Introduction To Compositing & Fix Deformed Faces In Playground Canvas

Playground Canvas 3 Simple But Powerful Tips. Hello everyone! Today I wanted to share with you some simple but powerful tips on using Playground's Canvas. There is so much more you can do with it but this will be a good start! While you are at it be sure to try out the new stable diffusion xl 0.9 model. We'll cover some basic navigation to get around the canvas smoothly, fix deformed faces and a quick introduction to compositing.
Time Stamps
0:00 Easy navigating on canvas
0:44 How to fix deformed faces
4:33 Basic compositing
Find us on
Playground AI playgroundai.com
Discord / discord
Twitter / playground_ai

[@rebelgonebad]

You and Playground just seem to get better and better. Thanks for the info!

[@playgroundai]

Love to hear that and you're welcome! 🙌

[@bhosaleranveer99]

Thanks a lot for showing how to fix the faces in full body shot images. Just what I was looking for 😊😊. Please keep doing more in depth tutorials especially for beginners.

[@playgroundai]

Glad to hear it was useful! As mentioned we will have a running series of beginner videos to come very soon! Stay tuned!

[@kemikemi756]

my background removal just wont work I click it it does nothing. Tried with various different works

[@fictionaddiction4706]

Love your tutorials, I've learned a lot. it was so random when I started I had no idea what was happenig.

[@digigoliath]

Ah, learnt a few tricks! The Face Fix will save many images that would otherwise be deleted. TQVM. Looking forward to the next tutorials.

[@playgroundai]

Yes exactly! Easier now to make full body images that are higher detail. 👍

[@kemikemi756]

It would be better even if I could remove the background for the copied image...

[@pedrodelacabra1855]

Nice. I'm really looking forward to the new video series you mentioned at the end. Keep up the good work. 👍

[@playgroundai]

Yes coming very soon! If you have any feedback on what you'd like us to cover feel free to let us know! 👍

[@hermitagetaiji9451]

Thankyou again for such an important tutorial. This is helping me develop my ability to create with Playground, not just 'play with it'. Using your compositing tip I was able to create an image that exceeded my original imagination of what I had hoped for; this is the jewel in the crown of AI - to go beyond your own imagination.

[@playgroundai]

Amazing! That's no nice to hear! We will do more in depth compositing as there are so many things to consider. But as you can see even with those simple techniques you can already do so much. Appreciate the feedback! 🙌

[@AIWizardTech]

Thank You very much, you are wonderful to show us all new and new possibilities of practical work in Playground, Good luck

[@playgroundai]

Thanks for the positive feedback! There is so much we will be sharing with you all soon!

[@user-ie7yt6ef1r]

Thank you for your videos! Very valuable information.

[@playgroundai]

You're very welcome, we're so glad they've been helpful!

[@luckyenigmas]

YES. I was WONDERING about this. Thank you for this clarifying!

[@playgroundai]

Glad it was helpful! And you’re welcome!

[@nn-db4fw]

Fantastic! This opens a whole new range of possibilities for me. I love this :)

[@playgroundai]

That's what I love to hear!

[@MindPersonality-fr7ru]

Wow that's amazing, it's really helpful to make an error design and have a blur effect

[@playgroundai]

Glad it was helpful!

[@stonkeris69tricaars23]

Thanks for this lovely ai. It's very helpful for me and i hope it's gonna be fine so people whit big ieda or fantasy can make a dream become a true❤❤. Thanks thad you guys existing whit this gold❤❤

[@ultra359hardcore]

GREAT! Thank you! 😍

[@playgroundai]

You’re welcome! 🙌🏼

[@Active-Vision]

Awesome content, thanks for sharing and waiting for new getting started tutorials!

[@playgroundai]

More to come!

[@wolfsangeleyes]

These tips & tricks are amazing and very helpful, thank you so much.

[@playgroundai]

You’re very welcome! I appreciate you consistently supporting us. 🙌🏼

[@richardchin1545]

Great tips. The face fix will be V. useful. Like also that cavas has layers.

[@playgroundai]

Indeed great additions! Appreciate the support 👍🏼

[@wolfsangeleyes]

And thank you so much for what to do with an image once you've removed the background. I was eagerly awaiting this tutorial. (and now I finally now what the 'bring to foreground' does😁)

[@playgroundai]

Yes we do have layers functionality which is very useful. You can also use background remover for other use cases. Of course I’ll be covering that soon too!

[@wolfsangeleyes]

@@playgroundai Looking forward to that 🤗🤗

[@digigoliath]

Have to consider the Lighting for realistic composites

[@playgroundai]

Yes very true but wanted to save that for another video to come. Good point 👍🏼

[@Dreamseerapp]

Thank for the walkthrough. There are very few tools that allow changing angles of faces and characters (i.e. pose/style-invariant reenactment). Would be great if your could give some tips on how this can be done in Playground AI. This is very valuable in visual storytelling.

[@playgroundai]

Generally you can prompt for different angles and perspectives as well as using controlnet. The main limitation in consistency with clothing. There are workarounds that can be done with some post production. I'll be covering this soon.

[@Dreamseerapp]

Thanks a ton @@playgroundai ! Will look into it and try.

