I think that your ideas around Phrasing is spot on. The da-da-da is Phrasing. Poetic phrasing. Anyone who writes extensively - lyrics or poetry - and actors/performers who acutely understand language in that way in order to interpret it using voice/song recognise this immediately. It’s like a creative writing course and you are talking about the structure of language, breaking it down in to its parts. I’d be very interested to see you interview a skilled writer about the expressive sophistication we have to develop to effectively interface with MJ. Soon there will be full courses at universities on how to write for these creative interfaces! You are an amazing teacher.
@FutureTechPilot I respect you and am grateful for everything you’ve taught me. Language is the ultimate interface. Looking forward to your first interview!
Love it! I wrote a prompt as I was watching along with this video and ended up having the perfect dada, dada dada dada dada, daDA prompt written with all your instructions. Thank you so much for this and making it so easy to understand with that technique.
Hi. The best Channel about Midjourney in UA-cam. Let's hope it continues to grow and soon reaches 100,000 subscribers. From my side I will recommend it
Thank you for the Rhythm bit. It can be helpful to remember. Once again an excellent video demonstration. Don't listen to some weird comments who didn't like your rhythm. After all its your videos and you can do whatever you like in your videos. There are many people who are enjoy watching you. Don't pay attention to the negative comments. If they find it silly, then they are free to watch other videos. You are putting to much effort and knowledge into these videos. don't let some idiots change the way you are teaching to us. If they are smart enough they can record their own video tutorials and we are here to judge them too if they dare. Keep up the amazing professional work you are doing Nolal. 👍👍 By the way, can you create videos for other Ai generators as well? not exclusively for Midjourney. Such as for Dalle-3, Leonardo, etc...
haha all feedback is valuable! I'm happy to know when my videos don't connect with people - it let's me know that I can improve in that area. But as always, thank you for the support! And I have nothing against other ai art generators, but it will be hard to make videos on them for one reason - my current video 'to-do' list is so ridiculously long (well over 200) and it can be difficult to switch focus and prioritize other content over the ideas I already have scheduled. But I really do appreciate you asking, so I know it's important and maybe I can get to that idea sooner than expected!
AWESOME video! The fact that you approached it from a rhythm point of view is not just creative and rare but also pretty darned accurate. Once I started using the 'v6 language, ' it was insane what I could get. Clarinet is _so_ incredibly helpful, it's unreal. I had started looking at it as bookmarks with books in between. Keep adding to the books until it stops recognizing. Great video. Very, very helpful. I was beginning to have Nolan withdrawal! lol
haha I'm glad people are connecting with the video. I really wasn't sure what the reception would be like. As always, thanks for sharing your thoughts! I'm so thankful Clarinet and their team took the time to write those guides 🤝🤙
Terrific Tutorial ! This is extremely helpful and i super appreciate your generosity by sharing this formula! The rhythm approach is very unusual and original. 1000 thanks!
Thanks, again and well done. As I was watching your video I was visualizing a traditional vertical list format to learn the new prompting--like a grocery list with number two being the main course. Once again, MJ and FTP are making my brain think different.
Cool idea! I've always loved the analogy of food/ingredients/recipe to explain Midjourney prompting and I think your idea of a grocery list makes a lot of sense!
❤MJ releasing model personalization. Get at least 200 image ratings and then add --p after your prompt. Personalization completely changes the way system interprets your prompts.
Hey man, good to see you. Thought you got lost in the post ( pun intended ). You can definitely be counted on to do some heavy lifting; excellent recipe. You always make great sense and you have nailed it with this one. Thanks for your commitment...
[ INT. ARTISTS STUDIO -- NIGHT ] An easel sits in a messy studio. A nearby window lets in the NEON LIGHT from a sign outside. The floor is flecked with dried paint. ^^ he's basically describing screenwriting style.
Midjourney leapfrogs Back to its No1 position in the AI Art scene. Nolan remains at his No1 position of producing the most comprehensive Midjourney tutorials available on UA-cam 🙏👊🔥
Dear Nolan, Curious as I am when it comes to images, I was able to follow an introduction about MJ a month ago. And soon I was blown away by what I saw appear! I now have a basic plan with MJ and have already been able to make a few nice pictures, but...I have also become very eager and I want to know much, much more about MJ! How lucky I was that I saw your "Future Tech Pilot" fly by ;-) And I have already learned a lot from you. Your way of explaining the material appeals to me very much, you do it so pleasantly and full of enthusiasm, I really appreciate that and it is easy for me to follow!, (with some help from Google translate!,..:- )) Many thnx!!! from Amsterdam! Btw, I thought I heard you say something in one of your videos about a complete(?) workshop or a course that you could provide?,...
