Exactly. The big question is why they won’t make it an option in SurfaceAppearance or something like that.
But like, realistically, how bad would this be?
Because most the time you can combine textures that are lower res aswell?
Like use a 512x512 emissive and the rest is 1024x1024.
A single new image cant make memory consumption that much more worse, developers just need this freedom.
Please read up before asking the same things.
Ah, my bad i didnt see it.
One thing that I find weird here though is why does TexturePack use a texture when we dont add a textureid to lets say Roughness/Metallic?
Couldnt you just use a slider/value of 0-1 if we dont provide a textureid for roughness/metallic?
Wouldnt that reduce memory usage for non-used slots on its properties?
Gotta bump this. If we had a choice, I’d choose stuff like this over the facial tracking and ai stuff
Praying roblox announces this at RDC in a few months and not just more AI slop.
been a while this would still be nice to have, so im bumping it
I’m still struggling with the lack of this basic feature which might have already been made internally but is not exposed to developers. Right now I’m forced to take pictures of my model from every axis and paint glowing regions on with decals which I then overbright. This is way more memory-intensive than just letting me upload an emissive map as it needs 6 textures per mesh instead of just 1. It also doesn’t always play well with lighting, and it has no concept of depth.
Not having this feature is like a car not having a windshield. When are engine developers finally gonna implement this basic option for everyone to use?
ah yes, facial tracking and lots of high poly face having avatars is definitely less intensive than making an already existing feature usable for normal developers. good thinking roblox! you sure do have the best priorities!
almost 3 years and nothing… bump
Let’s just keep bumping this… such a simple feature that we need!
Legit I need some sort of emission texture, neon is not gonna cut it all the time
And the facial tracking doesn’t even come with the ability to import shape keys… ;-;
Not having shape keys yet is really dumb, roblox is really behind.
Let’s just keep bumping this… such a simple feature that we need!
And atleast give certain users access to higher resolution texture maps, but still a limit on how many.
Sir yes sir,
Honestly we could use with some info about this. Could a staff member give updates?
I have zero expertise on this stuff, so please correct me here if/where I am wrong.
I can only really see this happening if we get a new class whose instances are parented to SurfaceAppearances, similar to SoundEffects, UIComponents, and PostEffects. If that is possible, it is the only way I could see emission maps not retroactively increase the complexity of all SurfaceAppearances.
Elttob described this really well what exactly this would do:
On unrelated node, @Elttob is a shadergraph still not viable at all? if we would limit it to What OpenGL ES 3.0 and DirectX 10 supports, wouldn’t we still have SO much more that we can do? Because I remember reading the post from zeuxg where he described why a shadergraph is very difficult, not just the shadergraph specifically but user created shaders.
But considering this was in 2016, we now could do much more couldn’t we? Since OpenGL ES 2.0 support was completely dropped (at least to my information), GL ES 3.x could do far more?
I know there are still underlying issues such as shader compiler security, couldn’t Roblox just ship a shader compiler for Studio, which is used to preview shaders inside Studio, where they also integrate something like Unreal where you can preview the shader on different graphics API’s, I forgot what exactly it was called in Unreal Engine but i remember it had a thing where you could preview different shader quality levels for different platforms such as Mobile (Metal, OpenGL, Vulkan) or Desktop, Console and so on. And then if a user ends up publishing a game they send over the shader that they want to integrate into their game, which is then compiled on Roblox’s backend (also is checked to prevent any illegal operations in the shader that could be used to exploit the shader compiler?) and is then shipped within that place? Meaning that you well couldnt dynamically add any new shaders, unless Roblox adds support for loading shaders from e.g the toolbox that were verified to be safe during runtime.
Like, what is the current state on a shadergraph? Its been over 8 years since that post from zeuxg.
Edit: I forgot to include the post from zeuxg:
I did indeed read and like that reply back when they posted it. My first sentence was more about whether or not the modular approach I proposed could even work in this case. I would not be surprised if that isn’t an option here the way it was for the other things I mentioned (unless the release of SoundEffects did make all sounds take up more resources whether or not they actually have SoundEffect descendants, for example).
If the referred post did account for my idea back then, I definitely fail to make the connection.