In this video I cover:
Timecodes:
0:00 – Intro
0:48 – Style of Studio Ghibli
1:20 – Text prompt
2:52 – Midjourney image database
3:22 – Text prompt revision – character details
4:35 – Text prompt revision – quality improvement
5:45 – Text prompt revision – fix background
6:35 – Concluding tips!
Part 2 of this Text Prompt Guide series with Midjourney AI. Create your own character design using AI (Artificial Intelligence).
Episode 1 link: https://youtu.be/2QTXTp-_5JY
Join my Discord, lets create together!
https://discord.gg/QRkMf5aJqD
Follow me on Instagram:
https://www.instagram.com/cyankatana/
Follow me on Twitter:
https://twitter.com/cyankatana
Links mentioned in this video:
https://haveibeentrained.com/ – Midjourney Image Database
https://www.midjourney.com/home/
https://midjourney.gitbook.io/docs/
Like and Subscribe for more videos on Midjourney AI!
I'm first
Fantastic video subbed for more really enjoyed this
My man! Great video, succinct and useful as always!
nice content!
very informational, thanks
Underestimated channel. +111111111111111
Great Inputs, Step by Step WE are there together
is it possible to generate different emotion/movement or place the character in different environment that we generate?
Fantastic insight to prompting. Thank you.
I did not get the same results using the exact same prompts, even after re-rolls (6:15). Even with "Unreal Engine" it shows mostly 2D characters. I'll add other 3D renderers to increase the probability. But the results are not clean and consistent. It's important people know that.
While I do not agree that placing the art style at the front achieves the best results, I do fully agree that the sequence of placement of words and phrases does matter.
The Midjourney Image Database site is very helpful for users to double check if midjourney understands the words and phrases we are using correctly. A spacing can change the entire meaning of the prompt.
E.g.
Liu Yi Fei = Old man Photograph
Liu yifei = Chinese Celebrity
If we just could create the same personw ith different emotional faces we could create comics
Studio Jeeblee
can you screen rec a live workflow? thanks so much for your help i've watched all of your videos and subbed
I freaking out that people are using ai art than human drawing which lead to the downfall of human art
Thank you so much for these mid Journey tutorials. I'm a person that learns visually so this is really helpful, but I don't always have the time or want to make the time to go back and rewatch videos for instructions again. Do you think you could create a PDF visual guide that you update as you go along?
Why would moving "ambient city background" closer to the beginning, make the character look more realistic?
I'm very skeptical.
Good job
Yes. AI prompting is definitely a discipline in its own right. Maybe usefull for standalone pictures. As a storyteller needing a consistent series of closely and precisely related images it is becoming increasingly more clear to me that the time I would have to put in acquiring this skill with such a vague results and quasi impossible continuity in detailing is not efficient at all. Commissioning the work to a real human skill is still far far more effective and quicker in the end. Hence even cheaper if your time is valuable. Neverning being more rewarding.
……………
This isn’t artwork, your just stealing so just quit you fraud.
AI-users = art thieves.
Pathetic you ghouls are.
Miazaky hates AI-users btw.
https://www.youtube.com/watch?v=ngZ0K3lWKRc
Dude how i can convert real photo to cartoon or 3d format in Midjourney AI
Is this technique also applicable for dall e?
Earlier words being weighted more heavily in the prompt is just not true. They are all weighted equally unless specified with different weight values. (according to their documentattion)
OK, this may be a strange question, but what is the background music you are using in your videos?
can we upload our own image and make it cartoonish?
I hate it. Not a fan!
That was a great tutorial, thank you!
Great tutorial,thanks
As with so much human communication, clarity and specificity are critical.
Regarding your question, why the guy looks so much like Miyazaki himself: I'm just guessing here, but – doesn't AI image generation work with connections (from words in the prompt, and their context, to images), at least roughly speaking? And now really, what person does have a stronger connection to Studio Ghibli than him? Wouldn't be too surprised if he showed up in the training data quite often, in that context
…
Love the succinct way you teach. Short and right to the point. I have learnt lots. I was using different prompts to create a celestial fantasy angel and didn’t get much success. Will try your prompts tomorrow and see what Midjourney comes up with. I feel I will get better results as I was getting a bit frustrated so I gave up. Thank you.
Ai is probably creating a database of the best prompts and tricks soon it won't even need people's input LOL.
I think it's probably better to leave out the style of at the beginning, instead after the "main idea" type "by (artist)". This works much better than describing the style such as "akira style" or "gihbli style", as you can target the exact style you want by the artist name. You can also mix artists, like "by (artist 1) and (artist 2).
I wish people that were doing these types of videos around AI would include the prompts, seeds, and settings that create these images (at least the base one) so that we can attempt to recreate them and have a starting point to then learn and grow from. But, they rarely do.
what's that background music it's so peaceful
Is it possible to create one same character in different poses??
It’s really hard to see the prompts you are typing in your video. If this is a tutorial you should either show them larger on the screen or include the prompts in the description
super helpful, thank you!
hello I love your videos, could you tell me how to do it if I want from a friend's photo to make him an animated version that looks like him please
Nice info thanks!
Bro please try out Midjourney v4 and make a video on it.