Do THIS to Create Amazing Poses in Midjourney!!!

[@christian.cb.bernhard]

Wow, best tip ever. This is a Game changer for my future projects. Let me share another tip: If you want a full body shot of a person, but MJ gives you a closeup, you can describe the shoes or legs in the prompt and MJ tries to get specific shoes or legs in the Image. And you'll get a full body shot.

[@rasserfrasser]

With regards to the ear shot, that's intuitively genius. What about a mood board using images of 90 degree shots via some kind of style reference? Then load that into the prompt for a character reference? Less editing, no?

[@damens-xn9pu]

Hey man as someone who just started on KZbin and is learning to tell stories, I think you do a great job with not just AI but storytelling as well

[@bigdada018]

Your hacks are phenomenal! Thank you!

[@cbeausoleil]

Finally!! This solves so many struggles I just gave up on. Now I can go back to some of my previous ideas and try again.

[@EZEonIon]

What a great tip! 👏👏

[@IversusAI]

This is PURE GENIUS. THANK YOU!!

[@MrAnthonyIII]

Thank you, this is one of the most helpful tips that I may never have thought off. I often have a hard time making characters in certain unusual poses or wearing unusual clothes, this helps give me so much more control.

[@murraymorison3924]

Really innovative. Clever thinking. Thank you.

[@kronos77777]

Great video. Really enjoy the outside the box thinking here.

[@maximumblackness]

Sounds like a great idea!

[@aikenpublishing]

Very cool way to build an image. As an artist, I can see first drawing out the image I want on paper, doing a rough sketch. Then find the area I need to generate first, as I begin to build the image one brick at a time, in MJ. Great technique.

[@jrbvdh]

Once again friend, you're a hero.

[@EmilyNilsen]

Great tip! This was a fantastic solution. I'm going to try this out. Thank you!

[@jesse8068]

Wow, this is really clever!

[@galefraney]

Great video, so refreshing to see something new, something that isn't covered by tons of other channels. You explanations are fantastic!

[@glibatree]

Awesome to see the effort is appreciated. Thanks for your kind words!

[@seantp3303]

So valuable. Thank you

[@jeremysomers2239]

well done, super creati e - appreciate you!!

[@CorporalDavis]

Gave it try worked really well.. thank you for tidbit of info

[@taoni8408]

Great tip! I like it - really unlocking the power of MJ

[@CatAndCloud]

This is priceless! It even works on the anthropomorphic animals :) I think I even quite close to make it work on the anthropomorphic cloud. Thank you! I would love to see your advice on how to work with the retexture feature: I often need to generate the same scene, but under different time of the day, e.g. whimsical house at night. I get very close resemblance, but not perfect: e.g. it changes shape of the windows or the shape of the path to the house. Kind of small details, but many and that makes it hard to fix in photoshop.

[@tasunke9555]

Pure genius!

[@kbd9765]

This is such a great trick. Thank you!

[@mnm.2015]

Dude, your channel is amazing! Great content, keep it up mate!

[@econ3000]

awesome video!!!

[@markrichards5630]

Many times you don't need to zoom out from it - just mention that thing and that thing will tend to "face" the camera (ears, bald spots, ponytails, open shirts, back pockets etc). Add in "side view profile" or "rear view" or "high angle" and you are golden

[@kevnar]

Imagine looking back in a few years, when AIs are way more advanced, and you no longer need to do all these work-arounds. Imagine being in an era where AIs just follow instructions and you can create your vision in a few generations with a couple of refinements.

[@zentai5076]

sooner than we may br ready for

[@IamSpidey007]

Just like we choose style we can choose camera angles

[@KreatedbyKrause]

So helpful! Thanks so much!

[@glibatree]

Thanks, glad you enjoyed it!

[@LorriEverly]

Truely brilliant! Thanks :)

[@Sebastian2018]

great tips

[@seanmaynard7417]

Awesome advise!

[@EveryDayLandMan]

Brilliant trick!

[@leodistefano7424]

Bravo!

