AI Reimagined Backgrounds: Create Stunning Product Mockups with Stable Diffusion | BG Inpaint
HTML-код
- Опубликовано: 29 янв 2024
- Welcome to a new era of creativity with AI! In this video, I'll show you how to effortlessly blend your products or photos into diverse backgrounds using the incredible capabilities of Stable Diffusion Automatic 1111. Perfect for photographers, graphic designers, and digital artists, this tutorial is a game-changer for creating stunning visuals.
🔍 What You'll Learn:
- Setting Up the Environment: We'll start by installing the 'Inpaint Background' extension in Stable Diffusion Automatic 1111 on Windows.
- Automatic Background Masking: Discover how to mask and replace backgrounds automatically, and learn to tweak settings for optimal results.
- Creative Background Generation: From rustic shelves for your products to romantic scenes for a rose, I'll guide you through various scenarios.
- Fine-Tuning with Denoise Strength: Understand the impact of different denoise values on your images and how to balance creativity with realism.
- Challenging Scenarios: We tackle complex images, like robots in a forest, and explore solutions for seamless integration.
- Photoshop Tips: I'll share some quick Photoshop techniques to perfect your creations.
Enjoyed the video? Hit like to support my work. It's a simple gesture that helps me continue creating content for you.
#StableDiffusionTutorial #ProductMockup #DigitalArt
Great tutorial! Really helpful for creating stunning product mockups. Thanks for sharing your process with us!
Damn! Great Tutorial! Nice work around with image to image to make the image look like its actually in the scene.
good tutorial, simple and amazing
Thank you Very much
amazing !!
Usefull, THX :)
nice, thanks
This will be great edition for those who use 3D softwares for preview!
thanks a lot for background removal tool it's quite a thing
thanks
Hi, you can circle the "blue contours" in inpaint, rather than regenerate the entire image, then you won't have to use Photoshop and waste time on post-processing
I've been trying the same thing, but with adjusted spectacular highlights, and accurate lighting on the subject that matches the environment.. basically AI rendering. I think we can get there once SD allows grey scale masks. Right now it still just a masked cutout on a AI background an an easy shadow. But nice video, you got my sub!
Spor la treabă 🤝🤝
Great tutorial! I just discover your channel and there is plenty of good stuf! love it! Can you help me with one thing, i'm trying to change the background on a picture with a car or a perfume bottle but when I prompt (car... perfume bottle....) SD generates the items, and it's somehow behind the object. I can get to render my images like you do, without these artefact or just with a clean (edges) subject. would you mind giving me any advice on that? :)
Did you tried to prompt only the background without mentioning the objects? And also depends on the noise strength if is lower doesn't have the power to create those objects. It happened to me a few times also, but if i mentioned only the description the background it worked sometimes, try it and see if it helps, so only background and play with the denoise. If not work you can also Photoshop quick a background similar to what you want and then is easier for Ai to guess what you want
@@pixaroma thanks will try that! if I found my magic recipe, I will let you know :) the thing for me is that I'm running SD on a MacM1, so everything is slow whenever I try things... still lucky to be able to try it :)
Great video, been looking for something like this for a while. Can this be done with tranparent PNG images that already have the backgound removed?
Not sure if it works, i am not at the computer to test it, just try and see if it gives any error
@@pixaroma I wish I could, but Im using RunDiffusion in the cloud and sadly cant load the "Inpaint Background" extension to test. Thanks for the reply though!
Amazing!!! I have a question, why when I do the step again in image to image my image is softer, less detailed? Thank you very much!
Yeah it can loose details if the denoise strength is too low, you can increase the denoise more maybe , usually at 0.35 it start to get more details, but depends on the image, try to increase it slowly and see if you can get a version that work, or you mask only the contour you want to fix, that way is still looks better even is a little bit softer compared with the original that didnt fit right
Thank you so much! 🙏
error when installing the ext:GitCommandError: Cmd('git') failed due to: exit code(128) cmdline: git clone -v --filter=blob:none --
Sorry not sure what that means, also i switched to force UI now instead of A1111
can you create a videos regarding lighting, perspective, ambience colour that can be control? thanks
Is hard to controll anything with AI since it tries to guess everything. The only way i was able to kind of control it is with photoshoot rough painting of how i want the light and then with img2img was able to guess it where it was and had to mention it in the prompt. It doesn't know well left and right so if i say top left light sometimes it does it from right, unless i paint with a soft brush the light how i want it and then it can see it
well.. at least is some control.. lighting is very important subject in visual.. if got solutions or event a bit.. i think it helps a lot @@pixaroma
Is that model good for creating blank mockups?
For me has been working ok, in the last month i use either that or juggernaut xl
Is it possible also with comfyUI?
Probably, but I don't use it so i can help you there, i am using only automatic and forge for now, is easier for me since i am a designer not a programmer
@@pixaroma Btw thanks for this tutorial 😊
Inpaint Background not showed up on Forge :(
There are still extension that don't work yet in forge
Did you find any workaround to make it work on forge?
Is there a fee to use stable diffusion?
No, is free, you just need a good video card, like nvidia with more than 6gb of vram to run it on your computer.
@@pixaroma awesome! Thank you so much ☺️
does not work for me
Sorry to hear that, at the moment i recorded the video it worked for me
doesn't work :(
What part? It worked when i recorded the video
Great video!
Can you help, Is it possible to do something so that when the denoise strength incr, the main picture does not change, but only the background changes? because the AI begins to draw the main photo along with the background if the Denoise lvl is high
Thanx
You can do inpaint, check my latest video about Inpaint , that should work depending on the background
@@pixaroma thanx