Hey thanks a lot for the kind words! Yes, I'm currently working on both a Beginner & a Master Course - teaching everything you need to know about Midjourney. I think it should be available by the end of the month, sign up for my newsletters - ftp.live/newsletter - if you want a special discount when it's released
I’ve been experiencing with this new version and photographic images are amazing. However, it seems that with illustrations is harder to get results as with 5.2 or even with the previous version. What would you recommend?
Hmm.. I think there are 2 things that come to mind. 1. try using multi-prompts - like adding the style you want with a little bit of weight /imagine prompt: (subject)(style):: (style)::0.4 or 0.6 2. try taking an image you like from 5.2, image prompt it into Midjourney - write a short description in your prompt and use --iw 2 or even --iw 3 secret bonus tip, it's okay to keep generating in old versions if you like the results! Hope that helps a bit
One vid' I would love to see from you is advice on controlling the art-direction of more subtle and specific changes to a composition you like, one that just needs some minor improvements and alterations. For instance adding glowing red eyes to a few background characters or craning the camera up a notch without affecting the image dynamic overall. I have tried various methodology such as vary region with prompting and vary subtle with prompt updates but I find it all a bit hit and miss and the knock on effect is countless iterations of trying to rework it or accepting a degree of randomness. So any tips on existing tricks would be great, otherwise look fwd to the next version where they offer more such post initial prompt round control.
Unfortunately, you nailed it - 'countless iterations' - it's a slot machine, and there probably aren't any special tricks to remove the randomness. You may have better luck just taking your base image into photoshop and making all your adjustments from there.
ok cool cheers yep I have actually just 'surrendnered' my wasted attempts at the more subtle improvements so photoshop comp cheats seem to be the best way ahead for now until hopefully improved updates to MJ 'post' tools@@FutureTechPilot
I stopped using MJ for a few months & am thinking of returning. Kinda worried about the new prompting style though. I really enjoyed using more aesthetic words and letting it do it's own thing and being utterly wowed by the result. I don't usually have any certain look in mind, just a vibe I'm going for because I don't use it to create specific images. I use it to create things I can use to create other things.
Curious where you would choose to put in an artists name if you were looking for a specific influence? Such as Frida Khalo or Diane Arbus for example? @@FutureTechPilot
clarinet is so helpful and this is another category of helpful that is even more-so. I was pausing to take notes - yes written on paper because that is another tool that works for me. Thanks for the pdf as well. I can’t wait to try this; you were so succinct with your explanation, much appreciated!
One thing about all AI image generators is this one main issue. Perspective. For example, Someone in a hallway in a home, looking up at a skylight.... Now... This is where the issue is. AI doesn't yet understand there are 2 parts to a skylight in some cases. I lived in a house that had a 2 part skylight. In the hallway you look up and there's the ceiling or attic portion of the skylight, then there is the portion on the roof of the house or ceiling of the attic. I like to prompt my dreams over the last 33 of 40 years of my life haha. Is there a way to prompt perspective to where the AI can understand what I mentioned above?
I have two suggestions ... first - kinda boring, but I would literally prompt in the way you explained it. "looking up at the ceiling. In the middle of the ceiling is a circular skylight" second - I would find a reference image that had the look you wanted, and I would use that as an image prompt! Hope that helps
Very useful.. love the improvements to mid journey, and I love knowing how to use them. Thanks! One trick I've come up with that may be helpful is what I call left turns. You get a momentum by selecting a grid evolution selection several times and it seems to bake in the latent factors related to those choices, and then you take a left turn in the prompt and it results in some very interesting changes. Sort of echoes of the original, which you can bring into completely new situations and get a subtle effect holographicly.
/remix prefer Maybe what he means, after selecting a few iterations of variations, using remix offers you ability to evolve a chosen variation via modifying the original prompt. ?
@@juliansimmons7700 if you ask for cat. Then select a few, then without restarting the prompt, change it to girl. You get a weird cat girl. Results may vary depending on version, but there's momentum to details everywhere.
