I was trying to figure it out by myself and was struggeling. I spent many hours of testing and trying and didn't get quite the results I could learn from. Thank you for the video ❤
First anime I thought of when you said old anime was Akira. Legendary anime. That bike stop sequence is everywhere once you realise. Latest one was in Nope
Thanks for the video, I've always written my prompts in the wrong order, this helps. A small request, could you put English subtitles on the video, it will help a lot for people whose native language is not English
Hey! RUclips creates automatic subtitles which I've tested and they're very accurate. However, those take up to 1 day to be generated - so tomorrow this video should have subtitles! Thank you for watching and glad you found the video useful.
Thanks for doing this video. I have 2 questions please. 1) Do you find there's a difference between using the "reroll" icon and entering the prompt again? Because I find the more I use the reroll icon, more often than not the images get worse. However if I enter the same prompt afresh, things don't get worse but just varies. 2) Also do you find reroll varies when you reroll the parent image vs reroll the child images? I find rerolling the parent image is better than keeping on rerolling the latest child images. Rerolling the latest child images keep on getting worse for some reason. Just curious if you experiences the same thing. Thanks again for the video.
Thanks for watching! Both of these are brilliant questions - in the next couple of weeks I'll be positing a video on exactly this. I'm currently researching these questions and have made some super interesting findings. That being the case, for full answers you'll need to give me some time to complete that video but the short answers are: 1. There is a difference absolutely but I don't think you ever need to enter the prompt again, reason is in 2... 2. Yes it absolutely does, rerolling the parent image vs rerolling latest image do 2 completely different things but are both required for a good end result - I'll explain it all in my upcoming video!
When you reroll, do you keep doing deeper variations or just keep hitting the same initial prompt.? I would have loved to see a speed through of your iterations
Hey! I keep rerolling until I get a decent composition thats quite close to being correct. Then I start vrolling to clear out the imperfections. I'll make a video on that soon :)
@@cyankatana Wonderful. Its really all about that last bit that we need to move to expert. When you have a good image with imperfection, and then feels like you start again and it vrolls away from the original. Thanks. Loving the videos.
Creating your portraits in vertical aspect ratio --ar 2:3 for example will really help with that. Other than that its just a matter of rerolling until you get a composition you are happy with.
Thanks for watching! It could work as you are still adding to the the prompt, no doubt. Its about following best practices though. When I was experimenting with placements, I tried a Rick and Morty style - as soon as I moved it to the back it went for a realistic portrait.
Copy n paste to save time. anime cartoon style anime , Cyberpunk city background style by Katsuhiro Otomo, female with long green hair, detailed face, determined face -- 3:2 --test
REMIX can help you with that! I think that'll be a good topic for the next video - I'll try to work on your example and try to show you how it could be done.
I managed to get your 3 legged alien! Join my Discord if you want me to walk you through it - discord.gg/C8HB45U5Rk (drop me a message on there if you join)
Yeah these are the problems with these so called AI wizardry 😁 They are called "keywords" which supposed to be derived from meta information of the sources they are trained from. Ideally the prompt should be able to derives what regular users uses and tap automatically into correct keywords but yeah seems like we are not there yet for now 😁
Yeah, midjourney probably aren’t storing all those prompts for shits and giggles. Enjoy it while you can, I suspect the end goal is a full self running content cycle not too dissimilar to targeted ads. The individual is dead, we’re never going to get another Michelangelo etc ever again and if that’s not depressing I don’t know what is
![How to Character Design in Midjourney AI [Prompt Crafting Tutorial] - Ep.2](http://i.ytimg.com/vi/Xs1IS6sO3MI/mqdefault.jpg)
![Noob To Max With DRAGON REWORK In Blox Fruits [FULL MOVIE]](http://i.ytimg.com/vi/LnBMOinoOvA/mqdefault.jpg)
Want to see more videos like these where you try out new styles.
I think thats a good idea, coming your way soon!
Most helpful tutorials so far I have found for Midjourney, thank you do much!
Really happy to hear that. Thank you for watching!
I was trying to figure it out by myself and was struggeling. I spent many hours of testing and trying and didn't get quite the results I could learn from. Thank you for the video ❤
Glad this was helpful - Thank you for watching! :)
Awesome tutorials keep on posting amazing videos like these 🤩
First anime I thought of when you said old anime was Akira. Legendary anime. That bike stop sequence is everywhere once you realise. Latest one was in Nope
this is really helpful :) thank you
Excellent tutorial! Your speaking and the pace are just right. Thank you.
Thanks again Russ!
Excellent. Again.
Thank you Sven! Listening to your Spotify at the moment 🤘
@@cyankatana Thank you!
Thanks for this. I'm going to try it out.
Good stuff! Would be interesting to see minimalist/line draw styles!
Thanks brother! Definitely will cover in coming videos
very interesting way with prompt, many thanks for this video - hello from 🇫🇷
Hi Bruno! Thanks for watching yet another video my friend - have a fantastic day! 💙
The fact you used Akira btw Leonardo DiCaprio owns the script. Thanks for everything!
This was gold, Thank you
Thanks for watching! Glad it helped you out :)
Your content is amazing buddy, keep it up !
Thank you for watching my friend. Episode 2 of this series just released ruclips.net/video/Xs1IS6sO3MI/видео.html
Great tip, thanks!
