Love it. I like how well this also showcases how you can simply ask for what you want and usually get it... all while using very simple prompting methods.
fisheye shot of an animated garden gnome dancing through a knolled mushroom forest, in the style of Pop Surrealism, photorealistic detail, gravity-defying landscapes, soft edges and blurred details, wimmelbilder, vray, playful character design, colour palette of verdigris, rust and cyan UHD image 32K --ar 16:9 --q 2 --v 5.1 Ed: Adding "Pop Surrealism" as the style really helped to define this image.
I’ve been playing with scale in knolling just for fun. I started with knolling the celestial objects in a section of a galaxy. And of course I’m now progressing to microscopic items. 😁
@@TheFuzzypuddle Haven’t tried yet but I might. Could be interesting. I’m a bit cranky with macro because it wasn’t doing what I wanted in another project recently. Lol
Thanks for the great tutorial, I've been playing with knolling since v4 days. I never tried doing them in different styles, so thanks for that! Amazing work as always
I would love your take on making mid journey produce pictures that it does not make naturally. Recently I wanted to try making a top view gaming map for a pen and paper game. I could not generate anything, maybe 1/10 got close to what wanted. After sometime I went and found three references pictures, used /describe three time on each picture. Then I took all the descriptions and fed them in chatGPT. Explained the problem and he analyzed the mass of word to give me the key words I needed. With those I was able to get much more reliable results.
This is so weird. This is already the second comment that suddenly popped up out of nowhere and I didn't respond. Natural language is useful only for describing scenes or context between subject and environment. As soon as you want something extremely specific with a very unique look (like your example), you need to find the right words that MJ has been trained on.
@@TokenizedAI that’s what I did. Using pictures of what I was trying to copy I used MJ /describe to generate a lot of descriptions. Then I asked ChatGPT to analyze that wall of text in order to extract recurring words. That was a way to find quickly which words MJ has been trained with for pictures that do no have clear /describe.
/imagine prompt: heroic shot of animated troll castle made of knolled mushrooms in the style of Pop Surrealism, photorealistic detail, gravity-defying landscapes, wimmelbilder, soft edges and blurred details, vray, inventive architectural design, colour palette of verdigris, rust, gold and Prussian blue UHD 32K, --ar 16:9 --q 2 --v 5.1 Ed: Adding "Pop Surrealism" as the style really helped to define this image.
Hi Christian...I haven't tried yet to write prompts in my language, Italian. Do you think the A.I. of Midjourney would produce the same results as if I wrote in English? Good job and thanks for everything
Excellent - for generic concepts. Some of us - however - ARE CREATORS OF ORIGINAL CONTENT and DESIGNS. CAN WE DO THIS WITH ORIGINAL PRODUCT photograph.? 1) What is the prompt and 2) how do we introduce our own copyrighted image s of a Trademarked product to create a “knolling” page with all my products?
Because I recorded this before v5.1 was released. I also frequently switch between versions. Just because it's the newest version, doesn't mean it's always better. The people on social media who write stuff like "Midjourney just released v5.1 and it's incredible!" just say that to get attention. Most of time they can't even tell both apart 😅
@@TokenizedAI understood. I agree with your general sentiment, new isn't always better, however I do feel that it is possible to narrow down the best version for different usecases. I feel like --style raw replaces --v 5 and --v 5.1 replaces --v 4 (less stylised vs more stylised). :)
This is the right video.😊 As cool as it is whiteout actual existing products it's usage is very limited. Why would I use non existing whiskies anywhere? Now if you can create a video how to create an image that would use existing factual products that has a commercial use.
I think you're underestimating the use case for generic stock photography on millions of websites. You might not see how this is useful but trust me....even without specific brands, it is very useful to many people. Either way, you're picking an arbitrary example as an argument, when it's actually about the technique.