Hi Nolan, thanks for the tips and the explanation of the 2-4-2 formula.Would you say this works well for clipart/t-shirt/stickers etc? It seems everytime I try and add more information on V6 it doesn't like it very much but if I start extremely simply and add details to it slowly it behaves however,it doesn't tend to do well if I get too wordy or add in a lot so I'll have a bash at this method and see what happens.I've had a go with the formula adding in "white background," "clipart", "ultra-high detail," in among the prompt but it's still not playing ball. I get the feeling that there's a lot of tweaking and iterations to be done to get what you want or am I over complicaitng things do you think? It's probably because I'm relatively new to prompting but I love your videos and how you make no fluff hit the ground running content.Thanks ever so much! :)
Hmm.. I'm personally a fan of shorter prompts when it comes to designs like stickers and t-shirts ... so I'm not sure if the 2-4-2 would work well because I guess I haven't tried it much. Actually, I did it once recently and I'm not going to lie, the pictures were so cool. BUT it's cause I had a really specific description. I'm not sure I would add in a lot of extra words if I didn't know exactly what I was looking for. I also recommend trying --style raw if you haven't and both a lower and higher stylize value. Like --s 60 and --s 400 aaaaand you can honestly just use something like --v 5.1 or 5.2 ... some of the older models were amazing at clip art and things like that. Hope that helps
@@FutureTechPilot Hi Nolan, yeah, --style raw is my go to! ;) V6 does better with --150 and occasionally --s100 but when I'm happy with it I usually whack it up to --s 250 just to not eat up so many fast hours.I did hear that V5.2 was better so I'll go back to that and have a bash at V5.1 as well.I've been trying to get consistent results when changing from one character to another but it does tend to gimp off a tad and produce some pretty random or totally unrelated stuff which can be frustating but yeah,definitely some of the images it produces just blows me away.It's so flipping addictive! Thanks ever so much for taking the time to reply and for the tips.It's very much appreciated and very nice of you! Have a lovely day! :)
What is the best way to give a reference photo and for Midjourney to use it properly. Say I want an Apple Vision Pro on someones face, even when referencing an image, Midjourny asks still makes up it's own headsets.
Thank you! Quick question 🙋🏻♀️ if I want to replicate a particular painting style by a particular artist (say I want it to look like a VanGogh whereabouts in the prompt (or rhythm) is it best to tell Midjourney to make it look as post impressionism painter by Van Gogh?
Definitely at the front! And honestly, art styles are a little finicky right now, I would even recommend saying it at the end as well. Like double or triple up on the key words haha hope that helps
I wish I knew how to pull the image out! I keep adding things about feet and legs, and it doesn't matter most times. Midjourney just gets stuck on medium shots and close ups.
Hmm. I was definitely going to suggest mentioning shoes. And you can try changing the aspect ratio. I think 2:3 and 4:5 have been good for getting full bodies. I know it's frustrating though
Currently tricky for just a single person, for more than one easier. I.e. Standing men and women. When zoom out appears we’ll have a solution to one figure, or maybe the beta release will help.
Would be very happy if you could give examples with reference images of faces. They seem to be almost always in portrait mode,no matter if you add wide-angle or street photo.
I'll make a bunch of 'Basics of Midjourney' videos soon .. but to put it simply - style raw is like a twin of the regular Midjourney algorithm. It was trained on the exact same things, it just interprets your prompt more literally. This can often lead to more 'boring' or 'ugly' images - but it's really useful if you find that the default version isn't following your prompt exactly
I have been using chat GPT along side it v4 the difference in V6 MJ is so much better, just re learning to prompt but that video I am experimenting with now
Thank u Nolan for this good video and for the keyword " 1980s film screenshot" i got more realistic images❤ can u make a video again how to generate photorealistic with V6. And do u know how to solve hand problems by V6? Thank u
hmm.. hands? I think that's definitely just a 'slot machine' thing, as in, you just have to get lucky. BUT, using Vary Subtle might help correct wonky hands (I've done that recently) ... and photorealistic images? I'll look into it but Version 6 tends to make things photorealistic without asking ya know? So I'm not quite sure which words will emphasize it
@@FutureTechPilot hi thx for ur fast answer. I asked the midjourney guru from suportteam. He told me to use just the keywords photograph in front of my promts. 😊
Thank you for the prompt tips, but there are still a number of issues that I have yet to overcome. For example, how can I avoid double barrel confusion? If I want an image of magnet fishing, how do I avoid all the unwanted angling images?
@@christopherd.winnan8701 I don't use other generators enough to make a recommendation. But Leonardo AI or Stable Diffusion can probably do what you're looking for
Thank you so much for this! I’m new and am really enjoying your tutorials. I’m wondering you have a tutorial on eliminating facial and hand/feet distortion. I’m having real trouble with that - thank you again!
Hmm .. I don't think I have any videos that talk about that :( sorry - but my suggestion would be to use Vary Subtle on any image that doesn't look quite right. I've found that to help fix hands (eventually)
It is still in alpha with limited functionality. Developers promised all those features, and more :) Latest update focused on default styling change and coherency improvements. Beta is coming soon :)
Thanks so much for this info. Will be busy today testing out new prompts. Your videos always provide great insight. Not sure if this is the right place to ask. Since V6 seems some of my midjourney images have noise or such. I use topaz upscale ai. Topaz worked great before. Is it me or did something change with v6
haha you know what, I've kinda noticed the same thing (I use Topaz too) - and V6 definitely changed, but maybe we just have to wait for it to get updated again
I continue to believe that fewer words produce a stronger result. You should always strive to be as concise in your description. Every word "dilutes" the result causes the AI to have to weld more concepts into a single image.
