I couldn’t agree more and if that doesn’t happen many people will search for a alternative. eg. YafaRay. Or, as the man mentioned any raytracer but Cycles. I mean the point for me is a believable material to the average viewer. Done with the least setup time. Usually and consider that underlined I don’t need the versatility of Nodesville. But, having it is sometimes a welcome addition. Meaning Cycles would be a great render engine for all of us with the addition of some basic materials with dropdown presets.
Then like harleynut97 mentioned Blender would have a one two knockout punch. A scanline for speed and certain looks and a outstanding raytracer. But, then sadly reality sets in. With one and sometimes two developers are we asking for the impossible.
So if this is just for someone wanting to start out in cycles, what if I or others made a startup file and posted it here (or stickied it) and in the materials drop down it would already have a long list of materials to pick from. Would that work? What materials would you want to see in a list of 10? List of 15? List of 20?
The argument has always been filesize, although I can’t be certain that was ever said by actual developers.
My solution to said problem is easy though: offer a barebones version, and one with ‘content’. This can include materials, 360 hdr images for background/image-based lighting, and a few example files to show off certain features.
This way bandwidth-conscious people and longtime users can still get the barebones, program-only download.
I don’t suggest to have full on materials which include the HDRs or even image textures at all. File size wouldn’t be affected at all, this would be closer to the mindset of procedurals… these would just be pre-wired, assigned to the right slots and if the user wants to add images or hdrs to the mix that’s up to them.
What if we didn’t even need to add an extra drop down menu but simply include a way to save node setups from within the material settings?.. sort of like how you can add presets to the render tabs (dimensions, codecs etc. with just a plus sign). Then only a few of these would ship with the Official release.
I think this would also help people take their work to the next level, they don’t need to be doing all the basic setups which slow them down. By the time it’s setup and tested, again and again, they’ve already spent too much time. Also this would help to encourage new and old users alike by giving them better results quicker:)
One day we could hopefully put some Material wizards at the helm, like I believe I already mentioned, to be responsible for the “trunk” version presets (extremely limited 4 or 5 types). This would then take some weight of the shoulders of the developers… we just need the tools.
I think this would be useful, but only if the materials have been vetted and are good. My definition of good is “physically plausible”. The problem with many material libraries is that the included materials often have little if any relationship to the real world. In a recent lecture John Carmack was talking about how when ID started building their material library they tried to use materials that shipped with commercial programs but soon discovered that most of them were useless.
There are certainly uses for nonplausible materials but artists should be consciously making the choice to use them rather than blindly choosing from a dozen almost, but not quite right, versions of metal or plastic.
I guess I’m perplexed on the file size issues people are bringing up… are they referring to the size of the download for the Blender install if a decent library is added to the trunk?
In my opinion…Who cares if it adds some weight to the file size of the download, if it is going to add to the ease for new users, and decrease the setup time for all users. It takes me about 5 minutes or less to download blender.
As soon as you start adding images to the materials the download size will go up. Taking you five minutes now for a 200 MB folder (pre-compression)? Are you willing to spend 30 mins to an hour to download it? I am not. (I download one every day though) A few hundred high res and HDR images and you’re easily up to that (Remember we can’t really use jpegs to good effect so it has to be a lossless format). I imagine the load on blender.org would break their site. So we really have to stick to basic if it’s going to be practical. Procedurals only. A hundred materials? The material list is not submenu’ed so wouldn’t that trip people up who just want a nodetree in the editor to get started?
This is my suggested list, all procedural:
Rubber
Plastic, polycarbonate
Plastic, polyethylene
Plastic, clear
Glass, clear
Glass, Frosted
Glass, mirror
Wood, furniture
Wood, lumber
Wood, paneling
Polished metal
Brushed metal
Car paint
Sand
Earth
Rock
Neon glow
Cloth, smooth
Cloth, coarse.
This is my idea for what might be workable. If you are about to reply saying that you couldn’t live without your ‘glassy alien hair’ or ‘100 year-old peeling wallpaper’ material in the list - well this is not to help anyone but those who are just stepping into the cycles bathtub and are intimidated by nodes. Especially not those who know how to put a material together but are just - well, perhaps I’ll be nice and say they are pressed for time.