@@TokenizedAI I've picked an example that is close to me, but I agree that it might be fringe. As a web designer however I always stay away from generic stuff. Whatever I build needs to be brand related. Exact city, exact building, exact people and definitely exact products. Any image that doesn't build upon that is just noise. That's why I rarely use stock photography in the first place. Now I'm interested in training AI on specific products so it can spit out great images with them, but I'm struggling, might be the limitation of technology where randomness of AI generation just works against it might be my lacking of technique. Best results so far I had with just ai generated backgrounds that I manually blend in Photoshop with the products. I wonder if you have any ideas for that?
This is awsome....i work as a graphic designer for 5 star+diamonds hotel...so know i can make various cuisines pictures faster, better and beutiful than our chefs...thank you Christ
You guys are reading way too much into these tiny details 😉 I recorded this before v5.1 was released. I also frequently switch between versions. Just because it's the newest version, doesn't mean it's always better.
Not sure whether a single example is necessarily a good benchmark for coming to the assessment that it "sucks" overall. Seems a little reductive, don't you think? 😅
@@TokenizedAI you're probably right. knolling worked out when I tried a bikepacking prompt, I was just focused on making EDC gear and got annoyed with the piles of misshapen bones and sticks it was spitting out.
Ich glaub die jungen Menschen nennen das heutzutage "Cringe" 🤪 Es gibt nichts, was mehr "Cringe" ist, als ein Deutscher, der versucht "Cool" zu sein (Anm: Bin mir der Ironie bewusst, lol), in dem er Worte wie "Cringe" im Deutschen verwendet und noch nicht einmal einen Bezug zum Ausland hat. Aber vermutlich bin ich mit dieser Meinung schon ziemlich "sus" 🤣 "Boomer" bin ich mit 39 sowieso schon 🤦🏻♂️
@@TokenizedAI ich bin älter und kann englisch reden, die Jugend die es nicht kann aber denkt, weil bei denen jedes zweite wort ein englisches ist, einfach nur peinlich, naja :) aber geilsten finde ich jedoch das "th" bei 90% aller yt Kanälen, ich sage nur: i like sis and dad, und one two tree 😁 viel englisches yt hören lehrt dich: wenn eine Sprache eine Hürde hat, die dir probleme bereitet, ignoriere es einfach nur lass es weg: "th". Geiles Konzept....
sieht sehr gut aus, der " Knolling" prompt mit allen Variationen ! 😎
This is awesome! Always appreciate your videos and the way you break things down. Very informative and valuable.
I appreciate that!
OUTSTANDING
Thanks again for another insightful MJ video. I've learned the power of aggregating objects in scenes through the power of Knolling.
Outstanding.. wow !! thanks so much for the info.. cant wait to try
Thank you Christian. As usual your content is amazing. You are a breath of fresh air.
As long as it's not one of those air fresheners....cool 🤣
Nice video Christian! This is really a great presentation of the practical use of MJ. !
Thanks 🙏🏻
Love it. I like how well this also showcases how you can simply ask for what you want and usually get it... all while using very simple prompting methods.
Yup, not need to overcomplicate things 🙂
Please continue your journey into prompt engineering.
Very interesting! I especially liked how you switched things up with illustration and lineart. Keep up the good work.
Always looking for that slightly different spin 🙂
fisheye shot of an animated garden gnome dancing through a knolled mushroom forest, in the style of Pop Surrealism, photorealistic detail, gravity-defying landscapes, soft edges and blurred details, wimmelbilder, vray, playful character design, colour palette of verdigris, rust and cyan UHD image 32K --ar 16:9 --q 2 --v 5.1
Ed: Adding "Pop Surrealism" as the style really helped to define this image.
Great piece of work - as ever!