Ghost in the shell or altered carbon
Thanks for the video, I've always written my prompts in the wrong order, this helps. A small request, could you put English subtitles on the video, it will help a lot for people whose native language is not English
Hey! RUclips creates automatic subtitles which I've tested and they're very accurate. However, those take up to 1 day to be generated - so tomorrow this video should have subtitles! Thank you for watching and glad you found the video useful.
Omg all my pictures were so wonky then I watched this video and now they turn out AMAZINg
Yay, this makes me really happy!
what are the "magic words" for photo-realistic images?
Try to add "photo realistic" or "realistic" at the very front of your prompt!
Ouaou! Impressive!
Thanks for doing this video. I have 2 questions please.
1) Do you find there's a difference between using the "reroll" icon and entering the prompt again?
Because I find the more I use the reroll icon, more often than not the images get worse. However if I enter the same prompt afresh, things don't get worse but just varies.
2) Also do you find reroll varies when you reroll the parent image vs reroll the child images? I find rerolling the parent image is better than keeping on rerolling the latest child images. Rerolling the latest child images keep on getting worse for some reason. Just curious if you experiences the same thing.
Thanks again for the video.
Thanks for watching! Both of these are brilliant questions - in the next couple of weeks I'll be positing a video on exactly this. I'm currently researching these questions and have made some super interesting findings. That being the case, for full answers you'll need to give me some time to complete that video but the short answers are:
1. There is a difference absolutely but I don't think you ever need to enter the prompt again, reason is in 2...
2. Yes it absolutely does, rerolling the parent image vs rerolling latest image do 2 completely different things but are both required for a good end result - I'll explain it all in my upcoming video!
@@cyankatana thank-you for your reply. I look forward to the video.
Not to get off topic but, Did you watch Power on Netflix?
can i make something like similiar anime like makima from chainsaw?? how
Yeah... Watch Akira. And Ghost in the Shell if you haven't. Not Stand Alone Complex, but the movie. Possibly the sequel.
Reminder to self: accurate writing is the key to a desirable image. 😎😎😎
When you reroll, do you keep doing deeper variations or just keep hitting the same initial prompt.? I would have loved to see a speed through of your iterations
Hey! I keep rerolling until I get a decent composition thats quite close to being correct. Then I start vrolling to clear out the imperfections. I'll make a video on that soon :)
@@cyankatana Wonderful. Its really all about that last bit that we need to move to expert. When you have a good image with imperfection, and then feels like you start again and it vrolls away from the original. Thanks. Loving the videos.
You don't know AKIRA? That's like saying i didn't know astronauts landed on the moon.
What words do I have to use to prevent Midjourney from cutting off the top of portrait's heads?
Creating your portraits in vertical aspect ratio --ar 2:3 for example will really help with that. Other than that its just a matter of rerolling until you get a composition you are happy with.
tell me you have since watched Akira!? It is one of my favourite films of all time
I want more videos with you))
Next video dropping on Thursday!
Good video but the screenshot of the prompt is hard to see with that video suggestions at the end
Thanks for letting me know! I did not notice that. Its fixed now 👍
@@cyankatana eyy awesome!! Thanks!! Keep it up!
@@theillogicalist 👏
A "Thesaurus" would probably help you when looking for "alternate" words.
Great tip!
I put "in the cartoon style of.." at the end it works fine. Akira was the original!
Thanks for watching! It could work as you are still adding to the the prompt, no doubt. Its about following best practices though. When I was experimenting with placements, I tried a Rick and Morty style - as soon as I moved it to the back it went for a realistic portrait.
@@cyankatana yeah I am going to try a comic with it so I got to get into this
I wish test and testp worked with image references :/
There is a workaround that can make that happen. I'll make a video tutorial on it 👊
Please, are there texts or links or .... something else to learn like the basic coding you talked about; --test --testp ...etc.
Thanks a lot
Why didn't you use Disney Tron legacy?
Had no idea about it. Thanks for sharing, looks great! But its not quite the Cyberpunk Edgerunners style I was looking for.
ti amo
Copy n paste to save time. anime cartoon style anime , Cyberpunk city background style by Katsuhiro Otomo, female with long green hair, detailed face, determined face -- 3:2 --test
No matter how hard I try, I can't get it to create a three-legged gray alien. Been working on it for days
REMIX can help you with that! I think that'll be a good topic for the next video - I'll try to work on your example and try to show you how it could be done.
I managed to get your 3 legged alien! Join my Discord if you want me to walk you through it - discord.gg/C8HB45U5Rk (drop me a message on there if you join)
@@cyankatana i will. Thank u
@@cyankatana I found your discord channel and was going to leave you a msg on it but it says I don't have permission
Look for Starblazer
Yeah these are the problems with these so called AI wizardry 😁
They are called "keywords" which supposed to be derived from meta information of the sources they are trained from.
Ideally the prompt should be able to derives what regular users uses and tap automatically into correct keywords but yeah seems like we are not there yet for now 😁
in my case, typing "in the style of x" at the front actually makes the image looks less aesthetically pleasing.
AI prompt writing is not safe from automation either, soon enough we will have an AI for that too
True
Yeah, midjourney probably aren’t storing all those prompts for shits and giggles. Enjoy it while you can, I suspect the end goal is a full self running content cycle not too dissimilar to targeted ads. The individual is dead, we’re never going to get another Michelangelo etc ever again and if that’s not depressing I don’t know what is
Déjà vu
You might want to spell check your English for future videos.