You just saved me life. Brought me back from the edge of a ledge. And back to the human race with VARY REGION which is REPAINT in Alpha which is what I needed because I generated the image in Alpha and it's a professional headshot so no creative upscales just stark realism. Thank you 🙏
I like the clear step by step explanations in this tutorial. Focussing on one image and working to achieve your creative vision with the tools and working around their limitations was a great way to structure the video. Nice work.
Can you show us how to do different angles of the same object? So if I want to do a car, can I get multiple different angles of the same car? Loved this tutorial dude. Subbed!
I have a question: the quality of the image keeps getting worse with each Vary (Region). In other words, after a couple of uses of Vary (Region) editor, the images get a lot darker and lose on quality. Did you experience this? Is there a workaround? I have image references if you need me to clarify.
Is it now possible with some AI tool to blend a picture of a product of a certain design , for example some jewelry and a picture of a person? And then put that product on the person? Something like replacement for real photography set
TOPIC REQUEST: Now that you have this tech-human character, how can you create more pictures that use this same character in different poses and environments? Can you do a tutorial on character consistency across different environments, poses, actions, etc.? Thanks!
I'm currently experimenting with it for the next video! Getting some really decent consistency but the image quality is awful. Hoping to improve that...
Great and helpful video! The music track in the background is very soothing (almost meditative) and complements your voice beautifully. Can you share the name of the track, please?
Very intersting video thx. I would interested into some explanation on how to create prompt with a specific artiste or mangaka like Toriyama/ Kishimoto or others but using a photo of your self as model. It would be like the artiste drew you. +1 sub :)
Question: when using the vary region it removes the object (like content aware fill) but it does not replace it with the new object i've prompted. I've also tried --stylize 1000 --chaos 25
Thank your for your response! The original prompt was something like: office guy behind desk. - sref (media link to another image) Then I upscaled it and tried vary region. Here I masked part of the desk where a laptop was standing and prompted: law books and law hammer. All the results where mostly an empty desk on that part with nothing replaced. Sometimes a tiny magazine maybe was recognizable. I also tried -stylize 1000 -chaos 100 but nothing changed much. I will give it another try with another themes, but your thoughts are of course appreciated :)
I found a similar struggle. I wanted to add a katana on the back of the character in this video. Selected the area behind my char and prompted "katana, katana on back" but it didn't add it to the image. I too tried --stylise 1000 and --chaos 100 which didn't help much. What I found to be more helpful was selecting a larger area and giving a more general prompt. I selected the entire body of my character and the area behind it & prompted "samurai with katana on back". That gave me much better results but image still looked a little odd. Hope this helps?
🔥 thanks for putting this together, super helpful. Any chance you make a video on non-charecter image generation? 2d logos or frame styles that could be exported as dxf?
Thanks for this great video! I would like a video about adding a real life image to an AI generated image - f.ex. taking a real image of clothes and putting an AI generated person in those clothes.
![Are you writing your Text Prompts WRONG? Midjourney AI [Prompt Crafting Tutorial] - Ep.1](http://i.ytimg.com/vi/2QTXTp-_5JY/mqdefault.jpg)
![Are you writing your Text Prompts WRONG? Midjourney AI [Prompt Crafting Tutorial] - Ep.1](/img/tr.png)
You just saved me life. Brought me back from the edge of a ledge. And back to the human race with VARY REGION which is REPAINT in Alpha which is what I needed because I generated the image in Alpha and it's a professional headshot so no creative upscales just stark realism. Thank you 🙏
So useful, thank you Cyankatana ❤
I like the clear step by step explanations in this tutorial. Focussing on one image and working to achieve your creative vision with the tools and working around their limitations was a great way to structure the video. Nice work.
Appreciate the feedback Fred!
This is just what I needed to get started with in-painting. Thank you.
You're most welcome! Open to video requests as well if you want to learn about other features etc.
Love your video style! Looking forward for more of those 😍
Thank you & welcome to the Discord brother 👏
One of the best video tutorial about Midjourney ever seen!
More to come! Thank you for the kind words brother.
@@cyankatana Don't see the time!
❤ Thank you from England 🇬🇧
Thank you for watching!
Very nice, is ot possible to change a part with an existing image ? For ex a woman hold a can with your own can
I thought this image looked familiar. Congrats!