[@johnleighdesigns]

Loved this and wondered if there is a way to integrate those three cutouts at 7 mins marker with the environment

[@playgroundai]

Yes, see our latest video here kzbin.info/www/bejne/mIeVdmp5dsllq5o

[@koendebelie197]

I am just getting into AI image generation and this Canvas feature really opens up so many new possibilities, it's awesome ! Is there any way to get realistic shadows into your scene from the characters you place in them? That would be great !

[@playgroundai]

Yes, using img2img with the new compose scene. I'm going to do a video on this soon!

[@salimbelaidi8774]

yesterday I tried to do an image sequence to make video but the process was stressful to do frame by frame .I think if you add image sequence in image to image option it would be awesome

[@playgroundai]

We'll definitely consider it. The focus at the moment is to make Playground a powerful image editor, who knows which direction it will lead us. 👍

[@Altzk]

Nice video, easy to understand!

[@Altzk]

is there any chances to adjust image dimensions?

[@Altzk]

nevermind, I've found the tutorial, thanks a lot!

[@playgroundai]

Thank you!

[@playgroundai]

Oh great! 😊

[@pablovidal3777]

Cuando removemos el fondo de una imagen se pierde la sombra del personaje, al pasar el recorte a un nuevo fondo pierde realismo. Lo mismo sucede con la iluminacion, existe la forma de crear de nuevo la sombra y que la iluminación sea acorde con el nuevo fondo?

[@veramaria6006]

Hi, I don't have the MODEL options in my canvas . What can I do to fix this? Thank you for your videos

[@playgroundai]

It's an older video so you don't have to worry about selecting model on Canvas anymore, playground will choose the correct one.

[@tony_dioro]

Hi. Can I use this same technique to fix hands?. Thank you

[@playgroundai]

Yes however starting with somewhat of a decent image of the hand is best. Some models do better with hands like rev animated, realistic vision or SDXL. I find as long as it has a decent shape, when you crop and upscale with image to image it will come out more defined. Takes a bit of work but very doable. I will cover hands specifically very soon.

[@coldgin2023]

Anyone have any ideas on how to get proper hands and feet? I have a ton of useless images from playground AI due to alien looking fingers lol

[@playgroundai]

The part where I corrected the face can also apply to hands. There are a few ways to get better hands, I'll go in depth about this soon. A few tips, start with a filter that does hands well like Rev Animated or Realistic Vision. You can always utilize seeds and img2img if you want to use a different filter after. Best solution but more effort, use reference photo or take a picture yourself with your phone. You can also utilize controlnet which is very useful.

[@MissChelle]

Wish I had a Chromebook, can’t do this on iPad 😭😭😭❤️🇦🇺

[@playgroundai]

Unfortunately we are still working on Canvas for mobile devices, I don't have a timeline at the moment.

[@hasp24]

This doesn't work for me. When I use the erase tool to remove the sharp lines around the new face, I hit g on the keyboard and the option to "impaint" appears instead of "generate". Then when I click impaint, the lines are removed but the new face gets distorted slightly. It seems to ignore the erased area and just reprocesses the image as a whole. Any ideas? I am using the free version at the moment.

[@playgroundai]

technically when erasing parts of the image is inpainting. Make sure as well that the generation box stays within the image

[@JasonAdam-gl9jf]

I don't see an option to select the Model on my canvas options on the right

[@playgroundai]

No need, just select the filter you want to use and on Canvas it will switch to the necessary model. By default if no filter is selected Canvas uses SDXL.

[@danrazART]

Please give us object rotation and zoom to selected

[@playgroundai]

Object rotation is on our list, I’ll see what the team says about zoom to selected 👍🏼

[@danrazART]

@@playgroundai zoom to selected will help working with canvas in a svap

[@TruthSpeaker-yj6xp]

Why canvas isn't available on mobile

[@playgroundai]

It’s currently being worked on. Mobile works quite different than desktop.

[@DailyInfo-tainment]

in canvas i dont have playgroung option on right side...

[@playgroundai]

Playground V1 will be obsolete soon, don't worry about it.

[@DailyInfo-tainment]

@@playgroundai so please make new tutorial with updated site ui

[@playgroundai]

Lots of other recent videos to check out. Besides that part you can ignore as the same concept applies.

[@SnoopyDoofie]

Had no idea that the images in canvas were different layers. I thought they were all just separate images.

[@playgroundai]

Yes it has layer functionality. 👍🏼

[@cekuhnen]

when you process his face - I cannot switch the model

[@playgroundai]

We've updated it since then, by Default it uses SDXL and it will detect any of the filters that use SD1.5. No need to switch models now. 👍

[@cekuhnen]

@@playgroundai ah so canvas uses SDXL ?

[@playgroundai]

Yes for the initial generations, however with outpainting it will switch to Playground V1 as we are still waiting for SDXL out/inpainting to be come available.

[@cekuhnen]

@@playgroundai ah I see. Btw I also see often SD 1.5 being used in canvas on my side when I look I to the image properties - but ain’t see playground v1 - maybe a mis classification ?

[@myohokyo]

Another great and helpful How-To video. You should do some AI image creation masterclasses someday. 😛

[@playgroundai]

Funny you mention that, been thinking of a masterclass dedicated to Playground. 😉👍🏼

[@SnoopyDoofie]

Please add a feature that lets users create a url that takes them to their profile the same way KZbin does. For example: [playgroundai domain]/[at]MyProfileName This will make it easier to promote a profile on other social media platforms and it will give you a lot more traction. Thanks!

[@playgroundai]

That’s great feedback, I’ll see what the team can do 👍🏼