It took me forever to figure out how to approach women and engage them in conversation and now I have to learn how to pick up midjourney? XD XD Holy Moly! Thanks for the awesome explanation. And yes, now it really makes sense. And it works.
Hi Nolan, thank you for the video, It's was super helpful. When I try to download the PDF, the mail returns without a Link Attach. Can you Check it out? Thanks a lot. Inbal
I think that your ideas around Phrasing is spot on. The da-da-da is Phrasing. Poetic phrasing. Anyone who writes extensively - lyrics or poetry - and actors/performers who acutely understand language in that way in order to interpret it using voice/song recognise this immediately.
It’s like a creative writing course and you are talking about the structure of language, breaking it down in to its parts. I’d be very interested to see you interview a skilled writer about the expressive sophistication we have to develop to effectively interface with MJ. Soon there will be full courses at universities on how to write for these creative interfaces! You are an amazing teacher.
Interviewing creative people is a dream of mine, so I'm glad to hear you would be interested in something like that! Thanks for the support :)
@FutureTechPilot I respect you and am grateful for everything you’ve taught me. Language is the ultimate interface. Looking forward to your first interview!
Love it! I wrote a prompt as I was watching along with this video and ended up having the perfect dada, dada dada dada dada, daDA prompt written with all your instructions. Thank you so much for this and making it so easy to understand with that technique.
Well I'm really glad to hear that! Thanks for taking the time to share your thoughts
Hi. The best Channel about Midjourney in UA-cam. Let's hope it continues to grow and soon reaches 100,000 subscribers. From my side I will recommend it
haha I appreciate the support pal!
Outstanding. So helpful for those of us just starting our midjourney journey. Thank you.
Da-da. Da-da da-da da-da. Da-da. 😊
haha I'm glad you liked it!
¡Gracias!
Thanks!!
your advice has already made the results of my prompts immeasurably better... thank you so much
Hey I'm really glad to hear that!
Thank you for the Rhythm bit. It can be helpful to remember. Once again an excellent video demonstration. Don't listen to some weird comments who didn't like your rhythm. After all its your videos and you can do whatever you like in your videos. There are many people who are enjoy watching you. Don't pay attention to the negative comments. If they find it silly, then they are free to watch other videos. You are putting to much effort and knowledge into these videos. don't let some idiots change the way you are teaching to us. If they are smart enough they can record their own video tutorials and we are here to judge them too if they dare. Keep up the amazing professional work you are doing Nolal. 👍👍 By the way, can you create videos for other Ai generators as well? not exclusively for Midjourney. Such as for Dalle-3, Leonardo, etc...
haha all feedback is valuable! I'm happy to know when my videos don't connect with people - it let's me know that I can improve in that area. But as always, thank you for the support!
And I have nothing against other ai art generators, but it will be hard to make videos on them for one reason - my current video 'to-do' list is so ridiculously long (well over 200) and it can be difficult to switch focus and prioritize other content over the ideas I already have scheduled.
But I really do appreciate you asking, so I know it's important and maybe I can get to that idea sooner than expected!
i just love this guy ! always come with best tutorials !
🤝 I'm glad I could help
Seems like it’s like writing screenplay description
Yah definitely something like that!
Great. Very helpful. I can't wait to try it out!
Cheers!
Unexpectedly helpful - at first I thought you were kidding but it actually makes a lot of sense. Thanks
haha well I'm glad to hear that! Thanks for the feedback
AWESOME video! The fact that you approached it from a rhythm point of view is not just creative and rare but also pretty darned accurate. Once I started using the 'v6 language, ' it was insane what I could get. Clarinet is _so_ incredibly helpful, it's unreal. I had started looking at it as bookmarks with books in between. Keep adding to the books until it stops recognizing. Great video. Very, very helpful. I was beginning to have Nolan withdrawal! lol
haha I'm glad people are connecting with the video. I really wasn't sure what the reception would be like. As always, thanks for sharing your thoughts! I'm so thankful Clarinet and their team took the time to write those guides 🤝🤙
i've been using midjourney for some time. ive learned puppeteering, early letters, lots through claire A net
What a great video!!!!
Cheers!
Terrific Tutorial ! This is extremely helpful and i super appreciate your generosity by sharing this formula! The rhythm approach is very unusual and original. 1000 thanks!
And thanks a lot for the feedback! I'm glad it worked well for you, cheers pal
Thanks, again and well done. As I was watching your video I was visualizing a traditional vertical list format to learn the new prompting--like a grocery list with number two being the main course. Once again, MJ and FTP are making my brain think different.
Cool idea! I've always loved the analogy of food/ingredients/recipe to explain Midjourney prompting and I think your idea of a grocery list makes a lot of sense!
What a great idea - I love grocery lists!
Like 5-coarse menu in a restaurant maybe? (I have to think about that first, to see if I even know that works)
I wonder what happens if you change the “rhythm” or the order of the “menu” or “grocery list”. 🤔 🫣
whats FTP?