I’ve been playing with scale in knolling just for fun. I started with knolling the celestial objects in a section of a galaxy. And of course I’m now progressing to microscopic items. 😁
that's sounds cool, fun, and more than just a little OCD :)
do you use "macro shot" for microscopic items? it could be cool if it works
@@TheFuzzypuddle Haven’t tried yet but I might. Could be interesting. I’m a bit cranky with macro because it wasn’t doing what I wanted in another project recently. Lol
This video is going to end up being _way_ underrated. I just know it. Thank you. 👍
Mind you this: 6:03 is worth everything! 😅
Most of them are underrated 🤪
Thanks for the great tutorial, I've been playing with knolling since v4 days. I never tried doing them in different styles, so thanks for that! Amazing work as always
Always gotta experiment! 🙂
Thank you, great😊you are so creative❤
this is amazing, you have a such a nice way to edit your videos and help us visualize everything
Thank you 🙏🏻
Thank you ❤
Excellent as always! Would be great if you could do a video on creating book covers.
I would love your take on making mid journey produce pictures that it does not make naturally.
Recently I wanted to try making a top view gaming map for a pen and paper game. I could not generate anything, maybe 1/10 got close to what wanted.
After sometime I went and found three references pictures, used /describe three time on each picture. Then I took all the descriptions and fed them in chatGPT. Explained the problem and he analyzed the mass of word to give me the key words I needed.
With those I was able to get much more reliable results.
This is so weird. This is already the second comment that suddenly popped up out of nowhere and I didn't respond.
Natural language is useful only for describing scenes or context between subject and environment. As soon as you want something extremely specific with a very unique look (like your example), you need to find the right words that MJ has been trained on.
@@TokenizedAI that’s what I did.
Using pictures of what I was trying to copy I used MJ /describe to generate a lot of descriptions. Then I asked ChatGPT to analyze that wall of text in order to extract recurring words.
That was a way to find quickly which words MJ has been trained with for pictures that do no have clear /describe.
You find the most interesting topics 💜
POW!erful indeed 🎉 great highlights bro 👏🏻👏🏻👏🏻
this is amazing, thx
Fantastic vídeo 👏👏👏👏 Just fantastic.
how cool, thank you :)
Graet vídeo. Thanks
/imagine prompt: heroic shot of animated troll castle made of knolled mushrooms in the style of Pop Surrealism, photorealistic detail, gravity-defying landscapes, wimmelbilder, soft edges and blurred details, vray, inventive architectural design, colour palette of verdigris, rust, gold and Prussian blue UHD 32K, --ar 16:9 --q 2 --v 5.1
Ed: Adding "Pop Surrealism" as the style really helped to define this image.
Its great for spacial engineering and design
Embellishments is also works as knolling, I think)
Hi Christian...I haven't tried yet to write prompts in my language, Italian. Do you think the A.I. of Midjourney would produce the same results as if I wrote in English? Good job and thanks for everything
You can definitely use other languages but I can't tell you wether the results are equally as good.
Cool Video 👏
What prompt we use if i want produce wooden block cube, straight line arranged, top view photography?
That's a very specific question 🤔
A Truly great consept! Thanks for my behalf and all those too lazy to thank.
Excellent - for generic concepts. Some of us - however - ARE CREATORS OF ORIGINAL CONTENT and DESIGNS. CAN WE DO THIS WITH ORIGINAL PRODUCT photograph.? 1) What is the prompt and 2) how do we introduce our own copyrighted image s of a Trademarked product to create a “knolling” page with all my products?
Just curious, how come you are using --v 5 instead of --v 5.1 or --style raw?
Because I recorded this before v5.1 was released.
I also frequently switch between versions. Just because it's the newest version, doesn't mean it's always better. The people on social media who write stuff like "Midjourney just released v5.1 and it's incredible!" just say that to get attention. Most of time they can't even tell both apart 😅
@@TokenizedAI understood. I agree with your general sentiment, new isn't always better, however I do feel that it is possible to narrow down the best version for different usecases.
I feel like --style raw replaces --v 5 and --v 5.1 replaces --v 4 (less stylised vs more stylised). :)
This is the right video.😊 As cool as it is whiteout actual existing products it's usage is very limited. Why would I use non existing whiskies anywhere? Now if you can create a video how to create an image that would use existing factual products that has a commercial use.
I think you're underestimating the use case for generic stock photography on millions of websites. You might not see how this is useful but trust me....even without specific brands, it is very useful to many people.
