Porky Gorky

Hi @arton Rotaru I realise you posted this comment two and half years ago but I see both you and the thread are still active so I am not sure this is officially necroposting. Anyway… Please can you go into more detail on this process using 3DS Max? I am unsure exactly what texture is being projected? Are we talking about rendered images of the final model on alpha backgrounds, one for each side? Also not sure what you mean by projected. When you say imposter model, these are the polygons that make up the lowest LOD i assume and do these polys need to be hidden in all the higher LOD models? What % of the UV space would you use for the imposter UV’s on for example a chair with a single texture at 1024x1024? Thanks alot for any info you can provide, it seems like an excellent idea, I'm just not sure of the best method to achieve it. Thanks. Edit: Here is arton's answer if anyone is interested.

I agree, ZBrush has better computational power and can handle multiple millions of polygons whereas Blender tends to struggle, ZBrush beats blender hands down in the level of detail that can be achieved. When it comes to digital sculpting the only thing Blender beats ZBrush on is price.

If you can afford the license for ZBrush and Substance then you should use them both. A common workflow involves sculpting a high poly model in ZBrush, then create the low poly model in your 3D modeling software, then use Substance Painter to bake your high poly to low poly maps, generating as many maps as you can including normals, curvature, thickness, AO and position. These maps will feed into the smart materials in Substance Painter offering you more functionality when texturing. You can also use Substance Designer to author your own tileable basic materials or smart/procedural materials for you to use in Substance Painter or you can import basic tileable PBR materials straight into SL.

I think a reasonable solution to the problem is for LL to deploy a new decent, default grid wide EEP setting that doesn't suck like the current one. We need a base standard default lighting model that both PBR and BP creators/users are reasonably happy with. This should offer consistency for creators and users and consistency is important for physical based rendering, People are still free to adjust their own EEP settings and creators are still free to support those niche requirements. Just give us an acceptable and consistent base standard to work from.

I arrived back in SL 2 weeks ago, drawn in by the PBR buzzword. I wondered if it was worth my effort to make my 3D PBR content compatible with SL as it’s generally made using a metallic/roughness workflow. After 2 weeks of testing and learning I fear you may be 100% right. Not a complete waste of time though, i’ve been gone 12 years, but I want to get some land, build something cool, as long as I have 100% control of the environment then the PBR will work.

Nope, but good idea, I should have tried.

I don’t have land to test EEP settings. A while ago I attached a bunch of objects with various PBR materials to my avatar and visited dozens of parcels with different EEP settings to see how they look. The results ranged from amazing to terrible, however the takeaway lesson I learned is that there is no consistency at all in the environments within SL apart from with the default settings which makes everything too blue. I was left with the question; How do I create and sell PBR houses when there is no standardized lighting model? This question led me back to these forums after many years away. I found content creators had already answered the question, they are going to ship their houses with recommended EEP settings. Seems like a reasonable solution given the problem at hand. So then I consider the new PBR prefab business. I sell a prefab house with recommended EEP settings. The end user is going to need furniture. If they go buy furniture made by another creator (who used different EEP settings) it may not look right, Ok I think, so I have to sell furniture too, oh and trees and landscaping and all the things! Then I start thinking about the avatars using my house. Once avatars start getting PBR’ed up to the maximum, will they look good in my EEP settings? Maybe I should sell avatars and every possible accessory they could ever want, that will solve the problem! So what I need to do is design and sell every single PBR product my house owner might ever need so that everything looks good in my recommended EEP settings. At this point I just continued making content for Unreal Engine where THINGS MAKE SENSE!!

On the first day of testing Reflection Probes I tried placing one large probe over a whole PBR house and it didn’t work. At first I thought it was working (when viewed from a distance) because I was able to reduce some of the blue sky color which was reflecting on the materials which made them look nicer and more accurate. However once I got in close the reflections were not accurate at all. Reflections were being captured from outside the walls and from adjacent rooms etc. I’ve tested more since then and learned that the reflection probe creates a volume. It is from the center of this volume that an image is taken of its surroundings which is then used to generate the reflection and ambiance effects. So there would need to be a reflection probe (or multiple probes) in each room or space within the building in order to achieve accurate reflections and ambient light bouncing.

