NEW Midjourney Update is Absolutely Amazing - Version 5.1 Full Breakdown
ฝัง
- เผยแพร่เมื่อ 12 มิ.ย. 2024
- Don't you love when you get to wake up to a Midjourney update? Take a look at Version 5.1 and everything it has to offer. I show off the stylize value and some direct comparison between versions. There is also a 'raw' mode available in 5.1 but that is very similar to regular version 5.
Free PDFs:
cameralens.futuretechpilot.com/ 200styles.futuretechpilot.com/ cheatsheet.futuretechpilot.com/
Links:
Merch - FutureTechPilot.threadless.com
Discord - / discord
Creative Fabrica - www.creativefabrica.com/spark...
My MJ Gallery - www.midjourney.com/app/users/...
![[PoE 3.24] Synthesis Implicit Modifier Weights REVEALED | The Greatest Mystery in Crafting Solved?](/img/n.gif)
I love how excited you are about all the results. Makes it super engaging
haha I appreciate the feedback! Thanks for the comment and I'm glad you enjoy the vids
What a time to be alive!
Exactly!
"Wooo, look there! Look here!! Look at THIS!!" - like a kid in the toy shop 🙂... same feeling here, mate.
haha I'm glad I'm not the only one, cheers buddy
Thank you for all your wonderful content
I appreciate the kind words pal!
Played with this last night and feel like if you generate images with a human figure 5.1 gives you more consistency of the person's faces across all the generated images and from prompt to prompt as long as their description is consistent. Seems it could be a big help for those seeking to do things like visual storytelling where consistent characters are needed and was hard to achieve in the vanilla v5.
Really? I'll have to try that out myself right now. Thanks for the tip!
Well done on presenting the styles and output !
I'm do glad i live in this era...and grew up with the 80's rock.
haha sounds like you've seen it all! Thanks for the kind words :)
Thanks! I'm straight on to it!
haha enjoy!
As someone who has spent their entire life, creating art and making art, although these images are really appealing to look at, it’s extremely sad to think about all the incredibly talented artists who spend 8-10 hour days mastering difficult skills and applying the talent, having to contend with this.
I hope they can learn to master this as well
Now imagine what those incredibly talented artists could do if they use these systems as personal assistants.
I encourage those artists to adapt
@yeah but it's just an art slot machine. You invision something then try to describe it and settle for what it comes up with. They need to make it so you can have full control. The seed thing doesn't work at all and the moderator bot is triggered too often.
Reminds me of using new aged recording software. Has almost unlimited sounds and synths but they all kind of sound the same. Everything Midjourney comes up with has that same thing going on, it's different but there's something the same about them all. Can't put my finger on it.
I enjoyed this video mostly because of you getting so excited about all the art, which made me excited about all the art 😂
haha I'm really glad to hear that! Cheers :)
my cousin got me to watch your channel… she calls you the “midjourney uncle” 😂 … very cool stuff
😂 somehow I think that's pretty accurate
Your videos are so much helpful. Thank you for your time that you spent researching and trying this cool stuff. Can you do how to do get consistent characters for a graphic novel or something similar?
It's funny, I finally filmed a consistent character video this morning and it will be posted on Monday. However, I'll let you know now, I advise against trying to get consistent cartoon/anime characters. It just won't work. But since you asked, I'll spend some extra time on it and see if I can find any new tricks
@@FutureTechPilot thank you! I tried to get some consistent characters and i did but it gets tricky when you try to get multiple characters together. Looking forward to your videos
You are really one of the best reaction channels! 😁
haha I appreciate the kind words
tried it this morning with --style raw, loved the results
Midjourney updates are the best
Poetry pormpts are deffinetly gonna kill a lot of artists careers. It used to be you could only do straight forward prompts with mid quality, but this is a total game changer.
Maybe artists will be the one to use it!
Playing with the poetry ones with some stuff i vaguely remember from high school. Robert Frosts line "The woods are lovely, dark and deep, But I have promises to keep, And miles to go before I sleep" gave ma something spectacular. Really fun to see what happens.
@@FutureTechPilot of course..my butcher has become an artist now..she is so proud of her work...like the police man with the stick, he put prompts in with it...wow
Can't wait to create new content with 5.1!! Thanks for the great rundown! :)
haha cheers!! Can't wait to see your art with it
I find in 5.1 the wide aspect ratios work really well if you are depicting more scenic stuff, instead of specific subjects
Yeah? That's a really good point!
