Hey Tim, 'Retexture' is awesome at turning your photos to a heavy textured oil painting or watercolor, using your subject as a character reference, just in one click. Now got to find all my old pics that need a makeover. Have a good weekend. Steve
Paul McCartney! 🤣🤣 The restructure one is blowing my mind. They _needed_ to give me (and others I'm sure) something to show traction before v7, because I was almost off. Thanks for the video...Paul! 😆
hey, Tim! Good, in-depth run-through of the features. At the very beginning, you show a pencil drawing of a dragon being "rendered" in various, more realistic styles. What feature or engine is doing that? I want to take my mediocre drawings and have them photo-styled! _LOL_
That's the Restyle feature-- I cover that a bit in the back half of the video. I know, I meant to actually try that with some of my 10 y/o kid's drawings, but they said No. haha, typical artist.
Well, it has a lot of potential. What might be a great feature is having like 3 layers in the editor, so that you can scale and compose different elements, like people or objects, in one picture. I mean bringing in a second or third picture on top of the main picture and then erase the parts you don't need... and then scale and place the subject to how you want it in the picture.
Layers are a thing I’ve been yelling from the rooftops for awhile now about! ALL of them need layers. There are a few that offer it, but I honestly don’t know why it isn’t a day one feature
Love your content I noticed you just did a kaiber video On the new studio did you know that they have just taken on runway and a few others If you didn't thought that might be some good content for you cheers
I don't think you got it right with the requirements. It seems that you need all three to qualify *People with yearly memberships (don't have this) *Members who have been subscribed for the past 12 months (got this) *Members who have made at least 10,000 images (got this)
Hmmm- that's odd, I'm monthly and seem to have gotten it. And trust me, I have no "insider" connection to the MJ gang. They're weirdly black box in that respect. But yeah, I'm right where you are...That's odd. Check w/ the Discord?
Hey, Tim! I was wondering how you got those images at 14:22 with the model being retexured multiple times? They are really cool & I would love to do something similar with my work :) Thanks!
Haha, the actor/musician we never knew we needed! Hmmm, I wonder what we can do w/ Suno, ElevenLabs, and a little AI Video magic? I think we need to bring this character to life!
WHy no one is talking about Krea new killer update where they basically brang all 4 big names of the Video Gen (Runway, Minimax, Pika, Kling) into their website ? While keeping their upscaler/ and Flux ? Is there a catch ? Like older versions of every models ? Curious to see a video about it ! Maybe compared versions with the "official" site model
I'm kind of disappointed because the one thing it says it can do supposedly is sref. But it seems that it can only do that based on a image reference and not an sref #, of which I use all of the time. If I'm missing something, please let me know. Keep up the good work, Tim!
Yeah, the way I look at it is that this saves a trip to an external editor. You almost always have to do little things in PS or another editor from that initial MJ output. This...I don't know, saves us a little time? I'm all for that!
Hello Tim Great Video do not seem to you that you really look like Makartny's floor? I joked. All these changes are like the wrong ones that have long been available in Lainardo?
I tried the editor, the faces in the generated elements were always quite distorted, maybe it’s just me, but I’m a bit disappointed with Midjourney for group photos and hands and extra legs and arms seems to have taken some backwards steps, I’ll maybe do some digging see if there are some settings I could use to improve outputs
Is it now possible with in paint to add several consistent character with using a character ref in the in paint prompt? I have struggled with this in the "normal" in paint editor, so it would be really cool if it works. Another things is that in paint (for me at least) have been really sketchy - often when prompting for an item to be added just filled in the background (like a removal tool). Anyone else seen this?
I've been playing with it all day, lets face it, it's currently dreadful. They would have been better off, regenerating the source image first, then tweaking the inpainted area from their own image.
Niji is a low key secret weapon in MJ! At some point I should probably do a Niji for Non-Niji users video. It's kind of a hidden gem! Kind of a slow autumn day here, might be good to break out the All Things Must Pass record!
@@TheoreticallyMedia I would really look forward to that Niji video. As I write stories and Niji is amazing to get them illustrated, just waiting on a few more updates to get a yearly subscription. Thank you so much, I really appreciate your efforts and channel.
