Chapters:
00:00 What is the anatomy of a prompt?
01:24 My biggest tip for modifying MidJourney Prompts
05:44 Using an “image source” with MidJourney
09:52 Writing & Stylizing a Text Description for your Prompt
29:01 Using Parameters & Arguments to modify your MidJourney Render
49:24 Using image seeds with MidJourney
What makes up the anatomy of an AI image prompt? In this second episode of my MidJourney Master Class, we cover the basics of promptcrafting – image prompts, fleshing out your wording with styles and inspirations, as well as parameters. In the next episode we’ll look at tools to help you craft prompts even easier, if you want a “do it for you” tool. This video contains a bunch of MidJourney Tips and Tricks, MidJourney Prompt tips, and how to get the exact image you want with MidJourney AI bot.
Buy my art: https://analogdreams.threadless.com
MidJourney Master Class – https://www.youtube.com/watch?v=FcRFQTMW_bo&list=PLgrJUN7TszTFNO9P6pNnXlS28Sw9rrg9W
Referenced prompt documentation:
MidJourney Docs – https://midjourney.gitbook.io/docs/
MidJourney Styles & Keyword References by WillWulfken – https://github.com/willwulfken/MidJourney-Styles-and-Keywords-Reference
Prompt Template Google Doc – https://docs.google.com/document/d/e/2PACX-1vRHOxyEb-ERGi-BdZM8Z_piEP54m4HwO0z8scjmEurEp2UZVA6rFxvyKd15elYVHUWfP1oSA4CQFwxr/pub?utm_source=docs.google.com&utm_medium=tutorial&utm_campaign=midjourney
Alt – https://docs.google.com/document/d/13c8Ci-8kU2PVZu6DKghlhOOrbf4kmtc9xxCJAnPqvC0/edit
Text in MidJourney – https://media.discordapp.net/attachments/992207085146222713/1010597544868458636/midjourney-text-placement-suggestion-shambibble.jpg?width=1596&height=905
Style Guides
https://docs.google.com/document/d/e/2PACX-1vSIWlr5F6J4yv4jxazaXcZaI-F3MxCsnSq4jwoaO3vaEYC911v2N7JpDgjuhAUVYhq4VmA1SCjF3rdK/pub
https://docs.google.com/spreadsheets/d/10PSLpFXiQiAiqsDQImW3f4Vrn6wYaNQL3ihHFeGZZ-M/edit#gid=0
https://docs.google.com/spreadsheets/d/117kRRXZFYkRM-QFt7yt6hRLQrg0n3mAMvk7RY3JyXhQ/edit#gid=1572798432
https://cattail.nu/midjourney/
https://docs.google.com/spreadsheets/d/1cm6239gw1XvvDMRtazV6txa9pnejpKkM5z24wRhhFz0/edit#gid=438712621
————————-
📧 Biz Email: eposvox@standard.tv
📬 Shipping: 303 E Court Ave # 459 Jeffersonville, IN 47130
👕 Buy my art: https://analogdreams.threadless.com
💻 Discord Server: https://discord.gg/eposvox
✈ Early & EXCLUSIVE content: https://curiositystream.com/epos
🎶 Music by Backing Track: https://backingtrack.gg/
📈 Patreon: https://www.patreon.com/glitchart
💵 Direct Tips: http://paypal.me/eposvox
💾 Main tech/streaming education channel: https://youtube.com/eposvox
📼 Alt Channel/Clips: https://www.youtube.com/channel/UC0n9X89gHg1kayykR6e3q8Q
🎮 Gaming channel: https://www.youtube.com/channel/UC-TRvAafLbCHVcszm7d8W-g
🎥 VODs Channel: https://youtube.com/eposvoxgaming
🐦 Twitter: https://twitter.com/eposvox
📷 Instagram: https://instagram.com/eposvox
📼 Twitch: https://twitch.tv/eposvox
#midjourney #aiart #aiartist
Very good video show. Thanks for the quote!
That v4 beta though lol
even i have no idea about idea about editing ,i start getting some very good results
thnks so much for infos
Hey Addie, not sure if this was intentional or not but your audio does cut out every few minutes for a few seconds at a time. First time it happened was around 6:30. Thanks for the awesome video though!