@@FutureTechPilot Yes sir, try something like this one, the outputs are gorgeous in landscape
'Capture a stunning rooftop garden in a futuristic cityscape, showcasing lush greenery juxtaposed against the metallic and concrete structures of the city. Emphasize the vibrant colors of the plants, illuminated by glowing neon lights and holographic projections. Use a Canon EOS R6 camera with a Canon RF 24-105mm F4 L IS USM lens, set to an aperture of f/5.6, ISO 400, and a shutter speed of 1/60s --ar 16:9'
It really is insane... I can't believe as a designer.. after all those donkey years.. I have to now learn to prompt! Those images are just mind blowing... it's overwhelming!
haha the world changes fast eh?
Thanks for all your great work and dedication on this journey through light-painting; for sharing your open-hearted, beginners-mind and contagious enthusiasm with us. Engaging. Inspiring. Poeticality.
Thanks so much for the kind words!! I'm grateful for all of you and I appreciate when you share my excitement.
Great again, go on. I don't know why you got "only" 16k. Go on, it will work. I do enjoy watching to compare styles, ideas and so on...AND I do write your information even down :-)
Thanks so much for the comment! I really appreciate it and I'm glad you like the videos
Wow, watching this video made me feel like I've entered the world of science fiction! 5.1 is truly amazing and the possibilities for creative expression are endless. Thanks for sharing!
haha I definitely know what you mean
The 'poetry prompts'! 🤯
That, is, awesome! Thank you Nolan. 🙏 My entire paradigm had been shifted to longer, weirder prompts because of v5. Have been playing with them since yesterday and yeah, it does seem like the widescreen ratios need some love! And as David Wiles in the comments has said: I feel 5.1 is allowing more 'creativity' and has a less robotic and inorganic feel to it.
It's a super weird feeling going back to short prompts haha but it's good that there is still the RAW mode if we need
Cool video! 👍
Thanks pal!
This is awesome, now I have something to blow my fast hours on ;) Thanks!
Our fast hours will never be safe for long lol
Thanks!
and thank you!!
Thanks for comparing the 3 different versions. Very inspirational again!
Glad you liked it!
Thanks so much for your knowledge and enthusiasm i sharing Midjourney updates and how tos and explorations. You are so fun to watch and listen too! Was there a reason you showed --ar 4:3 --ar 16:9 or -ar 16:9 --ar 16:9 with two aspect ratios back to back?
Thanks again for all you do here!
thanks a lot for the kind words!
And the extra aspect ratio is just because I have 16:9 set as my suffix. So it automatically gets added to all of my prompts. However, when you manually write one into your prompt, it will show up first and become the ratio that gets used
Thanks dude, super helpful for me PLUS Darth Thanos 🤯
😂 yeaah pretty cool for sure
the album cover for my new song WYNA was created with Midjourney! thanks for the video bro i just subbed! tonight i release the 5 minute + animation made with another, newer tech
That sounds sick buddy!
This is crazy. I was working a prompt this morning and unhappy with it, but after watching this tried 5.1 and it blew me away!
haha that's so cool. I'm glad you're liking 5.1 as much as I am
Ooo more toys to play with 😊back to say,Wow! It’s absolutely breathtaking
The rabbit hole never ends!
1:00
😆 🤣 😂 I love that prompt
haha yeah that one came out of no where
Miidjourney is back, yes!
💯
Wowwww man, #NEXTLEVEL!!
Yeah it's kinda mind blowing, eh?
Neon Batman is lit.
Yeeee
Gushing for good reason 😎
I dived into the pond only a month ago, with Playground, Leonardo, Dalle2 (via Bing) and SDXL on the Dream Studio Beta.
Have avoided MJ due to the quirky pathway, it's time to reconsider my options, amazing results 👍
You’re missing out
I was the same as you. MJ is all I’ve been using for months now. SO good.
@@goilo888 With 1.5 M users online typically at anytime, that's a lotta paying customers,
conservatively if most have the basic $10/month sub, that's a lotta monthly generated income!
BTW, I have not been able to try any FREE renders to try it out, are they getting greedy? or is it a bandwidth issue?
Either way, will probably take the plunge once I sort out my monthly $ commitments 🍺
@@SergeGolikov They stopped the free renders. Too many people taking advantage with multiple email accounts. I’m only on the Basic plan. When I start getting low I just do a lot of Image Ranking and then get a free hour :)
Just seeing this comment now for some reason, but I hope you've enjoyed your introduction to Midjourney!
Man, that poetry prompting set me off down a rabbit hole while I was watching this video 😅….. Side note, I wouldn’t mind betting that the paperback novel cover creation market has radically changed with the introduction of AI art.
haha I'm glad I'm not the only one who likes the poetry in Midjourney. And yeah I think lots of job markets have had to change recently
5.1 is impressive. I can't believe the results I'm getting from simple prompts.
