That’s interesting I didn’t know that was even possible using UI gradients. And yeah I frames not decals.
Really I’m just trying to figure out how fast editable images are. Because there are definitely some things I want to try, but if they’re not fast enough for doing a lot of real-time shader effects on UI and whatnot, idk.
Has there been any data published about the speed or, whether or not we can set groups of pictures at the same time in an array?
uh roblox? i think you made a mistake… this is too many good things at once!!! wooo
Quick question for any engineers reading this topic: will EditableImages on mobile be compressed?
I’ve got an open bug report centered around how ImageLabels are compressed when they really shouldn’t be on mobile. My use-case is pixelart-centered games.
My first tests with EditableImages have been genuinely mindblowing given how fast it is to render an image.
If the image data isn’t ever compressed when displayed to the user, this will act as a complete replacement for ImageLabel spritesheets for everything I could ever need one for, and will make my bug report completely obsolete 
As a bonus, here’s a clip of me and a friend working on our EditableImage codepath, only to realize we messed up the channels:
Do EditableMeshes work with standard Parts and Unions? Would be even cooler if we could potentially change the geometry of Parts and Unions too.
Also is the 1024x1024 resolution limit on EditableImages permanent? At this point in time i really don’t see a reason as to why that limits needs to exist, at least on runtime. I do know that roblox would usually downres any published image on the platform to 1024x1024 though that dosen’t entirely mean we could not get some more wiggle room directly at runtime…
Also are there any performance metrics for these features? Like how much time would it take for EditableImages to update a pixel or how much time would EditableMeshes take to create a triangle and etc etc… Also with the performance metrics, won’t the additional moderation features just kinda hit performance for these things? I understand the need for some moderation systems though i feel like as if moderation could entirely make these features next to useless if they just are slowed down to an insane extent just so they can keep up with moderation (Or at least force some back-end systems to constantly use resources just so whatever moderation system has the latest up-to-date systems).
As for moderation, what about false positives? These days it’s overall pretty easy to get hit by moderation actions for simply uploading assets that your AI systems believe are inappropriate content. Again, i get the need for moderation systems however i believe theres a chance these systems are made entirely useless due to the risk they pose in the face of moderation. Like what if you constantly update a EditableMesh/EditableImage? Would the moderation systems check each new version of said assets? If this is the case then those features would just be overall really dangerous to use as again, the moderation systems already hit people relatively frequently for one-time assets, and with some features that could potentially update assets at runtime (potentially even every single frame) there could be A LOT of situations in which developers get hit by the moderation system unfairly. There is a reason as to why people opt into uploading assets to the platform using alt-accounts.
Also what about players who create inappropriate assets? Right now its only and always the place owner who gets hit by the moderation if roblox deems something in their place as inappropriate, regardless of who made it. I don’t think there would be any easy nor any viable way for developers to simply just check and verify what sort of meshes and images player’s make in their games. I personally really cannot put my faith on whatever moderation systems you guys have in place for these things alone. Players will eventually find ways to bypass “safe replication” and “safety filtering” systems regardless.
Personally i believe that these sort of moderation systems are (and will still be) bottlenecking developer potential just to ensure the absolute “safety” of other peoples children. I don’t think children should be exposed to bad content but i also believe its not my job to ensure the safety of children, nor do i want to be the victim of situations in which whatever moderation ai deems a random mesh/image me or a player made as a random test or just for fun (nothing even against TOS) as inappropriate and therefore should ban/terminate my account. I believe if roblox wants to actually step up with their technologies and systems then one crucial step would have to be to tell parents to actually PARENT their children and not let multi-billion dollar companies or random internet people do it for them.
this is crazy! im working on a 2d game and i have to say this is a literal game changer, can we by any chance get some more functions for editable images? i’d love to have a function to get pixels of a specific color on the image? this can be very useful to replace a certain color for another, im sure we can do this using readpixel and iterating through the entire image but i think an internal function for this would probably be more performant, thank you!
the meshes just randomly disappear depending on how i look at them, the terrain is made up of chunks with a editable mesh for each one
THIS. THIS RIGHT HERE. EditableImages are cool as hell already, but holy hell would that make this update so much cooler!
Finally it’s here! Now I’m just waiting for you guys to make it so I can mess with EditableImages and SurfaceAppearances/Materials >:)
What is this “safety filtering”? Safety in what way? Preventing “inappropriate shapes”? Or simply preventing saturating the network connection?
please add the ability to write to mipmaps for editableimage, and also texture filtering options are long overdue at this point. In my case I’m using a low-resolution pixel art texture atlas with no padding, which ofc causes bleeding around edges, because of both the mipmaps and the texture filtering mode. Oh and obviously nearest neighbor filtering is needed for pixel art.
Amazing update! Now I will be able to create 2D games that wrap around 3D geometry!
So you’re telling me I can sort of paint meshes??
Custom vehicle liveries for planes, trains and stuff??
Is there a chance that EditableImages could be applied to SurfaceAppearance at some point in the future?
Other than that, this looks great and I cannot wait to see what people do with it!
I’m surprised by the amount of back to back updates roblox did today!
I can’t express what a huge leap forward this is for Roblox development! 
The ability to write pixels is already HUGE for things like rendering custom images, performant and quality spritesheets, and 3d or even 2d rendering systems within Roblox!
Editing meshes is also a huge win, things like terrain generation can have significant improvement to their performance, as well as visually look a lot better as developers have complete control over how it looks! Instead of changing a material in the terrain, developers could instead create meshes on the fly and possibly procedurally texture them to have different materials.
You could potentially even simulate damage to meshparts using this, say you’re wearing armor and are struck by a sword in an RPG. Not only could the sword slice through the metal armor in a convincing way, but it could potentially save! Reminding you of your past battles every time you play the game!
If it’s possible to have blank pixels on an editable image, then you can potentially use it for things like shaders. Like outlining a character a certain way for stylization, maybe custom water? etc etc
Lovely update! Keep up the good work! These past few updates have been a huge improvement, even if they’re in a beta still! 
oh my god shape keys oh my GOD SHAPE KEYS
I’m going to make Blender in Roblox.
Let’s freaking go! Thank you so much!
Great update! I will definitely looking at what developers do with this
Also, this must be the November update eh?
Its December 1st now… for me at least 
who will be the first to make a mirror using this?