I think I am starting to see the bigger picture thanks to the last 2 posts, so thanks for that! So far all my PBR tests have been done in sandboxes where I can’t control the EEP. The PBR reflects the sky color too much so I have compensated for this by using the reflection probes to make the materials look more accurate. I seem to have got into the mindset that the Reflection Probe volume has to encompass every PBR material in order for it to look correct. I think the way forward is to explore a balance between more suitable EEP settings and more strategic use of Reflection Probes. More testing required.

I look forward to seeing the PBR buildings that creators produce however I fear they will be too restricted by the limitation of the Reflection Probes to be that interesting, I like curves and organic and unconventional architecture but I cannot see how that can be achieved right now if probes need to be included.

Yep that is true. I meant that only PBR materials within that volume would be affected by the reflection probe, compared to PBR materials outside of the volume. I wasn’t considering non PBR things.

According to the SL Wiki PBR page the Reflection Probe needs to encompass any PBR surface that you want to affect. It says: “Select an appropriate probe shape for your scene. For example, if you want to create a reflection probe that covers a room, select a Box shape, or for other objects a Sphere shape is recommended. Move the reflection probe and resize it as necessary to fit your desired purpose. For the room example, resize the box so the probe just touches the walls, ceiling and floor of the room.” It also discusses using multiple Probes to fill a complex space: “At times, you may wish to place a reflection probe in an area where the reflection probe types (Sphere or Box) do not conform to the shape of the area. For example, a Box probe in a loft room has a triangular shape, which results in "probe bleed" wherein the influence volume affects an area larger than what is needed, and thus bleeds out onto the exterior roof, which is undesirable. In these cases, you may need to place multiple probes in order to blend together the sample volumes to achieve the desired result. In the loft room example, it's best to start with 3 box probes, which cover the floor, and each side of the roof. Then, use an additional box probe as "fill" to cover the gap inbetween the 3 probes. The main fill probe may need to be rotated at an angle to get the most coverage possible. If it is not possible to fill the gap between the 3 probes with this fill probe, you may wish to add more box probes to act as "secondary fill". It seems we need to use the Reflection Probes to create a volume and only PBR materials within that volume would be affected by the probe/s. If our avatar (or camera) is placed outside the volume then we can see the effect the probe is having on the materials. If our avatar is actually wearing PBR materials then they would need to step into Reflection Probe volume in order to benefit from it. These “volumes” need to perfectly fit interior spaces to avoid them bleeding into undesired areas, which I think is what I was moaning about in the first place. 🙂

Yes. I envision a SL in which 18 months from now, all the top prefab creators are competing to see who can make the best looking box shaped house. 😀

Yeah, it seems that if you want to sell PBR houses they need to be made of rectangular rooms so the Reflection Probes fit, they need to include lots of built in lighting to account for the Reflection Probe darkness at night or in enclosed spaces and you need to include a set of optimized EEP settings. It’s pretty game changing. Here is something worth thinking about. 6 months or a year from now when more people are on PBR enabled viewers, they are likely to start buying PBR products, maybe avatar accessories or home furnishings. They take their new PBR shinies home to their Blinn-Phong house that they purchased from you. Those PBR products likely won’t look like they did in the store because they are not being viewed within a Reflection Probe or appropriate EEP. So your customer has 3 choices: 1) Forget about using any PBR products at home 2) Set up their own Reflection Probes into the “legacy” house you built and change their environment setting themselves. 3) Go out and buy a new PBR house with preinstalled reflection probes with customized EEP settings i.e the optimal environment in which to view PBR materials. I guess my point is, if someone purchases ANY PBR product, they will need Reflection Probes and the correct environment settings at home to view it correctly in your Blinn-Phong house.