This helped me so much
That makes me happy to hear!
Why does the AI seem to equate Wizard/Sorcerer/Warlock only with elves? Even if I prompt with Human Wizard, the result is pointy ears.
maybe you could try --no elves ? something like that!
❤MJ releasing model personalization. Get at least 200 image ratings and then add --p after your prompt. Personalization completely changes the way system interprets your prompts.
Hey man, good to see you. Thought you got lost in the post ( pun intended ). You can definitely be counted on to do some heavy lifting; excellent recipe.
You always make great sense and you have nailed it with this one. Thanks for your commitment...
I really appreciate the feedback! I'm glad my style works well for you
[ INT. ARTISTS STUDIO -- NIGHT ]
An easel sits in a messy studio. A nearby window lets in the NEON LIGHT from a sign outside. The floor is flecked with dried paint.
^^ he's basically describing screenwriting style.
Yah that's a good observation! That mindset really works well in Midjourney V 6
Hey Nolan, I watch and appreciate yr vids on a daily basis. 2,4,2 - Sounds like a "double stroke roll" in drumming. 🥁 😁
haha okay it's cool that it's applicable across some different mediums
Midjourney leapfrogs Back to its No1 position in the AI Art scene.
Nolan remains at his No1 position of producing the most comprehensive Midjourney tutorials available on UA-cam 🙏👊🔥
Never lost it.
hahah well I appreciate the comment. Cheers buddy!
Ha - it's how scene descriptions in scripts are written. Of course. Wow, great insights in this video. thank you
haha yes, you got it. enjoy!
this makes sense! you are speaking our language
haha I'm glad I could help out!
Great video, your way feels like a script description of a scene so it makes a lot of sense.
haha cool I'm glad it clicked for you
Brilliant, thanks. I got it right away & it makes a lot of sense
Hey I'm glad it clicked for you!
Midjourney haikus. Love it. Very well explained. Thanks.
haha I'm glad it made sense to you!
Dear Nolan, Curious as I am when it comes to images, I was able to follow an introduction about MJ a month ago. And soon I was blown away by what I saw appear! I now have a basic plan with MJ and have already been able to make a few nice pictures, but...I have also become very eager and I want to know much, much more about MJ! How lucky I was that I saw your "Future Tech Pilot" fly by ;-) And I have already learned a lot from you. Your way of explaining the material appeals to me very much, you do it so pleasantly and full of enthusiasm, I really appreciate that and it is easy for me to follow!, (with some help from Google translate!,..:- ))
Many thnx!!! from Amsterdam!
Btw, I thought I heard you say something in one of your videos about a complete(?) workshop or a course that you could provide?,...
Hey thanks a lot for the kind words! Yes, I'm currently working on both a Beginner & a Master Course - teaching everything you need to know about Midjourney. I think it should be available by the end of the month, sign up for my newsletters - ftp.live/newsletter - if you want a special discount when it's released
I’ve been experiencing with this new version and photographic images are amazing. However, it seems that with illustrations is harder to get results as with 5.2 or even with the previous version. What would you recommend?
Hmm.. I think there are 2 things that come to mind.
1. try using multi-prompts - like adding the style you want with a little bit of weight /imagine prompt: (subject)(style):: (style)::0.4 or 0.6
2. try taking an image you like from 5.2, image prompt it into Midjourney - write a short description in your prompt and use --iw 2 or even --iw 3
secret bonus tip, it's okay to keep generating in old versions if you like the results!
Hope that helps a bit
Thanks, this gives me good ideas on how to start describing concepts for art I can draw.
Hey I'm happy to hear that!
Thanks, Nolan. This is great. Interestingly, these prompts sometimes feel like the prompts of days old.
Time is a flat circle 😂
Thanks Nolan, I always get great info and examples from your videos. This really helps with the new prompting.
And thanks a lot for the feedback!
One vid' I would love to see from you is advice on controlling the art-direction of more subtle and specific changes to a composition you like, one that just needs some minor improvements and alterations. For instance adding glowing red eyes to a few background characters or craning the camera up a notch without affecting the image dynamic overall. I have tried various methodology such as vary region with prompting and vary subtle with prompt updates but I find it all a bit hit and miss and the knock on effect is countless iterations of trying to rework it or accepting a degree of randomness. So any tips on existing tricks would be great, otherwise look fwd to the next version where they offer more such post initial prompt round control.
Unfortunately, you nailed it - 'countless iterations' - it's a slot machine, and there probably aren't any special tricks to remove the randomness.
