This is awesome! I’m super excited for mesh deformation and custom materials! I can’t wait for these features to be released!
these things used to be a dream to me
im so glad they arent anymore
now that i think about it, theyll most likely become overused, but im prolly gonna end being one of those people who overuse it anyway so yeah
It would be nice to have a way to make the Albedos and all other maps be scalable, we might need to use Seamless features in the future.
i could recommend taking advantage of that to make something neat out of it, but other than that, yeah i agree with you on those seamless features
I’ve been hoping for a mode gravel-y texture for use in gravel roads and gravel train loads, and I found a set of textures with diffuse, normal, bump, and spec maps. I applied the normal map to the SurfaceAppearance, and converted the AO map to be black and alpha (to create a colorable material in the translucent areas) and applied it as the color map.
I expected the color map to overlay on top of the part color in cases where the image is translucent, but that didn’t happen. I hope it will gain this feature so I don’t have to create different color maps for each different color gravel I need in my game.
Tiling is also important to me.
Thanks for continuing to improve the Roblox engine!
Nope. “Future Is Bright phase 3” is coming, but won’t be ready any time soon. And we decided to use metalness over specular maps since metalness covers most use cases and uses significantly less memory.
This property wasn’t meant to be visible. Internally it holds a reference to compressed textures as an optimization.
Most likely. Since SurfaceAppearance is a huge change already, we’re going to ship the minimum API that makes it work in v1 then expand to common use cases afterwards.
It’s too soon to say. We’re investigating multiple options to improve reflections.
Not at the moment.
Difficulty ranges from hard to impossible to change Roblox APIs after we’ve shipped them. That’s why we’re shipping the minimum number of properties that make this feature work instead of shipping a bunch of extensions all at once. There are plenty of improvements we want to make in the future, but we need to collect feedback and adjust our course along the way.
This is on our radar.
How feasible would it be for every part to have a SurfaceAppearance with all of the maps? How will this increase load times? I am definitely all for this, but I’m pretty sure the biggest part of game downloads come from textures…
For some reason it wont let me manipulate an imported mesh rig, that has already been boned and stuff. Help?
What about tiling for seamless materials?
It seems there isn’t a way for multiple anchored MeshParts to reference the same bone group because they don’t use welds or other joints intended for unanchored parts. An “anchored weld” that attaches an anchored part to an Attachment
would be useful for this, as well as for other use cases involving cframing anchored parts; It could even supersede the Massless
property.
Character customization is an extremely important aspect of many games, and texture compositing would make this so much simpler to implement. I want to add makeup/face paint options and other patterns that overlay over a SurfaceAppearance
. It also doesn’t look like it’s possible to apply a color multiplier to a SurfaceAppearance
’s color map which is necessary for inclusive character skin tones. I understand that the image processing for this is non-trivial, but I’m sure there’s a solution for gradually increasing resolution, reusing results from similar colors, or even just omitting overlays depending on system performance.
It also looks like Bone.Transform
is redundant in addition to Bone.CFrame
because the initial pose could just be baked in to the animation to improve performance. My game’s custom animation system does this because it removes the need for a toWorldSpace
operation for each bone. This likely doesn’t affect performance much on the C++ end, but it’s still not the simplest animation solution. If an additional transformation is needed, an intermediary Attachment
or Bone
could be used instead.
Thank you for responding. Will the Roblox roadmap be updated anytime soon, potentially with information about FIB v3?
It looks like PBR properties are clamped to [0…1] now.
Take a look at the interesting effect this user created by using values >1. Please consider leaving these properties unclamped. Unless there’s some dangerous technical issue with it, it seems like an unnecessary loss of creative freedom to clamp these values.
Really nice and is a wonderful edition to studio. Only problems are that I crash when uploading characters, vehicle imports become completely garbled, and it lags severely in run mode when deforming a car. Would love to see this feature in working condition though so I could utilize it.
Nice! The only thing this needs is some sort of way to edit individual faces. Not sure about moving away from humanoids, though. The current system needs some tweaking to reach the tightness and flexibility of the humanoid controller.
Do we have an estimated release date for skeletal skinning for use in live games? Or is it already live??
Also, no idea what’s happening, but I’ve been unable to get it work. Unfortunately the packages you included do not work due to having a tri count > 5000, and my own personal uploads are met with “cannot find [xyz] object” when uploading a custom import with the object as a vertex group? Any suggestions or is this a bug?
I’d love if there were an option to make the image tile (like the way textures work) so we can make our own seamless materials like in my material extension pack which uses a combination of textures and roblox materials.
An incredible update!! We’ve been needing these for a long time now and the only thing I can think of is… wobbly jello