It's super fun to use eh?
11:41
Yes, I can see that!
😂 I knew I could count on you
You are so funny, very entertaining. Thank you.
Where can I find a PDF or text information about 5.1 prompting so I can feed it to chat GPT
you'll have to search the docs and discord, put in a bit of work and learn it yourself too
I'm not sure if there is any documentation yet! I'd ask around on Discord, usually the Guides will put something together asap
WOW! My jaws dropped! 😲👍That was very impressive!! I've noticed that whenever you put "Chaos together with Stylize" in your prompts, the images are very tidy and perfectly clean and impressive. And you only have the setting for the "base quality" and the "Medium Style"... Ai fast evolution News drive my mind crazy each week.😜By the way why you put 2 different aspect ratios at the same prompt, is there a reason? Also I want to add that version 5 Understand way better the "White Background" than vesion 4. It gives me very clean white backgrounds without the vignette effect of v4. The only thing Midjourney can not do yet, were Adobe firefly can, is to generate Countries Flags.
haha thanks a lot for your thoughts!
The second aspect ratio in my prompt is just because I have it set as a suffix. So it shows up at the end of all my prompts. But when you manually place another aspect ratio in your prompt, it will appear first and be taken as more important.
Good to know about the white backgrounds, and I've never tried country flags haha I'm surprised to hear it's not good at them
Te sigo desde hace meses y me gustan tus vídeos. Pero creo que no deja en claro los cambios en la versión 5.1. Me gustaría que hubieras usado exactamente los mismos prompts de la version 3, 4 o 5 para ver Las diferencias más claras.
Son imágenes que ya tienes y que fácilmente puedes hacer comparativos.
Sigue con tu excelente trabajo.
Saludos desde Mx
I agree, I should have done a better comparison. I knew I forgot something. Thank you for the comment
I been playing with 5.1. I gives really good images. It can do 2 people really well.
Do you have any tips? I haven't had the best success with multiple people
If you add a seed value to the prompt on the first images you make you can use that number as the seed for rerolls or versions. But where does everybody get those stylize words from? Im seeing a lot of (term)core stuff going around is there a dictionary somewhere?
You can download my Free PDF which has a lot of 'core' styles in it - 200styles.futuretechpilot.com/ but I think people are finding them by using the /describe feature in Midjourney
In version 5.0 I was noticing that the generations with -ar 1:1 were having more quality and complexity to them, than 2:3 for example. And when I tried 2:3 there were more portraits and less detailed backgrounds. I dont know if its the case now.
No I think you accurately described the general differences between the aspect ratios!
Question: Is it possible to 'Save' specific images & their prompts? At the moment i am having to scroll up the entire 'Back Catalogue' of my creations, then download an image, and copy and paste the prompt into a new request....any ideas please?
Sorry pal, that's a pretty common problem and I haven't seen a good solution. Lots of people take notes on their computers; I use the website to look through my old prompts. I'm not sure if there is an easier way to do it, so let me know if you find anything!
9:20
👍 Actually. aspect ratio could
typically make a huge difference in quality of final render since at least v4.......at least with most of the prompts I have experimented with.
I guess there is always something I miss haha I hadn't realllly noticed until now
Dude! you have to make a short comic strip, how Thanos became Darth Thanos… 🎇 friggin awesome outputs
Great idea!! haha I can't wait for Midjourney to have some consistency tools in the future
@@FutureTechPilot well for now, the parentheses bracket can help make a batch of sequences if you plan the shots, it might be something as simple as : Thanos is transforming at an alarming rate into darker version of himself { close up shot, Ariel view, long shot },{muscles, muscular abnormalities, muscle bound beast} in dark wilderness . -ar {1:1, 16:9, 9:16}
@@pongtrometer That's a great idea! It'll eat up fast hours like crazy though lol
@@FutureTechPilot yeah 😅 I had to cut down the batch from 27 to 6 outputs 🤣 I haven’t got MJ money yet 🤑
Try bioluminescent blue with Batman and also add neon to the prompt. (bioluminescent neon blue batman --ar 9:16)
woah. Unreal tip lol I can't wait to try it. Thanks!
I am not showing RAW MODE in my settings, I was told to add --style raw, lets see if that works.
strange! I just checked and it's there for me. Right beside 5.1 ... let me know if you figure it out
This one is crazy.
Try to imagine what it will look like next year ... lol
Do I still need an active Midjourney subscription if i’m running it locally on my machine ?