You may have better luck just taking your base image into photoshop and making all your adjustments from there.
ok cool cheers yep I have actually just 'surrendnered' my wasted attempts at the more subtle improvements so photoshop comp cheats seem to be the best way ahead for now until hopefully improved updates to MJ 'post' tools@@FutureTechPilot
I stopped using MJ for a few months & am thinking of returning. Kinda worried about the new prompting style though. I really enjoyed using more aesthetic words and letting it do it's own thing and being utterly wowed by the result. I don't usually have any certain look in mind, just a vibe I'm going for because I don't use it to create specific images. I use it to create things I can use to create other things.
You can still do that for sure! That's how I prompt, but this video is more about showing you what is possible when you expand your descriptions
Brilliant use of rhythm and call back!
I'm glad you think so!
Curious where you would choose to put in an artists name if you were looking for a specific influence? Such as Frida Khalo or Diane Arbus for example? @@FutureTechPilot
Such a useful video and absolutely enjoyable to watch! Tara tara....I am still singing!Thanks for the neurohack!
hahah I'm glad it works well for you!! Thanks for taking the time to comment
Tara, tarataratara, tara ta.... got it! Thank you! :)
🎵😂
Version 9 prompts will have to be in the form of a haiku, with variations available through interpretive dance.
Thanks for another great video Nolan!
V12 and we imagine it in our minds and voilà!
😂😂 you're gonna have to sing an album to the bot
Solid video, keep it up!
Cheers buddy!
Really love the way you explain everything 💥
haha I'm glad my style works well for you!
clarinet is so helpful and this is another category of helpful that is even more-so. I was pausing to take notes - yes written on paper because that is another tool that works for me. Thanks for the pdf as well. I can’t wait to try this; you were so succinct with your explanation, much appreciated!
I appreciate a good note-taker haha I have lots of post-it notes on my desk. Thanks for the kind words and I'm glad my style works well for you
I didn't expect that technique was gonna get such good results.
Great stuff 💯
haha I know right?
very cool explanation 💪
Hey I'm glad you think so!
I love your videos! Thank you
And thank you for taking the time to comment!
0:40 Hey Nolan, I learnt from one of the promptcraft events that MJ does not understand words like "very" or "some" or "big". Not yet at least.
Fascinating!! Thanks a lot for letting me know
Thanks nolan, when is the song coming out to accompany the tune? da da da da da da??
😂 I'll hop in the studio soon
One thing about all AI image generators is this one main issue. Perspective. For example, Someone in a hallway in a home, looking up at a skylight.... Now... This is where the issue is. AI doesn't yet understand there are 2 parts to a skylight in some cases. I lived in a house that had a 2 part skylight. In the hallway you look up and there's the ceiling or attic portion of the skylight, then there is the portion on the roof of the house or ceiling of the attic. I like to prompt my dreams over the last 33 of 40 years of my life haha.
Is there a way to prompt perspective to where the AI can understand what I mentioned above?
I have two suggestions ...
first - kinda boring, but I would literally prompt in the way you explained it. "looking up at the ceiling. In the middle of the ceiling is a circular skylight"
second - I would find a reference image that had the look you wanted, and I would use that as an image prompt!
Hope that helps
amazing thank you, i just started learning 2 days ago your great
I'm happy to help!
Very useful.. love the improvements to mid journey, and I love knowing how to use them. Thanks! One trick I've come up with that may be helpful is what I call left turns. You get a momentum by selecting a grid evolution selection several times and it seems to bake in the latent factors related to those choices, and then you take a left turn in the prompt and it results in some very interesting changes. Sort of echoes of the original, which you can bring into completely new situations and get a subtle effect holographicly.
Could you elaborate a little more on that idea? You can email me if that's easier - futuretechpilot@gmail.com
/remix prefer
Maybe what he means, after selecting a few iterations of variations, using remix offers you ability to evolve a chosen variation via modifying the original prompt.
?
@@juliansimmons7700 if you ask for cat. Then select a few, then without restarting the prompt, change it to girl. You get a weird cat girl. Results may vary depending on version, but there's momentum to details everywhere.
I like it. Your pattern for a prompt.
Yah pattern is a good way of wording it too
Thank you. Effective and concise tutorial for great images. 🎨✨️❤️
Happy to help!
this prompting style feels like spoken word poetry!
haha yah I think you're right
This is insan, I love spending time generating pictures, i wish I could make it my job lol
😂 that's why I started a youtube channel
Ha wow alrighty my MJ Ai brother, we are talking a whole other process for NEXT LEVEL images!! 🎉
hahah cheers pal, I'm glad what I said made sense to some people lol
Hi Nolan, thanks for the tips and the explanation of the 2-4-2 formula.Would you say this works well for clipart/t-shirt/stickers etc? It seems everytime I try and add more information on V6 it doesn't like it very much but if I start extremely simply and add details to it slowly it behaves however,it doesn't tend to do well if I get too wordy or add in a lot so I'll have a bash at this method and see what happens.I've had a go with the formula adding in "white background," "clipart", "ultra-high detail," in among the prompt but it's still not playing ball. I get the feeling that there's a lot of tweaking and iterations to be done to get what you want or am I over complicaitng things do you think? It's probably because I'm relatively new to prompting but I love your videos and how you make no fluff hit the ground running content.Thanks ever so much! :)
Hmm.. I'm personally a fan of shorter prompts when it comes to designs like stickers and t-shirts ... so I'm not sure if the 2-4-2 would work well because I guess I haven't tried it much. Actually, I did it once recently and I'm not going to lie, the pictures were so cool. BUT it's cause I had a really specific description. I'm not sure I would add in a lot of extra words if I didn't know exactly what I was looking for.
