Nolan, MJ is a turnkey horizon event, but you keep coming up with new facets of what's possible. This is very nice. Thank you very much, this is generous. What's next with this sort of thing is the yellow pages, ranges, cataloging. It's kind of a foolish thing to suggest with something so vast, but I imagine this is where style dialing might evolve. Time periods, noir, rainbows, etc. Heh, now we just need to find a way to make ChatGPT generate it, right? LOL
Thanks so much for this video. --sref (whether used with codes or image reference) is my favorite Midjourney feature. It makes the possibilities virtually limitless.
Freakin' fantastic video! 💯 Would love to see more of these, and yes, if you wish to spend the money on a AI video generator on some of these would be brilliant. That dark gothic style...MAN that looks beautiful. Thank you. :)
Style mining is one of my favourite things to do in Midjourney for a quick hit of fun. Some styles can only partly effect an image, such as a color palette, contrast, line style or softness, etc., whereas other styles more strongly inform the entire aesthetic. If you like the look of a “weaker” sref, but are getting unpredictable results, prefixing the prompt with a specific type of image can help with consistency - e.g. “a cartoon of”, "a photo of” or "illustration,” etc., followed by the subject and the sref.
@@LASartor Pick a simple subject, and just run the same prompt paired with "--sref random" several times. A lot of the results will be relatively uninteresting, but about 1 in 10 are a keeper. Then try rerunning a code you liked with one or two different prompts to check that the sref is the main driver of the aesthetic.
Hi, thank you very much for your videos, I also learn with it. Tell me the Midjourney 6.1 change the way that we write some commands like style raw and ar? I had some problems with this yesterday.
Thank you, thank you. The catalog is great and has a couple of styles I usually have to describe in the prompts. I was hoping to find an SREF that gives an 8-bit pixelated video game style though. It can be prompted for, but an SREF would save prompt tokens and improve consistency between images. Dreamwave Horizon comes close.
Dude!!!! Craftlandia!!! This has to be one of my all-time favorites instantly!!!! Do you remember this prompt you shared: pilaf hypnagogic maximalism physicsbased paper collageaem, in the style of suburban flat shading, simplified figures, tantric manuscript, studyplace, expert draftsmanship, sculpture, oil pencil sketch, diagramic? Well let's just say if you have spend a little time crafting a little Cocktail with this prompt and Craftlandia, you get some amazing crazy wild images 🤣🤣 You're a legend man, thanks so much!
Hi Nolan! This was amazing! Thanks for your shoutout on your catalog!! I've used all that you found, its really great! ("Prism pop" i use a lot!!) Thank you.. do you think you'll keep adding new ones you find on the catalog?
Some will come to me naturally, but I like sending a bunch of pictures to ChatGPT and asking it for like 20 possible style names - most suck, but some are brilliant!
I think showcasing --sref without an explicit --sw is not very helpful. As we know, --s is the style persona strength (applies the --p nnnn style persona or the generic style persona if --p is not provided), and --sw is the (style) reference strength. If you use --sref, the default --sw is 100 but only with --sw 1000 you can set the full strength and quality of the sref. It's annoying how Midjourney uses the term style in 3 completely separate ways: --style is basically a different variant of the model. --s is the strength of the --p style persona or of the generic persona if --p is not provided. And --sw is the strength of --sref style reference. Style persona and style reference are two very different factors that influence the output. Style reference is very direct and kind of copies the style (including colors, textures and details) of one particular "image". Style persona is more general (because it's a set of preferences gathered across many images liked by a particular user) and affects the composition first before influencing the details.
Your reactions to just the colors and abstraction is fun to watch!
😂
Great stuff bro!!
Cheers!
YESS!!! I've been looking forward to this video! Great selections!
haha I'm glad you like em!
Many thanks Nolan!
Cheers!
always looking for new code to experiment, thx for sharing these! Great vid as usual
Cheers!
Tnx again for the recaps and the sref codes Nolan,..and get better soon!!
Cheers pal!
SO much amazingness!! I’ll def be playing w these today wooo hoooo
yaaa enjoy!
Great styles ! Thank you so much for sharing ❤
Cheers!!
Oh thank you for this, I cant wait to experiment!!!!
Enjoy!
I’ve only just found you but loving these tips! I’m always trying to get better, so have subscribed and will be binging your videos 😂
hahaha I'm happy to hear that!
Nolan, MJ is a turnkey horizon event, but you keep coming up with new facets of what's possible. This is very nice. Thank you very much, this is generous. What's next with this sort of thing is the yellow pages, ranges, cataloging. It's kind of a foolish thing to suggest with something so vast, but I imagine this is where style dialing might evolve. Time periods, noir, rainbows, etc. Heh, now we just need to find a way to make ChatGPT generate it, right? LOL
Yah I can't wait to see what sort of 'dials' get built around this technology!
Thanks so much for this video. --sref (whether used with codes or image reference) is my favorite Midjourney feature. It makes the possibilities virtually limitless.
Definitely an overwhelming amount of possibilities 😂
This is cool thank you
And thank you for taking the time to comment!
Thanks for the gum ball
Cheers!
I have so many codes I've been putting off organizing them I should probably go do that .
haha it's hard to keep track!