I often use 9:19 for a smartphone aspect ratio, those are actually the best for full body character shots
Look into
GFPGAN_inference.ipynb
I use it to clean up faces from midJ, they look so amazing afterwards, its meant to low res faces (like you were saying theirs don't look right at 100%)
Great video thanks, looking forward to the rest in the series
Super helpful! I just started on my journey 3 days ago. I am already 500+ imagine in. This video help clarify a few things very well. Idea for future video, if you haven't made it yet, is how get get specific text in the image we imagine.
DUDE!!!!!!!!!!! many many Thanks.. now i just have to work on whats best for head shots.
whats up with the shitty audio man
Could anyone tell me how I get a 4×1 grid instead the ordinary 2×2 (for the 4 initial images)?
Great info, distracting context. The impression that you are doing half your masterclass wearing wetsuit long johns. The exposure to typically untanned scuba divers' shoulders distracted from your message about making great images🙄/imagine shirt?
fuedal?
Hey, I find it really helpull, thx for all the effort you put making this series
Now if only I could figure out how to get a full body every time instead of a dang portrait!
Wow, excellent video. Here's a tip… windows settings> system> clipboard>clipboard history ON.
Then you can use Win+V to paste from a LIST of previously copied items!
Best prompt is "Jet Black haired Christina Hendricks as beautiful goth girl, with fit thick hourglass body, with huge chest, in skin-tight clothes, painted by artgerm, Alphonse mucha, Akihiko yoshida, sakimichan, krenz cushart, low angle shot, digital painting –ar 9:16 –test –upbeta"
Don't fight with me.
The new update is really good with the Test, Testp and creative modes. I don't even know what creative does but I usually give it a try if I get a promo that gives me something recognizable
This shit has totally fucked up my Google history. It's all best hot anime artist and now I follow pinup artists that draw mostly naked anime chick's because they have the human form and a interesting fluid pose and theme a lot of the time that the AI can use as a style reference. The other half of my searches are most recognized hot women and dudes in the world so it has a better chance to recognize them. Or at least this is my a rational excuse. I feel like hentai is the logical next step. I blame the AI. Turn back from the light!
Thank you for this awesome free Master Class. Many other yt would monetize it on udemy
Is the reference image possible with a locally running stable diffusion?
Is that style and keyword list also available for stable diffusion? Does it also apply there?
Everytime Iuse a URL it says it's an "unrecognized argument"… anyone can help?
any suggestions on merging/mashing/combining animals together? Example:50% spider 25% starfish 25% octopus
This really needs a great UI. Sliders and dials for everything so you don't have to key stuff out. Also add a Save As/Import/Export Params option. I am hooked on this and look forward to it producing poster size 8K images.
This part 12:50
Can I use an picture of my face or somebody else?
Thanks for your tutorial. Very helpful! 👍👍
I wonder how many artists will see this with mounting dread thinking about their life-bread literally stolen from them by a machine?
Recently noticed that Midjourney is used with Photoshop, Gimp or even Krita, to finish up their renders.
How would you create a superwide image or animation like for instance 'infinitief zoom'?
Excellent
This is the most fleshed out explanation I've seen. I'm just looking to make some references for airbrush art.
To get a "cinemascope" wide screen image of 2.4:1 you just need to type –ar 24:10 I also like –ar 23:9
Thanks for this, been looking for something like this. Can you do one for Stable Diffusion?
great video!!! can you do art of a word or phrase?
Ai art can be very complex and daunting at first. I have been toying around with it in the last couple of days but had no idea about these settings. Great video, beneficial tips.
Hey it would be nice if we can see a close-up view of your prompts while typing . Thanks 🙂
amazing
/Imagine spelling feudal correctly. Just offering to help. It might have had an impact on the output…
how dare you airquote "artistry" sir
This was super inspiring and informative. Thank you 👍
AI art can't be copyrighted so you don't actually own the exclusive rights to use those images for commercial purposes didn't put any labor into creating it.
This is the legal standards in the US and UK.
Australia and South Africa allow patent and trademark of AI generated images. However a patent and a trademark are not the same as a copyright.
So using those images on your merch is kind of risky if you're an the countries that don't support it.
Can you sell these arts on merchandise or web sites like Etsy?
Thank you So much for ur ti and support
You never know how much you have in common with so people.