I also recommend trying --style raw if you haven't and both a lower and higher stylize value. Like --s 60 and --s 400
aaaaand you can honestly just use something like --v 5.1 or 5.2 ... some of the older models were amazing at clip art and things like that.
Hope that helps
@@FutureTechPilot Hi Nolan, yeah, --style raw is my go to! ;) V6 does better with --150 and occasionally --s100 but when I'm happy with it I usually whack it up to --s 250 just to not eat up so many fast hours.I did hear that V5.2 was better so I'll go back to that and have a bash at V5.1 as well.I've been trying to get consistent results when changing from one character to another but it does tend to gimp off a tad and produce some pretty random or totally unrelated stuff which can be frustating but yeah,definitely some of the images it produces just blows me away.It's so flipping addictive! Thanks ever so much for taking the time to reply and for the tips.It's very much appreciated and very nice of you! Have a lovely day! :)
Very clear (your trademark 😊); and very useful
haha I'm glad I could help!
What is the best way to give a reference photo and for Midjourney to use it properly. Say I want an Apple Vision Pro on someones face, even when referencing an image, Midjourny asks still makes up it's own headsets.
You could try using it an image prompt. But from the sounds of what you want, it's probably way easier to do it in photoshop
Thank you! Quick question 🙋🏻♀️ if I want to replicate a particular painting style by a particular artist (say I want it to look like a VanGogh whereabouts in the prompt (or rhythm) is it best to tell Midjourney to make it look as post impressionism painter by Van Gogh?
Definitely at the front! And honestly, art styles are a little finicky right now, I would even recommend saying it at the end as well. Like double or triple up on the key words haha hope that helps
Good stuff. I'm sure I've used your tips and it's always good to get a refresher here and there 😀 🥂🖖
Cheers pal!
I wish I knew how to pull the image out! I keep adding things about feet and legs, and it doesn't matter most times. Midjourney just gets stuck on medium shots and close ups.
Hmm. I was definitely going to suggest mentioning shoes. And you can try changing the aspect ratio. I think 2:3 and 4:5 have been good for getting full bodies. I know it's frustrating though
Currently tricky for just a single person, for more than one easier. I.e. Standing men and women. When zoom out appears we’ll have a solution to one figure, or maybe the beta release will help.
Would be very happy if you could give examples with reference images of faces. They seem to be almost always in portrait mode,no matter if you add wide-angle or street photo.
I've heard that mentioning shoes will get a full body shot. Also using a 2:3 aspect ratio will help
It still can't seem to comprehend wide angle or full body etc. It always wants to do close-ups of things.
You're right, super frustrating. I've heard that mentioning feet or shoes is a good way of triggering it. As well as the aspect ratio of 2:3
The rest of us were doing "Yaba daba doo" no wonder it wasn't looking right.
😂😂 honest mistake
Hi. Thanks for the video. Can u explain style raw please
I'll make a bunch of 'Basics of Midjourney' videos soon .. but to put it simply - style raw is like a twin of the regular Midjourney algorithm. It was trained on the exact same things, it just interprets your prompt more literally. This can often lead to more 'boring' or 'ugly' images - but it's really useful if you find that the default version isn't following your prompt exactly
FINALLY YOU DO THE ASTRONAUTS!! THAT'S MY GIMMICK! LOL
😂 I love seeing reflections in the helmet
I did not receive confirmation email for pdf? Any help is appreciated.
Please check your spam folder! If not, you can email me - ftp @ futuretechpilot.com
@@FutureTechPilot Thanks...I have multiple spam folders, did not see it at first. Thanks for your work, Great stuff!
subscribed
very helpful, thank you
Cheers!
I have been using chat GPT along side it v4 the difference in V6 MJ is so much better, just re learning to prompt but that video I am experimenting with now
💯 ChatGPT provides some AMAZING prompts when used in Midjourney
Thanks, this explains a lot!
Glad I could help!