Needing to use Reflection Probes in interior spaces is a real hindrance. Completely ignoring the issue of needing lots of additional lighting, using the probes seems to be extremely restrictive. We have either a sphere or cube to choose from and they can be scaled and rotated and that is it. You can’t seem to overlap them either without undesired results. So if you want to use them in a conventional house with rectangular rooms they work ok. But what about tapered walls, curved walls or a curved roof or even just a corner that is not a right angle? Maybe I am missing the bleeding obvious here and I really hope I am, but it seems to me this completely stifles creativity when creating PBR buildings and designs will need to be restricted in order to conform to the limitations of the Reflection Probe’s shape. Also, is any creator who sells PBR prefab buildings expected to include the Reflection Probes positioned into the final product? It would seem like the right thing to do and would offer the best result for the end users.

I think we should be encouraging people to adapt their environments rather than the PBR materials reflectivity. The whole principle behind physical based rendering relies on surfaces approximating the way light behaves and reflects in the real world, generally relying upon data values derived from real materials. If a PBR material is well made then it should reflect the light in an environment realistically. If it doesn't then the fault lies with the environment and not the PBR material. Re. the OP’s problem, have you tried using a Reflection Probe and possibly a bit of local lighting? It can make a nice difference.

Would it be practical to provide demo versions of PBR products for customers to test first? Or would they ignore the demo just like they ignore the description?

It turns out this image was taken in a sandbox that had some sort of atmospheric fog effect in the environment which I am guessing is what caused both the light bleed and the extreme level of darkness shown. I’ve tried and failed to replicate the problem on the Beta Grid and in other sandboxes on Agni and it is now working as expected. The last time I was active in SL we were using Windlight. I’ve been reading about the EEP and I want to test how it works in relation to PBR materials as I'm relatively clueless. I assume in order to access the environmental settings I need to own or rent land. Back near the start of SL I used to own a parcel on the beta grid that persisted for years but it seems to be gone now. Is there any way to buy land temporarily on the beta grid? If not, can I just buy a small parcel on the mainland in order to get access to EEP or are environments controlled at a region level or something? Thanks for any advice you might have.

Hello, I am stood in an enclosed rectangular space with a reflection Probe filling the volume. There are no lights, it should be dark. My problem is I keep seeing light bleeding in along some edges. Why is this happening and how do I stop it? Thankyou.

Excellent, thank you very much!

Hello, I used to create content in SL a lifetime ago and recently read somewhere that PBR has been introduced to the platform which has piqued my interest. My question is, can normal maps be used within PBR materials in SL in the same conventional way they are used in Unreal or Unity? Thanks

Thanks, thats very useful to know. I shall invest some in experimenting. Cheers.

Normal mapping has arrived since I last looked. Been reading though threads about the various problems/ issues people have with their creation. So I was wondering how easy/hard it would be to import some stuff I created for Unreal Engine. The normal maps are generated either directly from Zbrush or using a standard high poly to low poly bake process in Max, all set to a tangent world space. The normal maps all work as expected in Max, Maya, Unreal and Unity. In theory, can I import these maps into SL and expect them to work out of the box or will they require some level of optimization? Thanks.

BlueTrianon wrote: I do remember that it was said that they were especially looking for people who have created experiences in SL. Beyond that, I'm sure being able to present skills in areas such as mesh creation (earlier groups were required to know Maya specifically, I believe, but I wouldn't be surprised if the next group were expanded out to 3DS Max, Blender, etc.) or C# would help. Has there been any official mention that other packages will be supported? I animate in Maya but hate modelling in it. It would be handy to have Max supported too.

I think he did a good enough job considering the circumstances and the state of SL when he took over. When he was hired I would not have believed he would take the company as far as 2014. For a while there he seemed to breathe a bit of life and drive back into LL even if it was partially focused on products other than SL. Unfortunately LL and SL in particular needed someone to do more than just a "good enough" job. SL needed a visionary that would revitalise, if not revolutionise the whole product in order to regain the mass appeal we experienced during the boom years. Rod completely failed to deliver there. I wish him the best of luck. I will have been in SL 10 years this coming May. I have seen SL blossom and then explode into life only to be mismanaged into a long and apparently drawn-out death throe. I am really excited that once again we stand upon the precipice of change. Once again there is the slim chance that some "visionary" will come in and completely transform SL into a product that has an exciting future. On the other hand I am also very worried that we will just get some suit wearing, puppet-caretaker CEO that is only there to oversee the ongoing milking of their once great and glorious cash cow.