I didn't think it was possible to run it locally ... I guess you should ask 'member support' on their discord channel
As someone who just happened upon your channel and someone new to MJ, I honestly ask you... this is amazing, that's for damn sure, but what do you do with these images other than for personal fun? Are there people who are buying these images or companies? If so, who? Specifically who? And where would you sell these images? Sure you can make logos, stickers, tshirts, mugs and all kinds of stuff for etsy, printify, but what about these complex images of an atom bomb, etc? What other applications other than say maybe wall art or posters?
thanks for the vid sir, cheers.
I don't personally sell my pictures but maybe I would one day. To be honest, I generate pictures because I enjoy them, but also to learn what Midjourney is capable of so I can teach other people.
Sorry I'm not more help lol you asked a lot of questions that I don't think of often
@@FutureTechPilot lol. All good man. Was just totally wondering... I love MJ and it's capabilities. I can't imagine what we're in for in the next year when it comes to AI. Mindblowing in what we've been able to accomplish in such a short time. I appreciate your time.
I did get a word to spell accurately with v5.1 !
Amazing lol
hahah awesome!
in 50 to 100 years you wont search videos on youtube, you are gona type in what you wanna see and get that😶
hahah I wouldn't be surprised if that happens in 3 years lol
Has anyone worked out how to prompt for stickers without the white surrounding border?
You could try something like --no border, mockup
Is raw mode only out for certain accounts? I sitll haven't gotten it yet. (Edit: typing style raw helps me use it)
@@marceloregent5003 oh ok. Thanks !
Strange! Did it show up?
@@FutureTechPilot Yeah it showed up after I saw marcelo's comment.
what iseed and number after that . i tried the same prompt ( i guess it is a prompt ) but got a message invalid prompt
Every time you make a generation, Midjourney selects a random seed for you. But you can write the seed you want in your prompt if you want repeatable pictures. You should ask one of the 'prompt chat' channels on discord
Regarding Text Try 4 letters or less. I am getting 80% success in generating 4 letter words in the V5 and above. They may need to be common words MJ may have seen. I bet we are a month away from near perfect text like Floyd!!
4 letters definitely seems like the sweet spot!
I'm waiting for the text thing and I reprise my pay plan 😂
I can't wait for that feature either!
5.1 is better and is more "creative" with the images that it creates. 5.0 always struck me as a little too sterile and lifeless, like Stability Diffusion. However 5.1 is overly sensitive to uploaded images that it deems as NSFW. Some of the images that I uploaded for /describe earlier this week and worked perfectly are now deemed NSFW.
_Exactly_ my thoughts and experience with it. 🙂
Strange! I guess it's hard to iron out the quirks before making it available to everyone
That means: sophisticated prompt not works any more?
You can change the settings to 'RAW' mode and that's where longer prompts work better
deep content
cheers pal
this is a basic question. Are you using a paid account or the free version of midjourney?
I have a paid subscription
@@FutureTechPilot thanks. That’s helpful to know. I will look into getting a subscription. Cheers.
what are seeds? never seen those before
Each prompt gets generated from a random seed. There are over 4 billion seeds (4 billion pictures it could make)
but you can specify a seed number and get a repeated set of pictures, so that allows you to alter the prompt a little bit and get the best representation of how it changes the generation
Have you tested personalized images. My goal is to make fantasy images of people.
hmmm .. I think if it's a celebrity, it has a good chance to work because their face is probably in the dataset ... but for a random person? I don't think it does a very good job at all lol
Holy crap.
I love updates haha
I hope 1 day midjourney can make airplanes....fighter jets.
haha I'll keep my fingers crossed for you buddy
Dusty, evil princess mummy chained up in the Tomb of Imhotep, soft God Ray light --style raw --s 76 --v 5.1
I'll try it out!
The resolution still needs to improve.
I think it's a priority for them
More greeting cards 🎉😂
🤣 oh yahh
Oh... OH. oh.
😂
Awe so we don't need "I can't believe how beautiful this is" anymore!!? 😂
😂 the life and death of a midjourney strategy
I can't even get a bull and a bear in the same picture with it.. something so simple but fails 95% of the time 🥴
haha it definitely has trouble with multiple objects/subjects mentioned in a prompt! It likes to blend them together for sure
What is this seed number thingy ?
The seed number just lets you get a repeatable picture. If you don't specify a seed number, Midjourney will use a random seed every time
@@FutureTechPilot I got it thanks
If i want to create realistic people, like a family at the beach. Or a pregnant lady. Or a selfie. So not realy creative. Should i use 5 or 5.1
5 but check out 5.1 --style raw (5.1 RAW is close to version 5)
the subtitles is disabled
Strange! Sorry about that
Very nice. I just wish it was a proper web app and not a Discord channel.