Thank u Nolan for this good video and for the keyword " 1980s film screenshot" i got more realistic images❤ can u make a video again how to generate photorealistic with V6. And do u know how to solve hand problems by V6? Thank u
hmm.. hands? I think that's definitely just a 'slot machine' thing, as in, you just have to get lucky. BUT, using Vary Subtle might help correct wonky hands (I've done that recently) ... and photorealistic images? I'll look into it but Version 6 tends to make things photorealistic without asking ya know? So I'm not quite sure which words will emphasize it
@@FutureTechPilot hi thx for ur fast answer. I asked the midjourney guru from suportteam. He told me to use just the keywords photograph in front of my promts. 😊
Very helpful ! Thanks a lot
Cheers!
Feels like to teach tow to write sentences. 101. back to primary school :)
haha something like that!
Thank you for the prompt tips, but there are still a number of issues that I have yet to overcome.
For example, how can I avoid double barrel confusion? If I want an image of magnet fishing, how do I avoid all the unwanted angling images?
I'm not sure I understand your question. Can you rephrase it please?
@@FutureTechPilot When I try to generate images of magnet fishing, I only get images of normal fishing. Can you think of a useful solution?
@@christopherd.winnan8701 Try using a reference image as an image prompt
@@FutureTechPilotSadly, I do not have access to MJ. Do you know of an alternative generator that is capable of accepting reference images?
@@christopherd.winnan8701 I don't use other generators enough to make a recommendation. But Leonardo AI or Stable Diffusion can probably do what you're looking for
Thank you so much for this! I’m new and am really enjoying your tutorials. I’m wondering you have a tutorial on eliminating facial and hand/feet distortion. I’m having real trouble with that - thank you again!
Hmm .. I don't think I have any videos that talk about that :( sorry - but my suggestion would be to use Vary Subtle on any image that doesn't look quite right. I've found that to help fix hands (eventually)
@@FutureTechPilot thank you for the suggestion - I’ll try that!
Thank you so much for this!!!! ❤
and thank you for the support!
wow,!, thank you so much!
Cheers!
Cool examples, good information.
Thanks for the feedback! I was worried I didn't explain this in the best way possible, I hope I helped a little
I found it very helpful.@@FutureTechPilot
Great video
Thanks! 🤝
Another amazing video- as always. I have tried v6. I am wondering why you cant zoom out, custom zoom, vary region etc. Perhaps eventually?
It is still in alpha with limited functionality. Developers promised all those features, and more :) Latest update focused on default styling change and coherency improvements. Beta is coming soon :)
They're working on it! Maybe around 10 days from now we'll see it
Excellent especially the dash dash and space for specific instructions
I'm glad I could help!
Very informative
I'm glad I could help!
Super! Thank you
Cheers!
Damn gonna have to do an english course now, miss the old way :(
😂 I know what you mean, but this video might help you a little - ua-cam.com/video/FVBMSE9mLyI/v-deo.html
2 - Subject
4 - Description
2 - Atmosphere
Makes sense.
Yah well put!
great video, so now prompt more like you would in chatgpt, more a statement than a command
Yah I think that's a fair way of describing it!
Thanks so much for this info. Will be busy today testing out new prompts. Your videos always provide great insight. Not sure if this is the right place to ask. Since V6 seems some of my midjourney images have noise or such. I use topaz upscale ai. Topaz worked great before. Is it me or did something change with v6
haha you know what, I've kinda noticed the same thing (I use Topaz too) - and V6 definitely changed, but maybe we just have to wait for it to get updated again
Thank you !
Cheers pal!
i like video
I continue to believe that fewer words produce a stronger result. You should always strive to be as concise in your description. Every word "dilutes" the result causes the AI to have to weld more concepts into a single image.
Yah I like shorter prompts as well! But now with V6 MJ is capable of a lot more
It took me forever to figure out how to approach women and engage them in conversation and now I have to learn how to pick up midjourney? XD XD Holy Moly! Thanks for the awesome explanation. And yes, now it really makes sense. And it works.
😂 I'm glad my style works for you!
Thank you
Cheers!
Thank you so much
Cheers!
Sooooo helpful!!!!!!!! xoxoxoxoxo
I'm glad to hear that!
Hi Nolan, thank you for the video, It's was super helpful. When I try to download the PDF, the mail returns without a Link Attach. Can you Check it out? Thanks a lot. Inbal
I just checked and it should all be working properly. Is there a button in the email to take you to the download?
@@FutureTechPilot THank's It's work now
Why is describe still using V5.2 to create images.
I think it has to do with what version you have set as a suffix. I think ...
Love it!
Hey I'm glad to hear that!
Looking for the link for the PDF...
ftp.live/242 !
Ironically, I have been prompting in this way since v4 with mixed results......sometimes the results were extraordinary sometimes terrible
haha oh my gosh, now I want to try some of these prompts in V4 ...
V cool ... I like the notion of 'rhythm' as it makes it easier to remember the structural ptmopts that work best
Yah I was hoping that idea would click with other people