Blender nicely gives you a choice of preview items: sphere, plane, Suzanne the monkey head, cube, and hair strands. IMHO, at least three of choices like that would be helpful. Like sphere, hair (because that's so specific), complex mesh.
Sorry in advance for the lateness and length of this reply. I spent too long messing with it. I realize the thread moved on, but I still wanted to respond. Feel free to ignore it.
Soon after I've got a good skin shader for FF, I switched to Octane. When I got a good skin shader for Octane, Cycles/Superfly came in. For those from DS there is also 3Delight and Iray. None of these rendering engines understand each other's materials, except for FF and SF. In spite of me being an Octane aficionado, I find myself using more SF nowadays exactly because it understands Poser materials directly. More recently I have realized that my custom FF/SF skin shaders don't work in Octane because it doesn't support Math nodes (only "multiply" is not enough).
That's when I came to know the Physical Surface that works in both FF and SF, and tremendously simplifies skin shaders, plastic, metals and glass materials. My only gripe is that it didn't exist in P10, so it's P11 only, meaning I cannot use it in store products. The scary part of all this is that if one wants to "fully" support both Poser and DS, we would have to create materials for 4 different rendering engines, where at least there is some common ground between FF and SF, but even then with limitations because of backwards compatibility. To make it better, Iray MDL nodes were undocumented last time I checked.
From all these renderers, it appears that Cycles is the one being more actively developed, while Octane seems to be the slowest to change from the bunch because they apparently have only one programmer. But even then, I saw people discussing version numbers in the Cycles developer conference video, where features added in Poser don't synchronize with the ones introduced in Cycles itself, so version numbers have no meaning when it comes to features.
It's great to have options, and many of us always wanted to have higher quality alternative rendering solutions for Poser/DS, but as things stand now, it's complicated.
Oh, I definitely think that it all depends on your interests, strengths, weaknesses, and goals.
I've been able to quickly create the kind of skin you get from the PhysicalSurface node in most SSS supporting renderers. Mitsuba, for instance, took me about 15 minutes to make that kind of skin. It's just SSS and blurred reflection. What usually makes good skin hard to do well is the multiple layers, varying relationship between subdermal and dermal scattering depending on areas and individual skin type, lack of internal flesh in areas like the nose and ears, and (in older people) the gathers of just plain skin, away from the flesh. Most serious skin shaders have multiple layers of scattering and back scattering to consider. And most of those don't handle anything but pale, pink skin well. They don't really take into account the role of melanin in making the skin more reflective and less permeable (but not opaque) to light. But the PhysicalSurface node only supports one layer of scattering.
If all you're talking about is SSS and reflection, wax, plastic, marble, and skin are all pretty much the same. Which is cool for many, if not most, people, just not me. And either as a PhysicalSurface, Firefly, or Cycles material, not at all hard for me to make. For the latter two, you need one node for the SSS, one for the reflection, and one for the Fresnel. That's it. It's finding the sweet spot for the settings that's the trick.
There's lots of reasons you might want to take your materials beyond the PhysicalSurface node in both Firefly and Cycles. Needing to use refraction, for instance. Or just wanting to understand how to put together a material in general. All Blender material tutorials I've looked at go a bit beyond what the PhysicalSurface node can do in terms of features (like issues with Fresnel and roughness). But sure, if all you're trying to do is make something that works, and it does, then use it. It especially seems useful for Firefly. I can't find a way to get colored transparency in Firefly without the PhysicalSurface.
In my experience, the tricky part of materials comes from particulars of the renderer. For instance, Cycles already has a Glass node. Problem is that it has weaknesses. If you're fine with those weaknesses, go to. But if you don't want dark artifacts in your glass in Superfly, you're going to have to learn how to get around its Fresnel refraction problems. And Firefly, well, Firefly doesn't have caustics. So its refraction is either overly shadowed or not shadowed at all. Glass, gems, and water take tons of tricks to handle without problems I find glaring. And that's not even addressing the fact that Blender's "Full GI" settings for Cycles gives you 128 Transmission bounces which render just fine, where I've never been able to go beyond 5 or so in Firefly without massive slow-down. And I've seen some serious problems with plain vases and such just below 16 bounces. Iray, on the other hand, has no problem _at_ all with refracting materials and renders them like a dream, even with dispersion. But I'd rather do skin reflections in Cycles or Firefly than Iray. There's pretty much no good solution for proper blurred and Fresnel controlled skin reflections in Iray.
