sl-service-account
commented
1 year ago COCA Yven commented at 2023-02-13T21:59:28Z, updated at 2023-02-13T22:21:33Z
Dear Kyle
The graphics engineer, that build the materials set up when mesh came into SL was a genius, don't break his work , enhance it.
He build the interface and the code for it to be upgraded in time to more complex solutions and nodes.
I would make the point that You do already have an ocean shader, which You can easily use if You wish to bring metals with reflections in SL.
I would recommend that before endeavouring into metalness workflow, You asked Yourself a couple of questions.
- Where are we going with this in 15 years?!
- Are creators who are keeping the platform alive going to conform or retire.
-
Why not link the current diffuse channel to be a light emitter that cannot be picked up by clients cam directly but by objects, and use the specular channel as it intensity, with no decay?!...You can very well do what you are trying to do with projectors, without breaking anything and in half a day of coding.
I would recommend You first started to work in:
- Sun sky lighting system which's current dynamic range is anaemic
- Weather system, with superior clouds system.
- Plane, spherical and IES lights, with colour OR colour temperature system, and with robust dynamic range.
- Animated lights in the lights editor, so one can emulate caustics and or create animated light effects *(currently You cannot change the lights to change textures in a loop )
- Give people ocean shader, so they can install it in surface mesh prims of eg swimming pools, baht tubs, mirrors door handles, glass, plastic etc etc.
- 8K terrain textures with displacement support.
- 2 to 4k texture maps with Dynamic scaling of objects material maps in respect to a clients cam distance radius**(the further he object the smaller the materials maps will be rendered, while closer rendered at full resolution).
- Pick up the phone and give a call to Sony in Japan (I do not think Epic is the owner of the technology) and ask them to license their Nanite technology, so You get rid of poly count limitations once and for all. https://www.youtube.com/watch?v=eviSykqSUUw
- Negative scaling of mesh objects with automatic normals flipping.
- Allow to plugin environment and glossiness and specular colour texture maps... Refine Your default shininess which is by all means supposed to be a clear coat, but it is not working as such(chrome objects should be black with a smooth surface.
- Work on refraction to make glass or transparent plastic using current graphics.
- Pathtraced shadows and AO are a MUST.
- Halve the way objects detail sis rendered..... currently lots of creators especially for avatar customization, crank up the lods so for an object not to disappear, when it should be in the CLIENT SIDE to choose the rendervolumeLOD factor just as the draw distance...this dialog should be in basic settings right in the face of the client when logging in..... a. client logs in, b they get a blurred background with basic setting as at the bottom side for cam lods and draw distance, keyboard bindings etc... c. client clicks done and spawns in an already rezzed environment.
-
A Graphics Engineer is a Graphics Engineer, a Project manage will always be a Project Manager, but You will never be like a VFX artist. You need to think of You guys as the F1 Team in the paddocks that need a Test Pilot(Archviz or VFX master) because Your SL clients are the Pilots...a graphics engineer without a test pilot will never know what he is needed to accomplish, he can only guess that a car needs to go fast, but will never know how it has to drive.
Once you figured out this stuff, then You can know where you want to be in 2035, and on what new API.
SL is help up by a handful of creators who excel at what they are doing, I wont make names, but its the usual ones that make heads bodies, homes furniture, there is a best in every type of content, and they are BY no means going to move to a metalness workflow, hey are cash cows, they will certainly not update their products, and will simply stop making content, and SL as You know it will be gone forever.
While I didn't want to make names, but I'll make one to make an example just Lelutka decided to retire all of a sudden... just this one creator, You can start counting the days for the platform to stay alive.
The most successful content creators do NOT use Substance painter to make their own materials, most work on Zbrush and photoshop, I personally make My own and use V-ray Quixel Ndo.
Another issue You will encounter by adding metalness workflow, besides 99.999999% of content creators knowing nothing about Photogrammetry, is that You gonna end up with Adobe sueing the butt out of Linden Lab, because it will be inevitable, that a lot who sell textures, will just buy shaders off Substance, bake them on a 1k plane, and drop them for sale in MP, I am sure You don't wanna end up in court Vs Adobe.
Beside this, as most of Substance shaders while acting as PBR are by no means PBR, You are going to end up with a Second life that is all shiny but even more cartoonish, the ultimate goal of 3D is photorealism, SL clients purchase what gives then the closets detail to that.
If You want to upgrade graphics, You should Hire first of all a master in architectural visualization, and then have Your graphics engineers conform to what He wants and needs to make things photorealistic, starting with the lights, because lights are the most important feature for photorealistic rendering..
I hate to say it, but I saw the video, the chrome doesn't look like real chrome, the plastics neither, nor the satin aluminium, and You cant get to cast shadows off the fully reflective surface...You will end up with an SL looking like a game engine of 7 years ago, and by the time You done setting it up you will be the old stuff.