If you have generated over 10K images, you have access to the Alpha Creation Website
👋
🤝
12:00
Yes, for 99% of the images I have MJ render for me, I use a chaos of no higher than 14.
That's because I already have a good idea in my mind of what I want MJ to produce.
That said, there is a case to be made for very high chaos such as where I've very little clue as to what I want. So just now I prompted MJ to render:
"Imagine the most beautiful bizarre multicolored
city-nature landscape
+ Trippy-Esoteric-Fantasy-Fractal-Tryptamine-Astral-Hallucination-DMT-psychedelicCore + I can't believe how beautiful this is + Extreme Detail + High Dynamic Range saturated luminosity-colors
--c 100 --v 5.1"
I did it once and the results were quite awesome and I'd very little preconceptions of what results I'd find satisfactory. Chaos 100 seemed okay here.
hahah so maybe the crazier the prompt, the more chaos makes sense .... maybe?
@@FutureTechPilot I'd say, the more vague a prompt combined with the less of a concrete notion of what the prompt could mean is call for higher chaos
Still can't deal with display text, disappointing
One day!
6:40
Yummy!
Digital Delicious
is this free? $$
haha I wish buddy
Does AI just do hot young chicks? I can't see it being much use it thats all it does? Ok..it does planets, great.
it can do basically anything! I love the cars for sure
if every monkey can create a fantastic art (stolen from all artists in the world) ..where all this make any sense? i cant get it...
It's a crazy world
It’s interesting how you call “cool and creative” images that are mostly repetitive and mainstream. So far, I haven’t seen (though waiting for) true creativity. I see a great deal of remixing of images that are vastly and abundantly available on the web, sometimes even royalty-free. Since there are tons of ‘artists’ that can produce (with effort, sure, but all the same) images that look exactly like many others, I ask myself: where is creativity heading to? Is to be creative to know and use well the tool, then producing more or less the same output, or is it to produce some ground breaking striking images that are also dissonant, maybe even a bit disturbing, thought provoking? I don’t see yet a problem with true artists being at stake here. I don’t see any Giger, or Picasso, or John Eaves or Matt Jeffries. I see a lot of mundane and boring mainstream imaginary (because that is the biggest part of the dataset midjourney and others feed on). Not sure of course of what the future holds.. but for now.. well..
I wouldn't bet against A.I
5.1 is not good.
haha beauty is in the eye of the beholder!
Hmm…why only young girls? 😶
I think there is more Thanos in this than anyone
Changing the DPI also produces amazing results.
What do ya mean? What's an example I can try
yes I agree is nice to use Midjourney very good quality even that there other sites which can be used free of charge
💯
When are we going to get nudity.
Never, Stable diffusion is the way
Isn't there enough porn around already if you need it?
Art is spirit lifting, porn is destructive.
Probably never from Midjourney
Not talking about porn.
Hey buddy, I hope u understand that you are not a real artist or designer or concept designer. Just a reality check for u. 😊
When did I call myself an artist?
Thanks for the amazing examples. I’m hoping that ‘describe’ will still work since I’ve been using it about 70% of the time with my own art and getting killer prompts and freakin awesome art. I haven’t seen anyone using it the way I do though. Keep on making awesome videos, every one is inspiring.
What do you mean by "describe"? Not familiar with this feature
@@olherb9633Use /describe and then upload an image. Midjourney will generate 4 different descriptions for your image, like it's trying to guess a prompt that would create a similar output. It's a neat feature! Especially if you feed your own previous generations into it, might provide inspiration for new prompt ideas!!
@@olherb9633 /describe will let you upload an image and give you four prompts as it describes what it thinks it is. You get 16 images. At least one will be useable and the prompt it produced provides a lot of keywords I bet you never thought of.
Thanks for the kind words! Do you have any tips on how to use describe? Do you image prompt with it or change some of the words before you generate?
@@FutureTechPilot I input an image and just run all four suggestions. If I get one that’s close to what I want I tweak it and re-roll it until I’m satisfied. One of my tricks is to just put in one of my hundreds of sketches and let it do its thing. I’ve gotten the most amazing pix with simple line sketches. I had no idea that was even possible until I decided to give it a try. Isn’t David working on something similar for future generations? I can promise him it works. I wish there was some way to share a few that I’ve done. You’d be AMAZED at how cool they’ve turned out.