[@rafaelohana5812]

Thats A very cool technique!!!! ❤

[@glibatree]

Thanks so much!!

[@vincenzosalvia]

This technique is truly brilliant. It's disheartening, however, how far behind Midjourney is compared to its competitors in terms of understanding the prompt. I also work in parallel with Ideogram and it doesn't miss any single comma. I hope they fix it soon because features like cref and sref are unique to Midjourney and allow incredible control of style and character.

[@OmniEblis]

Layering this idea atop the processes of conventional photo manipulation strategies seems helpful to me

[@TellMeSongs]

That is ridiculously smart. Well done, can't believe I didn't think of it earlier.

[@Mone-t9n]

Bravo bro

[@oemerka8908]

You are producing great content anyway 👌, would you have a chance to make another detailed video about the consistend character?

[@glibatree]

I have got a lot on the docket, but it definitely could happen- any specific goals you want to make sure I help accomplish?

[@PerFTNilsson]

Good one! Thanks!

[@zentai5076]

great tip! subbed for more

[@LASartor]

Me, too.

[@kaiforging]

excellent ,thank you

[@Mooooty]

great vid!

[@garthcorden1625]

Really clever idea. How did you get your website layout like that? The bigger image area is a winner!

[@KostasAspiotis]

Amazing, thanks a lot

[@kaiforging]

awesome

[@armondtanz]

U could use flux or ideogram, its really good at understanding precise prompts. Then use as image & style reference in mj and crank the --iw to 3. Sometimes it works. I'm gonna try this, great idea , and really cool hack. The worst is that dystopian image where the guy looks out over the decaying horizon😂😂😂, MJ loves that 1.

[@gustavo-brunofernando]

Чувство юмора у автора огонь :))

[@SonicDephect]

Such a great idea! Unfortunately the strict filter on MJ blocks my prompt in the editor because I often work with photos of women wearing boho attire that it may misconstrue as inappropriate...?

[@MrSzwierzak]

OK- I admit: this looks neat. Simple yet effective. I will test it now and come back to comment if it works on my astronauts.

[@glibatree]

Awesome! Good luck with the astronauts, it often surprises me how good the results are.

[@MrSzwierzak]

@@glibatree yup. its working even on astronaut. I started from tiny symbol, then I expended picture and symbol with great detail was on side of astronaut helmet exactly as I wanted. New trick in my toolkit. Thx!

[@heartshinemusic]

Fantastic video! Still hope that we'll get a better tool for referencing poses and camera angles in Midjourney. You got some amazing results with this technique, but I think it's hard to do this when you also want consistent characters, or do you have a work around for that too?

[@nugatron]

I'm working on a graphic novel style project, and I have some aliens, females, etc. That constantly get rejected in editor. Usually, for a denied image. Absolutely nothing nsfw, not even close. Even when mj created the image. Normal looking males are usually not an issue but aliens and females no dice! Any thoughts on how to work around?

[@alwaysinout]

For a beginner it would be great if you showed an example from A to Z, you said take a shot of the angle you like , then what !! what should i write in the prompt !! Thanks

[@paullenoue8173]

Damn, I wish I knew this months ago. Would have prevented a lot of frustration.

[@cocacolabeast]

I LOVE GLIBATREE

[@myrrhsolace5875]

That depth of field where everything except a small portion of the image is out of focus is actually a low depth of field, not a high depth of field. In a depth of field photo, everything from foreground to horizon is in sharp focus.

[@paulrossjones2234]

thanks for this. I have been doing similar techniques for a while trying my best to control the images created when using only midjourney created images. However, the problem I really want to solve- is how to do the same kind of thing with a supplied images- not a mid-journey generated one. I am a photographer- and I would really like to integrate these kinds of techniques with real photographs I have taken- or use my photographs within. I have had limited success with Leonardo- but mid journey is a far better generator. Do you know how this can be done?

[@EridanusYT]

Check the new Midjourney tool they added called "Retexture" and you can use cref with it too I think that's what you looking for and good luck.