I want to finish by saying, that I say what I say because I want You to succeed, but You need to understand that Triple A games are al rendered in Vray Corona occasionally Arnold, and the game play itself is very very very old looking graphics, so You should NOT by any means try to follow, If You working on new GI (which requires at minimum 3 GI bounces and not 1 like in ur Inworld demo), You need to look beyond what i snow on the market and think 7 to 15 years from now with a well planned project which will go one step at a time to that goal. (sorry to say it but You guys are still thinking lambert blinn and phong, which nobody even considers nowadays)
It doesn't Hurt to contact Chaos Group, I know they was working on a gaming graphics engine, they will for sure solve Your ''light portal issues''(how to stop lighting to pass through a mesh object), Im sure u can budget a few million dollars to have them set u up properly, they are in specular workflow , but can support metalness and substances and have experience with true Archiviz VFX Motion picture and gaming master creators.
Please do not share what You are doing to SL creators, that was on Sansar too, get professional architectural visualization artist to work with your graphics engineers...not people that don't make a cent in SL and that think that it is cool to know this or that linden and push their creative agendas or likes to the biggest content platform that was ever made. They are not the ones who are keeping SL on its feet, and to some degree neither is Linden Lab.
Let Me say that SL as is now, if one client was to build a sim with the best of each content, no triple AAA tittle could rival its detail, photorealism, animations the list is long....
Start by fixing SL lighting system first, possibly add some environmental volumetrics, then mirror reflections is a piece of cake to do, You already have all You need.
See if You can get Nanite in SL, so We will all get 10 times the performance and need smaller system requirements.
And You will have already made 99 steps out of 100 for next gen API keeping the same workflow as always but enhancing it with more features little by little as You make improvements in performance and latency.
Just do not bring Sansar mistakes to SL, it will be the end of it guaranteed.
What just happened?
Linden Lab has decided to move SL on Metalness workflow to implement PBR, instead of keeping Specular workflow.
What were you doing when it happened?
x
What were you expecting to happen instead?
x
Other information
Metalness workflow in SL with texture maps capped at 1k is NOT viable, You would need at minimum to allow 4 k textures and quadruplicate the amount of materials ID. You can instead stick to Specular workflow, and allow to slot in maps the, environment, gloss channels reflection colour channel and add where u deem necessary. Also suites like substance painter, while having materials that work in a GI ecosystem are far from being true PBR, 99% of those libraries are made by amateurs who have no idea how to make accurate PBR materials. So what will happen in time, is that instead of getting towards photorealism, SL will gradually turn to look like poorly executed Cyberpunk games instead of looking like having a Game of Thrones looking Dragon. By the time You debug all the headaches you will have into switching to metalness, which was a massive flop on Sansar which you are about to repeat, most of the content creators will abandon the platform. Stop listening to the usual suspects that were in Sansar that are so good to voice for feature requests.. Sansar flopped because of them. if You want to implement metalness workflow then pay Adobe to allow substances in SL, if not then stick to specular workflow, which is a million times more accurate and non destructive
Original Jira Fields
| Field | Value | | ------------- | ------------- | | Issue | BUG-233371 | | Summary | Strictly for Kyle Linden: You do NOT need to implement PBR using metalness workflow | | Type | Bug | | Priority | Unset | | Status | Closed | | Resolution | Unactionable | | Reporter | COCA Yven (coca.yven) | | Created at | 2023-02-12T21:43:29Z | | Updated at | 2023-02-15T18:30:06Z | ``` { 'Build Id': 'unset', 'Business Unit': ['Platform'], 'Date of First Response': '2023-02-15T12:29:26.045-0600', "Is there anything you'd like to add?": 'Metalness workflow in SL with texture maps capped at 1k is NOT viable, You would need at minimum to allow 4 k textures and quadruplicate the amount of materials ID.\r\nYou can instead stick to Specular workflow, and allow to slot in maps the, environment, gloss channels reflection colour channel and add where u deem necessary.\r\nAlso suites like substance painter, while having materials that work in a GI ecosystem are far from being true PBR, 99% of those libraries are made by amateurs who have no idea how to make accurate PBR materials.\r\nSo what will happen in time, is that instead of getting towards photorealism, SL will gradually turn to look like poorly executed Cyberpunk games instead of looking like having a Game of Thrones looking Dragon.\r\nBy the time You debug all the headaches you will have into switching to metalness, which was a massive flop on Sansar which you are about to repeat, most of the content creators will abandon the platform.\r\nStop listening to the usual suspects that were in Sansar that are so good to voice for feature requests..\r\nSansar flopped because of them.\r\nif You want to implement metalness workflow then pay Adobe to allow substances in SL, if not then stick to specular workflow, which is a million times more accurate and non destructive', 'ReOpened Count': 0.0, 'Severity': 'Unset', 'System': 'SL Viewer', 'Target Viewer Version': 'viewer-development', 'What just happened?': 'Linden Lab has decided to move SL on Metalness workflow to implement PBR, instead of keeping Specular workflow.', 'What were you doing when it happened?': 'x', 'What were you expecting to happen instead?': 'x', 'Where': 'x', } ```