| 2 Jan 2024 |
 David Huculak | I guess to start you could take the metallicness to scale how much the reflection is visible and the roughness map to scale how much the reflection is blurred | 17:28:00 |
| David Huculak | Just like in IBL/environment maps | 17:28:11 |
| David Huculak | I assume you're talking about SSR | 17:28:27 |
| David Huculak | 
Download 1000006411.png | 17:28:35 |
 norlock | Euh I have a very simple reflection shader now where I use the roughness to determine the reflection. But I guess I have to blur it, do you use mipmapping for that or irrediance map? Because I'm not sure what that really suppose to solve | 17:39:51 |
| norlock | https://github.com/Norlock/sparticles/blob/64c91fd73298ae7b2096b528d1f7f90180ed6498/crates/sparticles_app/src/shaders/particle.wgsl#L111 | 17:40:32 |
| norlock | Is what I do now | 17:40:38 |
| norlock | * Euh I have a very simple reflection shader now where I use the roughness to determine the reflection. But I guess I have to blur it, do you use mipmapping for that or irrediance map? Because I'm not sure what that really suppose to solve. But I tried to pass the env_color through the brdf function but it result was not good. I guess I make some mistakes | 17:43:39 |
| David Huculak | I can't remember the exact specifics, but learnopengl has a great tutorial on pbr/IBL that explains it | 18:20:41 |
| David Huculak | You basically create a cube map as a lookup table for each environment which lets you look up the lighting for each roughness value and then another lookup table for the metallicness I think, then you combine them with some fancy math that I don't understand | 18:21:51 |
| David Huculak | And yeah mipmaps are involved for the different blur/roughness levels | 18:22:17 |
| norlock | ah ok thanks, I will take a better look at the tutorial. I didnt really get the convolution map it was way too heavy to run even just for one frame it completely freezes for a second but I can improve that probably | 19:53:24 |
| David Huculak | No probs! You only need to generate the lookup tables once, meaning it can be done offline and saved to disk for quick loading. You could also time slice it (split the process up into small pieces and limit how many pieces you run per frame) if you want to prevent frame spikes | 19:58:14 |
|  szatang joined the room. | 21:42:44 |
| 3 Jan 2024 |
| norlock | Yes i might need to split up because it kind of freezes my pc everytime i resize. Or use less precise sampling | 16:02:53 |
 Ralith | why are you recomputing it on resize? | 18:44:42 |
| 4 Jan 2024 |
| norlock | Yeah i dont have to | 14:49:51 |
| 6 Jan 2024 |
|  @sofia-snow:matrix.org left the room. | 22:02:12 |
| 10 Jan 2024 |
|  pverlind joined the room. | 15:37:51 |
| 13 Jan 2024 |
|  psydroid changed their profile picture. | 20:52:21 |
| 14 Jan 2024 |
|  gaast changed their profile picture. | 20:17:00 |
| gaast changed their display name from gaata to Vash. | 20:17:33 |
| 15 Jan 2024 |
| psydroid changed their profile picture. | 08:29:59 |
| 18 Jan 2024 |
|  @atlv:matrix.org joined the room. | 03:56:34 |
| @atlv:matrix.org changed their display name from atlv to vero. | 03:57:45 |
| @atlv:matrix.org set a profile picture. | 03:57:49 |
| @atlv:matrix.org left the room. | 03:59:13 |
|  om1kron joined the room. | 16:15:20 |
| om1kron | guys what crate do you use for stuff with keyring or storing some login credentials securely | 16:15:40 |
| om1kron | https://github.com/hwchen/keyring-rs This is what I was looking at | 16:15:53 |