Just at time for me, thanks !🎉
Yay! Thank you for watching :)
Can you show us how to do different angles of the same object? So if I want to do a car, can I get multiple different angles of the same car? Loved this tutorial dude. Subbed!
prompt like "car, full body, multiple angles"
I have a question: the quality of the image keeps getting worse with each Vary (Region). In other words, after a couple of uses of Vary (Region) editor, the images get a lot darker and lose on quality. Did you experience this? Is there a workaround? I have image references if you need me to clarify.
I never known about inpaint thanks
Is it now possible with some AI tool to blend a picture of a product of a certain design , for example some jewelry and a picture of a person? And then put that product on the person? Something like replacement for real photography set
can you make video on warrior or character hold a sword? may be different poses
I'll explore similar ideas. Stay tuned!
love your videos, it helps me a lot, can someone help me to understand, why - - sref does not work, i try ewerything cannot understand whats wrong ???
Had to rewatch it 2 times, guess I am a slow learner lmao.. the bullet points helped to pause and replicate, thanks for the video!
Thank you brother!
TOPIC REQUEST: Now that you have this tech-human character, how can you create more pictures that use this same character in different poses and environments? Can you do a tutorial on character consistency across different environments, poses, actions, etc.? Thanks!
Love this! Could be like a series following the same character for different stuff
Thats a fantastic idea, I'll play around and see if I get any decent results worth sharing!
Hopefully you find something on pose contril, pretty much the same character but with a different pose@cyankatana
Upload picture and use --sref and the url also seeds
Great video. Just subscribed. Consistent characters is a great future topic. You may have already covered this.
I'm currently experimenting with it for the next video! Getting some really decent consistency but the image quality is awful. Hoping to improve that...
Great and helpful video! The music track in the background is very soothing (almost meditative) and complements your voice beautifully. Can you share the name of the track, please?
Thank you! Absolutely, its Silent Saunas - Mandala Dreams
Thank you!
Awesome video thank you 🎉
Thank you for watching :)
@@cyankatana subscribed
Very intersting video thx.
I would interested into some explanation on how to create prompt with a specific artiste or mangaka like Toriyama/ Kishimoto or others but using a photo of your self as model. It would be like the artiste drew you.
+1 sub :)
Oh thats an interesting one! I'll explore 🙌
Have you played around with the media link add some descriptive prompting and in the style of “artist’
Question: when using the vary region it removes the object (like content aware fill) but it does not replace it with the new object i've prompted. I've also tried --stylize 1000 --chaos 25
What exactly are you trying to do? More context you can give me the better
Thank your for your response! The original prompt was something like: office guy behind desk. - sref (media link to another image)
Then I upscaled it and tried vary region. Here I masked part of the desk where a laptop was standing and prompted: law books and law hammer.
All the results where mostly an empty desk on that part with nothing replaced. Sometimes a tiny magazine maybe was recognizable. I also tried -stylize 1000 -chaos 100 but nothing changed much.
I will give it another try with another themes, but your thoughts are of course appreciated :)
I found a similar struggle. I wanted to add a katana on the back of the character in this video. Selected the area behind my char and prompted "katana, katana on back" but it didn't add it to the image. I too tried --stylise 1000 and --chaos 100 which didn't help much.
What I found to be more helpful was selecting a larger area and giving a more general prompt. I selected the entire body of my character and the area behind it & prompted "samurai with katana on back". That gave me much better results but image still looked a little odd. Hope this helps?
@@cyankatana thanks for your insights! Will try that out too, with a bigger area!
The alpha site has in-painting. But great great tutorial.
Oh really? I couldn't find it, how do you access it?
🔥 thanks for putting this together, super helpful. Any chance you make a video on non-charecter image generation? 2d logos or frame styles that could be exported as dxf?
Glad I could be of help. I'll look into it!
I'm surprised Midjourney didn't censor this prompt or image result as "graphic" yet
Thanks for this great video! I would like a video about adding a real life image to an AI generated image - f.ex. taking a real image of clothes and putting an AI generated person in those clothes.
I'll look into it! Sounds interesting but might be challenging given where Midjourney currently is. Maybe it could work as an image reference...
where have you been since a year ago
Hi brother! Life. I'm back now.