Either way, you're picking an arbitrary example as an argument, when it's actually about the technique.
@@TokenizedAI I've picked an example that is close to me, but I agree that it might be fringe. As a web designer however I always stay away from generic stuff. Whatever I build needs to be brand related. Exact city, exact building, exact people and definitely exact products. Any image that doesn't build upon that is just noise. That's why I rarely use stock photography in the first place. Now I'm interested in training AI on specific products so it can spit out great images with them, but I'm struggling, might be the limitation of technology where randomness of AI generation just works against it might be my lacking of technique. Best results so far I had with just ai generated backgrounds that I manually blend in Photoshop with the products. I wonder if you have any ideas for that?
The majority of the results are still too jumbled up and crowded. Is there a prompt to adjust the distance between the objects?
Sounds like your only real solution is Photoshop then. Sorry.
This is awsome....i work as a graphic designer for 5 star+diamonds hotel...so know i can make various cuisines pictures faster, better and beutiful than our chefs...thank you Christ
That's exactly where this comes in handy so much 🙂
Спасибо! Очень помогаешь
The grandmother is grippy. Made me laugh
I know, right? 🤣
May I ask, why you seem to default to --v 5? Thank you
You guys are reading way too much into these tiny details 😉
I recorded this before v5.1 was released. I also frequently switch between versions. Just because it's the newest version, doesn't mean it's always better.
@@TokenizedAI "Aaah, I see" (said the blind man). Thank you.
Mr Token, I’m not getting alerts for new videos from you 😢
Check your settings 😱
awesome
I'm in the EDC community, I tried knolling a month or 2 ago on v5 and it sucked because midjourney sucks at knives. none of my attempts worked out.
Not sure whether a single example is necessarily a good benchmark for coming to the assessment that it "sucks" overall. Seems a little reductive, don't you think? 😅
@@TokenizedAI you're probably right. knolling worked out when I tried a bikepacking prompt, I was just focused on making EDC gear and got annoyed with the piles of misshapen bones and sticks it was spitting out.
You should try it now with v5.1. Sharp pointy things have improved 😅
Wow, MJ really has it done.
Endlich mal ein deutscher der gutes Englisch sprechen kann, Halleluja 🙏
Als jemand, der in Kanada aufgewachsen ist, wär alles andere auch peinlich 😅
@@TokenizedAI hört man 😏 die ganzen anderen Kanäle die ihre reichweite nur erhöhen wollen und daher einen "auf Englisch machen", Ohren-Krebs!! 🙉 😁
Ich glaub die jungen Menschen nennen das heutzutage "Cringe" 🤪
Es gibt nichts, was mehr "Cringe" ist, als ein Deutscher, der versucht "Cool" zu sein (Anm: Bin mir der Ironie bewusst, lol), in dem er Worte wie "Cringe" im Deutschen verwendet und noch nicht einmal einen Bezug zum Ausland hat. Aber vermutlich bin ich mit dieser Meinung schon ziemlich "sus" 🤣 "Boomer" bin ich mit 39 sowieso schon 🤦🏻♂️
@@TokenizedAI ich bin älter und kann englisch reden, die Jugend die es nicht kann aber denkt, weil bei denen jedes zweite wort ein englisches ist, einfach nur peinlich, naja :) aber geilsten finde ich jedoch das "th" bei 90% aller yt Kanälen, ich sage nur: i like sis and dad, und one two tree 😁 viel englisches yt hören lehrt dich: wenn eine Sprache eine Hürde hat, die dir probleme bereitet, ignoriere es einfach nur lass es weg: "th". Geiles Konzept....
Knolling grandmother 😂
😅
Why does it almost 2 minutes to tell us this is a prompt for midjourney? Too slow.
There's a big fat Midjourney logo in the thumbnail and it not only works in Midjourney.
You sure you're not the slow one here? 😘
/imagine Knolling a V8 engine --v 5.1 --style raw --ar 16:9
/imagine knolling the human body --v5.1 --style raw
Ew