[@paulrossjones2234]

@@EridanusYT Hi, thanks for the reply! Yes, I’ve been glued to my computer learning to use this new update- this is exactly what I’ve been waiting for, and it’s now an incredibly powerful tool.

[@ryanw5569]

I have struggled getting midjourney to create someone throwing a dart. I've tried to get it to produce it when the arm is bent at 90 degrees, the follow thru, anything. It never looks like a normal person throwing a dart (even the darts are weird, sometimes being a dartboard).

[@samsaraAI2025]

One question? And how is possible to combine this Editor trick at the same time that you upload a picture for character reference??

[@EridanusYT]

Check the new Midjourney tool they added 2 days ago it's called "Retexture" and you could use cref with it too.

[@oybrj]

For a long time I have tried to create frontal image of a SR-71 pilot in a yellow pilot suit, doing a closed, tight somersault, with his knees at he’s chest, and his arms around the legs. But it’s seems impossible to get Midjourney to understand how a somersault looks. I even bought a 3D model of the pilot and rendered a reference image in blender. It simply won’t work…. but got a lot of very strange images in the process. If anyone has any tips, please let me know 😂

[@TomHimanen]

Could you make a video about using consistent character in Runway (or other video2video)? It would be cool to be able to act the scenes on a video myself, replace myself with the MJ-created consistent character and replace the original background with virtual environment made using Midjourney. Is it possible?

[@glibatree]

Agree that'd be super cool. Unfortunately, right now, the best video models are text-to-video or image-to-video. But the video-to-video options haven't really caught up yet... So you can put characters from midjourney and scenes from midjourney into original ai generated video, but not yet replace characters or scenes in footage like your describing. At least at the quality of the latest Gen models that I'm sure you'd like to see them at.

[@TomHimanen]

@@glibatree I tried yesterday replacing the background of the video that drives movement, and got quite good control of the background and figure via prompt. So creating completely virtual character on conpletely virtual background with very natural motion workd very well in Gen 3 alpha. Unfortunately in very complex moves hands and face seem go glitch, but 100 % natural slower / less complex movement with virtual visuals is very much solved. Anyway, I was wondering if I could add consistent character in desired environment as first frame of reference video. Perhaps, if it matched the prompt, the video2video model would subliminally get an idea of continuing to imagine everything like the first frame. Gotta continue hacking it today after work,

[@glibatree]

Yes, that is pretty doable in my opinion. Consistent characters and selected, as a starting, frame can work great if it's well-prompted for. Which it sounds like your great at. Often I find there is a very quick loss in quality a few frames after the starting, especially if there is a lot of movement, but not so much that it ruins consistency. Still agree with you, that it is certainly the best way to get consistent character video at this time. Too bad on not having great footage replacement though, cause that'd be sick hahaha

[@TomHimanen]

@@glibatree Thanks for your insight, also thanks for ChatGPT assistant and great content generally. I will continue experimenting. I'll let you know if I have mentionable success. Keep hacking in the free world!

[@KimberlyLetsGo]

I want to be the woman in the orange dress with 5 fingers and 1 thumb. I could type so much faster! (4:53) If I have extra toes, do I have to pay a surplus price on my pedicure? (4:58) While I'm poking fun, this info is really helpful. TFS

[@heredownunder]

I always get side views and have to negative prompt --no side view

[@danielrodriguez-aleman8797]

do you get too often the stupid message : generate image fall outside our community guidelines? for creating normal images?

[@ThisFrameDoesntExist]

Great reverse engineering ❤

[@silversurfer4195]

Excellent, thank you. I think I still need to get the hang of the editor and when to delete or not something, but I managed to build a replica of the boy reaching out for gold, with the shadow of a dragon against the back wall. As you said, @glibatree, most generations were useless, but one of them-a Sherezade-like character-was very good. Great videos!

[@yolojc]

You should try just being better at art. Don't let the robots do all the work. AI c.r.a.p

[@glibatree]

Ok