Nice video! would like to understand if, by uploading a real image-say, depicting a person-and only changing the background or a part that doesn’t involve the person’s face, once the software applies the requested modification, the face remains identical to the original or, in the case of an upscale, it ends up looking a bit different or rubbery, as often happens with Midjourney AI generated images...🤔
nope! It actually leaves your original face totally intact! Now, if you run that "suggest prompt" it might call you Paul McCartney (that's in the video haha), but it keeps the same face!
I always look forward to your videos. I need a new hobby lol
14 часов назад
Not really there yet: 500 times of 1 time the editor strikes prompts like, man on a beach, man in New York etc. Ideogram has this feature better integrated if you want to become creative.
I'm going to try to block out some time for Ideogram next week! I love them too! I think MJ and Ideogram might make for a really powerful stack! And yeah, MJ's editor is...well, it's Day One. But, it's a long overdue feature for them, but everything about it (so far) seems to be on the right path. I just want that upscaler to take it up a notch!
Yes well, me too then.. I use light mode, I think, only the psychopaths use the dark mode.. 🤣 I always learned that if you put your images in a dark frame, you are trying to hide something.. but hey it was art school in the 90's before internet 😂
They're busy w/ Firefly. And to honest, from everything I've seen, I don't think MJ has any interest in selling. No doubt that David (MJ CEO) is pretty well off at this point, and from what I've gathered, I don't think he's interested in joining that Elon/Zuck level of money. ...one of the reasons I like him.
Their moderation and censoring is still too heavy-handed to make MJ a viable tool. I'm not a free speech absolutist either or interested in NSFW - however MJ's so afraid of any attempt made for adult material that so many innocuous words trigger the filter that using their AI becomes a chore and frustrating - rather chill and relaxing experience.
Yeah, I hear you. It’s funny, there are often ways around it. I once needed a bunch of dead zombies laying in the street and it wouldn’t give it to me. So I changed it to “sleeping zombies” and MJ was like: oh, ok, cool. Haha. Just gotta get creative to trick it.
Midjourney is going down the drain fast: Everything shown in this video is already available with Flux and comfyUI / Invoke for free(and from the looks of it, way better implemented). Midjourney still can't do proper hands and eyes or write text properly. Image to 3D objects is already available on multiple sites. They are like 6 months - 1 year behind everyone else and all their future projects have all already been done succesfully elsewhere. Their version 7 better be the best thing we have ever seen or I think they are finished.
Why do these things still feel like its a slot machine? You put in your tokens and pull the lever and nope. You put in a better description and WTF is that? So maybe fill up the "Donts" catagory, maybe put a lil "English" on the lever and your dragged further and further away and the negative prompts has just given it ideas about all the crap it can put in there. Its gonna make what it makes and your input has nothing to do with it. Its three years later and ya lost $200 billion and the wife has questions. You tell her it finally does speech bubbles and it wil put real words in there. Yes Honey, its still gonna misspell everything, Its smarter than us and is too smart to learn how to spell stuff. Yup, its 3 years and about $200 billion per model, art stolen from 6 billion people and we can make 3 second long movies or slow it down and make a standard 5 second long clip and none of it is what you wanted. I think with all the hype it will make it another year before we call it a pump and dump scheme. The stock will keep on rising but you should cash out early cause the majority wont be able to cash out at all after Bill Gates dumps and leaves the country
Chicken Dinner to you!! I was pretty late to the MJ Editor video party-- but I really wanted to cover this one, despite the fact that it was a super hectic day...but hey: CHICKEN DINNER!
Oh, look-- for sure there are tools out there that are better than MJ's Editor. BUT-- let's also give them some breathing room, as this is like, Day One. Also, I'm still one that thinks you shouldn't stay on ONE platform. Generate in one place then kitbash it with other tools. Still, the closer you can get to your vision with that initial spot, the more time you'll save later. So-- I don't think of this as being a "be-all" aspect of MJ at all.
@@TheoreticallyMedia The main thing for me it's, if that new tool it's not ready or not as good as a FREE tool (Fooocus/SDXL), Why not release it as a FREE BETA TESTING first and then start to charge money to people after they do a better job than the FREE one. That's all. For example, Runway and Kling are paid because there are no free ai video gen that do that level of quality right now, and that it's totally fair.