank you bro very useful
Incredible, thank you
I saw those images when i was in Midjourney the other day! Super cool
Now that's value, thank you man
Damn i've learn so much thank you!
Thanks for the updates Matt, really enjoying the updates to MJ. I did try DaVinci, for whatever reason the result wasn't what I was expecting so went back to Adobe AE, a lot cleaner but you can still see a very slight stutter between frames - struggling to get that buttery smooth transition, but that's part of the journey :)
I have the same problem
Again, very informative cool video!🎉
Amazing! Love it!
Thanks for the DaVinci tips! This is so helpful. 👍🏼
I've already generated a stream of pictures was going to use capcut on my phone, bit that divichi look pretty good. Appreciate the info. We are all learning together wish I could have a chat with ya. Have a good one!
For upscaling, I use Upscayl- it's open source freeware, does excellent upscaling, is very fast, and can also do bulk.
Great tutorial !! Tkssssssss
Great Video, thanks
At around 6:40 when you get Martha Stewart mixed with Snoop Dogg, I've found that you can take Snoop Dogg out of the prompt entirely once you have him the way you want him. Or whomever or whatever one is using in place of Snoop Doog LOL
Nice one Matt. On it like usual
Thank you sir!
Thanks for sharing this Matt, v5.2 is crazy, and not sure we will be ready for 6.0 and the web interface.
After watching your video yesterday I made a music video for one of our songs with the v5.2 Zoom technique and a few LeiaPix renders. Finished the edit this morning and sent it to some creative friends - it blew them away and they couldn't figure out how I'd done it. Been trying to do something like this in Kaiber for months. Thanks for the info!
What's the link so we can watch it
@@chris-ib2lc Thanks for the interest. I tried adding the link twice but TH-cam deleted it since it takes you off their site.
Ty, again!!
BEST how to video. thanks Matt
This is so orsum Matt. Love your content🤙🔥
Great job!
fantastic !!
You did it again you mad man !
2 minutes in and already know this is a quality video, so much information
Another incredible video
What a great video tutorial, Matt. Besides making cool-looking videos we can post on social, what kinds of business applications do you see this being used for? Games come to mind. What else? Thank you again and keep up the amazing work.
you're the best dude. love ur videos
Really awesome! I find that you've made great progress compared to your works from half a year ago.
Man I even learn Davinci Resolve things here, truly the channel that keeps giving!
Oh my!!! Imma hop on my midjourney now!!
What I noticed after getting spoon fed AI news since March, is that:
Real innovations usually come out towards the end of the month, or every 2 months for really big announcements.
Middle of the month is regular updates, light improvements.
Beginning of the month is pretty regular.
Oh and open source….well, new sht just keeps popping out.
Good thing is, when there is a big announcement, you won’t need to try to guess if it’s a big one, it’ll be all over your feed with appropriate titles.
And Matt really found how to improve his news videos by adding tutorials, kudos for that, makes my morning coffee tastier !!
Exciting! Maybe I’ll start the subscription again
Thanks for the shoutout ❤ Great vid!
Amazingness! 💯🎉
Hi Matt. Thanks for your great videos. My turn to make a suggestion: for free and local up-scaling (including in batch and up to 16X), get Upscayl !
Great tutorial! The first thing I thought about when I saw this new Zoom Out feature was, can we make a video with this? Thanks, Matt!
One of your best!
DaVinci Resolve should actually pay you for the advertisement, because you just soled it to over 381K of your subscribers! I spend 2 day of researching the newest TH-cam videos to know how to do these things, only to realize that you posted it 12 days ago, and if I had caught up with you sooner, I would have known it right away - bravo! I will NEVER spend my time watching any other AI/Midjourney content creators before you. Thanks ^^
Matt is an absolute pro...hands down!👌
Any chance you have a follow-up video on how to keep MidJourney from Frankenstein-ing subjects when you zoom out too far? I literally had a generation that redid a body as a giant sneaker??
Truly appreciate all the information you are presenting. It's amazing you can keep up with it all.
It's unreal what v5.2 has done. The --shorten feature is great as well. They're going to perfect these features and add more by v6. It's powerful and impressive. Thanks for the video. 💯
Great video
i needed that zoom tutorial for davinci ty
AWESOME!!!!!
I feel like this could also be used to zoom in in a similar way, I mean we are talking about Recombination group theory here, which technically I only learned about like, an hour ago, but theoretically you could do that
Bro i just wanna say thank you... 👍👍👍👍
You are awesome ❤
Thanks for the great video! I knew the minute the "zoom out" tool was introduced, that this is what it meant. I was already creating "zoom out" stills for Resolve 18. I was running up against problems in Resolve though, and your quick tips on the program really helped! The thing that bothers me now, though, is, you would think Midjourney would assign consecutive numbers to each zoom, and I assumed that was the case... But when I bring the sequence into Resolve, and drag them onto the timeline, they aren't in the correct order. Depending on the subject, it can be hard to re-align them visually. I now have to take the time to save each image with consecutive numbers (001,002,etc.). This works, but really slows down the process... Am I missing something here?
DragYourGAN (DragGAN) can reposition faces really nicely. It's on hugging face.
Thanks Matt! Gonna do all of these! You da man🥳🤩🦾
This is awesome
This update is huge. I've been begging for outpainting for months (and millions of others probably, too 😉). It's still a bit complicated with the resolution restrictions, but I'm having an absolutle blast with it.