Photoshop's generative fill combined with all its other features is for me, the most significant and useful addition for all the years I am using it, mostly for photography and some graphic editing. In the other hand I use for couple of months now Stable Diffusion models in WEBUI 1111 for AI generated images, inpaint and img2img and so on, and I find myself going from the WEBUI to Photoshop and back again many times for an image. I would say that both Midjournay as an alternative to SDXL and Photoshop complement each other pretty well. Great video and the comparisons helped especially when I do not use Midjourney but is under my radar for some time now. Thank you
Photoshops generative fill usually needs at least a few rounds with more variations, before getting a good result. Comparing only the first three results, doesn't really do it justice :)
Using the website instead of discord is a LOT easier since you can organize things a lot better. Especially given I keep the face of a character the consistent and re-prompt everything but the face. You can organize putting things in folders etc instead of in discord where you have to scroll through a lot of generations unless you create manual folders to start the generations in.
2023-10-26 Importent! Now you´ll first have to write " /prefer remix " (in prompt window) hit enter...and then you can do as described in curious refuge´s vid.
Do you mean from your midjourney gallery ... definitely yes. It might have came out after this video but if you are able to work from their website you definitely can.
im quite dissapointed with both, and even other platforms, what you see in social media is amazing but prompting something so simple generates garbage, I wanna add buildings to the background and I prompted 10 different ways with different models on leonardo AI and got terrible results, came here to see if photoshop was better but I guess is just as bad.... only great for removing stuff but adding feels like a long way to go. sadly. I'm trying to set extend a frame of my film and its just not doing it.
Curious, i can’t type in any prompts after I select an area, as soon as I start typing the light box goes away. I’m in Chrome, not the app, anyone else have this issue? So odd Great video!
The midjourney hair looks a lot more red than brown. Also the one you went with it looks more like the hair changes color than being affected by the light. Especially that middle group of strands where it looks like it should be under the same lighting but is desaturated on one side and reddish brown on the other. The photoshop one actually looks more like it is part of the same image. and same lighting conditions. There is an orange light coming from below her perhaps from a monitor or indicator light off screen. But midjourney put a warmer color temperature behind her just on her hair. Photoshop remained consistent with the lighting. But Midjourney did better on detail. You also skipped on a HUGE advantage with this for use in AI style film making. Which is to make consistent characters by repainting everything except the face. Making small changes is great for sure ... but being able to have consistent characters in different locations is the HUGE part.
You realy are one of the best teachers out there.
Photoshop's generative fill combined with all its other features is for me, the most significant and useful addition for all the years I am using it, mostly for photography and some graphic editing. In the other hand I use for couple of months now Stable Diffusion models in WEBUI 1111 for AI generated images, inpaint and img2img and so on, and I find myself going from the WEBUI to Photoshop and back again many times for an image. I would say that both Midjournay as an alternative to SDXL and Photoshop complement each other pretty well.
Great video and the comparisons helped especially when I do not use Midjourney but is under my radar for some time now. Thank you
Hello, can you post a tutorial on writing amazing Midjourney prompts?
Photoshops generative fill usually needs at least a few rounds with more variations, before getting a good result. Comparing only the first three results, doesn't really do it justice :)
Using the website instead of discord is a LOT easier since you can organize things a lot better. Especially given I keep the face of a character the consistent and re-prompt everything but the face. You can organize putting things in folders etc instead of in discord where you have to scroll through a lot of generations unless you create manual folders to start the generations in.
This video is almost a year old :)
2023-10-26 Importent! Now you´ll first have to write " /prefer remix " (in prompt window) hit enter...and then you can do as described in curious refuge´s vid.
Thanks for the clarification. You can also simply select the remix mode in your settings and it will automatically pop up as well.
I think another big thing to state and why MJ is king in my opinion, is I can use MJ anywhere I want or need to from mobile via iPad or iPhone.
Yes, MJ is a powerhouse right now and is leading in many ways. Thanks for the comment!
Hello, great video, a question: is there a way to do the inpainting in midjourney but from images in our gallery?
Unfortunately no. I really wish you could though.
Do you mean from your midjourney gallery ... definitely yes. It might have came out after this video but if you are able to work from their website you definitely can.
can you do the inpainting with non midjourney images in midjourney?
We'll consider it!
@@curiousrefuge is it possible im saying
Man , don’t stop giving news about the world of AI. Take care 💪
im quite dissapointed with both, and even other platforms, what you see in social media is amazing but prompting something so simple generates garbage, I wanna add buildings to the background and I prompted 10 different ways with different models on leonardo AI and got terrible results, came here to see if photoshop was better but I guess is just as bad....
only great for removing stuff but adding feels like a long way to go. sadly.
I'm trying to set extend a frame of my film and its just not doing it.
Sometimes it's difficult to get net new assets but fixing issues is sometimes the better use case.
Curious, i can’t type in any prompts after I select an area, as soon as I start typing the light box goes away. I’m in Chrome, not the app, anyone else have this issue? So odd
Great video!
Make sure you turn on remix mode in your settings. That should fix it up for you.
The midjourney hair looks a lot more red than brown. Also the one you went with it looks more like the hair changes color than being affected by the light. Especially that middle group of strands where it looks like it should be under the same lighting but is desaturated on one side and reddish brown on the other. The photoshop one actually looks more like it is part of the same image. and same lighting conditions. There is an orange light coming from below her perhaps from a monitor or indicator light off screen. But midjourney put a warmer color temperature behind her just on her hair. Photoshop remained consistent with the lighting. But Midjourney did better on detail.
You also skipped on a HUGE advantage with this for use in AI style film making. Which is to make consistent characters by repainting everything except the face. Making small changes is great for sure ... but being able to have consistent characters in different locations is the HUGE part.
keep in mind, this video is almost a year old ;)
@@curiousrefuge Is it? I didn't notice ... I seen 9 month old posts afterwards.
helpful, thanks
Glad it was helpful!
SUPER COOL!!!!! ❤ ✨
Thanks Roger! :)
Photoshop is so so behind, and midjourney becomes more and more useful, and it's just the beginning
I am so curious to see where Midjourney will be in 1 year!
In photoshop, you can impaint anything whereas in Midjourney you can only impaint what you produce on there. That's why photoshop is more practical
Yeah, I was about to say the same thing... Midjourney currently does not support inpainting on your own images
First?