Beyond an initial learning period, I don't find making initial, renderer-specific shaders that time consuming. Plastic and metal are pretty easy, as is the fake, non-refracting glass the PhysicalSurface node supports. About 80% of my work has always been compensating for a renderer's weaknesses, with the other 20% being the time spent getting to know the range of responses of different settings _really_ well through repeated testing (which is less of my time because it's mostly a one-time thing). For instance, most of my time learning Iray was trying to get rid of the awful edge artifact blurred reflections have. I assumed it was me, and that I was doing something wrong. I managed to isolate the issue (that's how I know it's the blurred reflection setting), and learned that there's nothing I can do other than not use that setting much and rely on bump and normal maps for softer reflections. I'm messing now in Blender/Cycles so I can get a handle on how normal maps work (they need to plug into a lot of stuff to work properly) and find a good work around for that Fresnel refraction problem.
I'm pretty clear on the physics shaders mimic. But that's me. I know most people aren't that into the material physics/optics side of things, and don't have any need to be. My advice is for those who either want the power to make their own robust or complex materials now or think they will in the future. Or just want to be able to make skin in the next awesome new renderer that comes out.
I enjoy making materials in good renderers. Except for having to deal with the DS interface (so not my thing), multiple PostgreSQL instances taking up processor power and memory for no benefit, and lack of a PFE equivalent, I don't find it hard to support shading in Iray. I'm pretty happy with Cycles, and I was happy with BI. Luxrender frustrates me a little in that it shares Iray's weakness with living SSS materials, but I'm more put off by its render times (and my inexplicably terrible experiences exporting scenes with Reality). I generally find material work more of a pleasure than anything. I usually enjoy Firefly, too, though I wish it could do more than it does. The FF GI blotch problem has never stopped _immensely_ frustrating me, and the lack of caustics makes a lot of stuff impossible to do accurately. By far most of my fakes and crazy material work has been trying to compensate for Firefly's lack of caustics. My tiaras were a joy to experiment with in Iray. Firefly, much less so. That said, I can't handle DS/3Delight at all.
In my experience, assuming equal shading feature support, the difference in procedural textures and layering methods are the most frustrating aspects of translation. That's killer for me. I don't think I can make my satin/velvet/lace layered material in Iray, for instance. I think _someone_ could very easily, but I don't think I'll be that person. It will be a big enough deal just to get to a velvet and a satin I'm satisfied with, because so much of both of my FF satin and velvet is procedural texturing and color math. Universal shaders would be _so_ cool, but what I'd _really_ love is consistent procedural texture implementations.
Cycles is being actively developed, but there's only so much that will be meaningful to the content community at this point. They're not touching on caustics or the Fresnel problems. That same video only mentioned about three things that will matter to most of the Poser community. One is the "Disney" node that will combine all aspects of shading into one node. I'm rather skeptical about that working well because I've seen lots of versions of this in the Blender community. Blender's node system allows you to make custom nodes with custom inputs and outputs (like Poser's compound nodes). So you can already make a node like that yourself, and many have. They've all been so complicated and unwieldy that most people still use particular materials. But you could duplicate them in Poser right now, if you wanted to. Maybe the built-in uber/disney node will be better, but considering the community is so involved in development, there's not usually a huge quality difference between what you find as random add-ons and what gets added as features. In fact, the former often become the latter. The second is the metal node. Again, I'm waiting to see how that stacks up to what's already out there in terms of custom metal nodes. Third is displacement, which is already in Cycles. I figure we'll have to see how Poser handles that.
So you've got two advancements you can find on your own right now very easily (look up Blender UberSurface, IIRC, and Cycles PBR) and bring into Poser today, and one advancement that Poser will almost definitely announce. Other than that, maybe I'm totally wrong, but I don't think the content community is going to use new motion blur features or any of the planned camera features.