RAD video! Really needed this. Thanks!!!!
and just pay more attension. I envy you mate.
!
what if i wanna save a project but i use free trial? can soone help
ong
"feudal."
Learned so much from this! (Probably missed so much too 🙂) Thank you!
video's.
Really enjoyed this .. I subbed.
Thank you 🙂
yeah
QUESTION! Do u have to do — or can it be —
I use my iPad & phone. So With — I have to put a space and then toggle in front of the 2nd dash and backspace to get that exact look but with — I just do 2 dashes in a row, no spaces…
Can we use the Job ID for anything?
Amongus Style lmao
Best MJ explainer video on YouTube, nice work
This was a great tutorial!!!
Do they ever plan on breaking out from Discord? And putting this into their own platform? Because I can’t stand the discord interface and searching frantically for my renders.
OMG this is what I was looking for!
Fantastic tutorial. Thanks so much for all this.
5:51 "1 or 2" I've done as many as 20+ . In the "Image Prompts ONLY" thread most of us do 1 or 2, yeah. 5 is hectic. 8:08 I've never come across any url references that weren't from a MJ generation 🙂
From fractl (Mod) == "When you only have one image, the ai can’t infer what about the image you want (the subject, the style, the composition)? I suggest adding a few images with something in common to re-enforce the part you want while showing that other contexts can be flexible"
Another note is DavidH has said that MJ is better at interpreting AI generated images.
12:46 26:37 Back in May, I read from AmirKerr (Mod) that "In order to use these render engines you should paste the name of render engine before your prompt" but I have never, ever seen anyone ever do that lol.
Thank you so much for this awesome teaching video and the links you have provided. I just got my subscription with Midjourney and I am so happy I did. I do tee shirt designs on Merch by Amazon, Redbubble as well as other things. I am having a problem with how to write a prompt for a tee shirt design. How do I write a prompt to show just the image so I can put it on a shirt without a background? It gave me a background color with the image, and I was having a problem trying to erase the background without erasing part of the design. Any helpful information on how to do this – will be greatly appreciated.
My friend, you have a gift for teaching!! I wish all YouTubers taught as well as you do. You took something confusing and made it make so much sense.
Re Video Thumbnail: Did you notice the guy on the left has 3 legs? 🤔
Thank you, you are awesome, what a great teacher! I am having so much fun with mid journey and you are helping that process be even more fun. Thank you thank you thank you.❤
Just saw "The ULTIMATE Beginner's Guide to MidJourney AI Art – MidJourney Tutorial & Explained Part 1" but where da heck is your part 2 to that?
Great video, dude! Very helpful as a beginner!
You can't do 2.35:1 but you can do 235:100
Best Result u have when input: "one color Wallpaper"….its awesome.
cubical
I do alot of post Photoshop work and I love MidJourney. Where can I sell my art?
With Topaz AI you can get the movie screen lol
You are incredible….. watched the first two videos… looking forward to the others… so much to digest. Thank you!
Thanks
This might Has anyone figured out prompting specific words and call font styles? I’m late to the game but I mad3 it already doing ridiculous artwork series’ what dangerous rabbit hole this tech is.
🔥 Great video learnt a lot
….. to make better people say detailed face, and detailed body's explain what type of body type it is hyper detailed body or face too helps render full figures
I made the whole matrix crew referencing there real names [name] works out like a dream ..SO life like! ♾
it's really cool! Thank you!
This doesn't help newbies. It's a waste of time. Been on for half the day and not one prompt has worked.
I'm realizing more and more that Elden Ring is basically an AI generated game
add — video in your prompt and send envelope after – you'll get the video this way
This is just amazing, I don’t have enough thanks for you ❤
Lots of great info. Thanks! Looking forward to watching the other parts of the Master Class.
17:23
that's what i was looking for. Thanks bro 🙌
Is there a way to recreate the same ai generated character we create and get it in different poses – so we can create a story. Similarly would like to save an ai created wallpaper and add in a saved ai character, get Midjourney to combine the two? Is that possible?
Thanks for the videos I think I will be watch the rest of this series…
Thank you so much 🙂 very helpful. You are awesome!!
Great video- great sound! What mic are you using at the start? It's not even visible and the audio is great!
how did you get a dark mode in the github. my eyeshurt from white background??