Neat! I was actually thinking about this concept just a couple days ago, what a coincidence!
1 Like
Number one:
And also, there is another person who did this for images:
I was able to do this with it:
4 Likes
The lag would be pretty bad though! Since you need so many frames… Brr, me making a raytracer be like 100 gigs ram usage
1 Like
Yeah, it tested it on a 37x34 image, HD and 4K images would be near-impossible due to memory limits.
2 Likes
Not the first time that something like this has been made, just never publicised.
https://www.roblox.com/games/3917247723/
https://www.roblox.com/games/6047892253/
1 Like
I only imagine if someone managed to store individual image files within a file in a form of a UUID to make a video. Problem is that youll have to request the server for a new frame. Or stream the video from a hosted server to roblox by requesting it every frame. But it all depends on how its done, i would do it in a similar manner but with a reduced resolution of 400-200 pixels for each frame. But idk really, seems like a good way to make a video streaming system for your own game.
1 Like
I’ve been working on that 
Its possible, it could most likely be done with the same way the op has done this, but it would probably be super low resoluion and would be like 10fps
If i am correct, if i use replit to save an image with an UUID and save it to a folder which is meant to be an anchor for a video.
Too bad my brain hasnt completely learned python, so making this possible is a stab in the eye.
However i would say splitting each image frame into pixel packets would be a viable option.
Using the power of my pc and patience. I have managed to render a 1080p image and a 1440p image. (Scripts won’t be open source since it can be abused - can show videos.)
Most of this was done through c++.
1440p
1080p
Tried to make this one with depth. (As you can see, it’s can be abused)
https://gyazo.com/c67aad42457833aa0a4c9da58cb97a9e
Ram Usage:
7 Likes
Are you not worried about getting your account banned for creating TOS breaking functionality?
Will you feel bad at all when little kids are exposed to inappropriate content due to your shared code?
You can’t control what people do with your code but you are helping enable workarounds to image moderation. (which seems like an incredibly bad idea)
Don’t do this (but here’s the git repository for how to do it).
Can I know how many instance count you have when the image has finished rendering?
And do you use GuiObject or do you use BasePart?
I did make something similar except that I made a python script to parse images and turn them to Lua tables and I would need to embed those inside the game on a modulescript, then a normal script parses the modulescript and renders the image
Unfortunately this used GuiObjects and I keep getting hit with Roblox’s limitation
If you read my post, you’d see this is a proof of concept, not a ready, working-out-of-the-box resource. It also isn’t technically TOS breaking since it is theoretically possible to manually make frames yourself with nothing but Studio to mimic an image.
I’m not trying to control it, naturally, I can’t. I added a disclaimer multiple disclaimers because you obviously shouldn’t use it in your games unless you do want to get banned. In addition, as others have already pointed out, this exists already. Don’t know why you’d pick on the fourth or fifth post of something like this.
Also, if anyone competent spent around thirty minutes trying to do this, they could easily achieve the same result or better.
1 Like
I like your effort, but using GUIs is a no-go for image renders in roblox. They’re not efficient at all and there’s a limit to how many gui instances can exist on your screen at any given time.
Now as for how to go about it better? BaseParts and color-blending your run-length data into EDIT: less* parts that are longer with averaged color values. Here’s an old screenshot of my renderer (not even hosted locally, it’s just that efficient. It uses JS as a backend.)
That image was Full HD, 1920x1080 but was optimized to cut that pixel count down to just 29% of the original.
3 Likes
What do you mean by color-blending?
Oh my bad I missed a bit there, I meant to say “comparison color-blending”
And essentially what I mean by that is comparing the RGB (or in my case, HSV) values to one another and if they are within let’s say 15 (/255) of each other, blend them together, adding up all the values and then averaging them once you encounter a color that does not meet these conditions.
I won’t go too in depth considering this took around a year to get to that state of optimization (with changes to both the front/back ends) and I have no plans on releasing it to the community unfortunately.
1 Like
I may have posted something on those previous posts, I can’t recall.
Sure, players can stack up blocks in the shape of something, but your(and others before you) ‘tech demo’ is aimed at making it trivial to work around image moderation. I guess I don’t get the purpose of it. (if it can’t or shouldn’t be used).
Anyway carry on… it doesn’t have to make sense to me 
The purpose of the developers who make these is rarely just “bypass moderation” if it even includes that at all. For me, it was a test of how much optimization and compression via encoding I could manage as well as how much abuse can the Roblox engine take with these bulky tasks with its high-level focus.
Elaborating on what I mean by high-level: It’s a framework that doesn’t allow you to access the core of the environment, in this case, we must use parts and cannot individually draw pixels to the screen.
1 Like