Superb
Freakin' fantastic video! 💯 Would love to see more of these, and yes, if you wish to spend the money on a AI video generator on some of these would be brilliant. That dark gothic style...MAN that looks beautiful. Thank you. :)
haha yeah the next time I experiment with a video generator, I'll use those pictures!
Style mining is one of my favourite things to do in Midjourney for a quick hit of fun. Some styles can only partly effect an image, such as a color palette, contrast, line style or softness, etc., whereas other styles more strongly inform the entire aesthetic. If you like the look of a “weaker” sref, but are getting unpredictable results, prefixing the prompt with a specific type of image can help with consistency - e.g. “a cartoon of”, "a photo of” or "illustration,” etc., followed by the subject and the sref.
Now that's a good tip!
how does one style mine?
@@LASartor Pick a simple subject, and just run the same prompt paired with "--sref random" several times. A lot of the results will be relatively uninteresting, but about 1 in 10 are a keeper. Then try rerunning a code you liked with one or two different prompts to check that the sref is the main driver of the aesthetic.
@@LASartor run the prompt: "" --sref random --repeat 3
Hi, thank you very much for your videos, I also learn with it. Tell me the Midjourney 6.1 change the way that we write some commands like style raw and ar? I had some problems with this yesterday.
Nope! Everything should be the same. But I know the service was having some outages recently, so maybe that's the problem you were running into
Love it!
Cheers!
Thank you, thank you. The catalog is great and has a couple of styles I usually have to describe in the prompts. I was hoping to find an SREF that gives an 8-bit pixelated video game style though. It can be prompted for, but an SREF would save prompt tokens and improve consistency between images. Dreamwave Horizon comes close.
I haven't seen a Pixel art one yet!
Dude!!!! Craftlandia!!! This has to be one of my all-time favorites instantly!!!! Do you remember this prompt you shared: pilaf hypnagogic maximalism physicsbased paper collageaem, in the style of suburban flat shading, simplified figures, tantric manuscript, studyplace, expert draftsmanship, sculpture, oil pencil sketch, diagramic? Well let's just say if you have spend a little time crafting a little Cocktail with this prompt and Craftlandia, you get some amazing crazy wild images 🤣🤣 You're a legend man, thanks so much!
hahah oh jeeze, okay I'll try that now!
very cool!!
Hi Nolan! This was amazing! Thanks for your shoutout on your catalog!! I've used all that you found, its really great! ("Prism pop" i use a lot!!) Thank you.. do you think you'll keep adding new ones you find on the catalog?
Maybe not adding to the catalogue but I'll definitely consider making a new one!
@@FutureTechPilot Fantastic, Thanks!!
Love the show, Images are also breathtaking, my own problem is that Batman and Harley gets too many shots and Superman or Hulk nuttin'
😂 I'll try to spread the love around
Thanks for sharing these codes and the pdf Nolan - style code mining is fun! How do you come up with names for the style codes?
Some will come to me naturally, but I like sending a bunch of pictures to ChatGPT and asking it for like 20 possible style names - most suck, but some are brilliant!
love ref codes. does anyone know a resource/library with an overiview of like a lot of different styles?
This is a pretty great website - www.stylesafari.xyz/
Thanks for sharing, btw i noticed image generators can calculate and translate laguage. Prompt: an image of 2+2= translate in french
can you elaborate?? I'm not sure I understand
@@FutureTechPilot put my prompt in ideogram, you see
Can I get the thumbnail image please🥺 It's so gorgeous😍
Just try the quantum dynasty code with a character reference!!
@@FutureTechPilot thanks a lot man 🙏🏻
Where did you get these codes? Did you invest time in developing these yourself? Or did you copy them from discord?
I used --sref random!
Any style code for image like basquiat art style?
Not that I know of off the top of my head!
Hi Nolan, You know what SREF I also like because it's fun and very simple to remember?!? --sref 10 (its a poppy colorful lego style)
haha I'll have to try it out!!
-sref instantaneous subscriptilius
You really like your nightmare fuel, eh? Really getting into the Halloween mood there.
😂 baha the rest of the catalogue is much more tame I think
Hi 👋
🤝
I think showcasing --sref without an explicit --sw is not very helpful. As we know, --s is the style persona strength (applies the --p nnnn style persona or the generic style persona if --p is not provided), and --sw is the (style) reference strength. If you use --sref, the default --sw is 100 but only with --sw 1000 you can set the full strength and quality of the sref.
It's annoying how Midjourney uses the term style in 3 completely separate ways: --style is basically a different variant of the model. --s is the strength of the --p style persona or of the generic persona if --p is not provided. And --sw is the strength of --sref style reference. Style persona and style reference are two very different factors that influence the output.
Style reference is very direct and kind of copies the style (including colors, textures and details) of one particular "image". Style persona is more general (because it's a set of preferences gathered across many images liked by a particular user) and affects the composition first before influencing the details.
In my experience, --sw 1000 isn't *that* much different than --sw 100
@@FutureTechPilot How odd. In my experience the difference is massive, especially with very low --s of course
--sref random is annoying as it finds the same "Randoms" all the time.
ya a lot of styles look similiar!
Great vid bro 💯
Cheers!!