I would love to see more of Octane. It's obviously an amazing renderer, and I've seen such great work done with it. If it weren't so expensive, with at least two licenses to keep up on, I'd have gotten it a long time ago. I noticed that even in its beta and maybe even alpha phase, Octane always seemed to have a very balanced focus. And everyone says it renders _so_ fast.
I agree things are complicated now. I think my vote would have been to implement the Cycles nodes in the background of the Poser nodes, add the Cycles nodes that didn't have Firefly equivalents to the general node list with an asterix or something, and internally address the switch between Superfly and Firefly and the input multiply vs. input override issue. Three root nodes and an unnecessarily difficult selection path to get to Cycles nodes is far from ideal. I think the PhysicalSurface node kind of helps, but I also think it hides lots of basic properties, and makes others, like anisotropic reflections and refraction, impossible to implement correctly. There's no good way to transition from understanding it to understanding either Firefly nodes or Superfly ones. In their tutorial video, Smith Micro said it was essentially a modern version of the Simple tab in the Material Room, and I think that's accurate. I think it would have helped people if the tool to make it easier to make materials in both renderers also made it easier to understand how to build materials in general.
It's funny your experience is so different than mine, although I image I know why, and I think I have found one more reason to maybe start moving away from google as my search engine. I suspect that since I spend so much of my time on sites related to Unity3D, my search results are 'tilted' towards things that reference gaming. And blender is definitely used massively in the indie game sphere, precisely because small studios don't have to worry about buying a license for everyone on their team.
You've obviously spent a lot of time looking at the various PBR renders, so thanks for all the info.
I think a lot of it comes down to how you use Poser. I was afraid of the Material Room until Poser 11 came out. Actually I was still afraid, but I loved what Superfly could do, but I had too many things that looked completely wrong if I just tried to use the Firefly materials. So I started to learn how to adjust them to work in Superfly, and the Physical Surface Node was so much easier for that.
But when I find some time, learning cycles is definitely on my list. Heck, I really need to sit down and learn how to do more than two things in Blender period.
Oh, I think I've miscommunicated. I wasn't saying there aren't lots of people in the indie game sphere that use Blender. I know there are. I was saying that most of the Blender community isn't indie game producers. Cycles was explicitly designed by Brecht for video and film as well as stills. Rendered animation seems to be the main focus when it comes to Cycles development. You go to BlenderArtists, the main Blender forum, and in, say, the general WIP forum, game content is in the minority (below 10% I'd say from the first page). But you see a lot about movies, sculpts, and stills. When I looked up the issue of caustics, I saw a whole bunch of people argue against them as unnecessary on the basis of them being too slow for film and therefore unnecessary. Games never came up. There's a whole section at BA for games, and the BGE is where a lot of simulations started, but the two big professional communities most (but far from all) Blender developers and users seem to focus on are the Zbrush/sculpting community and the animation community (with something of a bias towards Maya). The most popular resources for material rendering knowledge,
BlenderGuru and
BlenderCookie, have (in my experience) focused on stills and short animations. One of the best free material resources I've come across (not all physically accurate, but all good as a start),
Chocofur, seems to be focused on stills.
I totally and completely agree with you about it being about how you use Poser (or any other 3D software). And I know I'm not the norm. That said, I've worked a _lot_ with materials, in a lot of different renderers, to the point that I'm very comfortable translating basic shading principles to new renderers. I figure the roadblocks and best practices I've encountered on my path might be useful to someone out there who, for whatever reason, wants to have similar skills.
I think most choices are about knowing the pros and cons, but for some reason most material stuff seems to become a discussion of right and wrong. I've spent way too long making wild shaders to compensate for Firefly's weaknesses, and seen too many _incredible_ but totally fanciful shaders (my favorite ATM is the
crystal turtle) to be a realism/physical accuracy purist. I can make myself crazy in terms of realism, but I that's more of a weakness than a strength.
Let me say kudos! for overcoming your fear of the Material Room. I think fear is the main thing that separates kids from adults in terms of learning. The more comfortable you can be with not knowing what you're doing and making lots of mistakes, the more you can learn. I'm pretty much in a perpetual state of not knowing things I want to learn.