Wouldn’t bother me a bit… especially if I knew I was getting good materials and setups out of the extra time.
The average user doesn’t download any builds other than the official versions…again I say average user. You bring up a good point on maybe it would tax blender.orgs bandwith… but from my experience when a new release comes out its a little slower that day or even a couple days after… after that business as usual. I don’t know what the additional cost would be to them, but a decent material library would definitely attract and keep more users.
Truthfully I’m impressed with very few materials that use procedurals, There is nothing wrong with using them, but I would think that you should include some setups that also are imaged based. What would be wrong with including the set up but not the texture?..you hit a preset that says wood, image based… It builds the node tree for you, puts an image texture node in where it belongs, and the user just adds there own image. Give me a set up to add a decal easily and quickly.
It’s a shame because that online material library was really nice, but the developer Peter comes and goes, and really is not active here. His builds break with each release, and than he has to fix them, but it’s a hell of a nice add on and he did a wonderful job on it. He just needed a bigger selection of materials.
For one thing, the best way to use images is with a UV unwrapped object… That’s one step up from brand new user IMO. So the material I use will always have a UV texture coordinate in the node tree. When a brand new user tries to use that - well I’m not even sure what will happen. Got to try that out when I have time to waste.:evilgrin: I’m guessing - purple object. Lots of posts about that on here!
I totally agree that file size when including images within the materials is a major concern… especially in parts of the world that have very limited bandwidth. Do Yafaray, or Lux include images within their materials? I don’t think so.
We should be very careful as well that we don’t ask for everything, and a pony If we don’t make a modest request that’s easily doable it’s less likely to happen, in my experience anyway. The best way to go is find the solution, then present it.
The point is not to make the materials for you, it’s just to get started. Again I’ll using my mesh object analogy… we start with a cube and work from there… it’s just meant as a convenience.
Oaky so I think there is even a simpler way to do this: In my default scene that loads up every time I launch Blender, I have a bunch of material groups already in that scene. These are just basic starter materials like Skin, fabric, glossy + diffuse, glossy + diffuse + transparency, glass, fake metal, etc. So then whenever I start modeling, I can just go grab one from the materiel groups menu, set the color and add textures and so on. I think all we need to do in order to get this going is find a collection we all agree on (keep it small and simple) and have it loaded into the default scene that ships with Blender. That way, it doesn’t have to involve the developers who are already very busy with other things. Of course, we would have to get permission I guess. We could simply use this thread to create a scene with all the different materials in it, test it out and when we think it’s ready, submit it to the devs.
Actually, I don’t know how we have gone so long without this happening yet.
I have a similar system, but I have the groups in a different file so I just append the ones I need (I have enough now to force scrolling in the groups menu).
I also have been sharing this file on this forum (see my sig for the link), so that everyone can use it, my main goal though is to make them more generic than a skin shader so as to give more flexibility in how they can be used. (though along with shaders that can be used alone, there’s a lot that deal with masks and textures too).
Perhaps this would be better solved by doing rather than talking/writing.
Make a list of presets that would be useful. Then make those materials. Share them here and everywhere… let people tweak and refine them. Then, with the materials made and a large-ish group of people vouching for their usefulness, make the pitch for that specific set of presets to be shipped with Blender. If you want to go the extra mile, find a coder to cobble together the handful of lines of code that would make these presets easily accessible… then submit the whole mess as a patch (or add-on, if you find it can be done easily with Python).
This stuff doesn’t happen on its own… and putting together a set of materials is something that doesn’t require a developer. Put the time into making those materials. It will make your argument far more compelling.
“Like in the USA. I’m stuck with 10Gbs per month!! and it already tacks me 30 minutes to download blender”
The fact that ISPs are getting away with that galls my ass but that’s another thread.
I for one am getting a little confused here. When I mentioned presets and materials I wasn’t thinking about forty different car colors. Nor procedural textures. It was like maybe five presets: Glass, Glossy, Defuse, etc. and sub settings might be nice but are they even necessary.
Actually what I was talking about was a way to set the parameters of any given material right in the panel. Reflection - Refraction - Translucency , etc. And, in a effort to see that in action away I went to Utube and a Vray tutorial. In the picture below the guy has picked ‘Clear Reflective’ but I saw a guy use a default grey and turn it into excellent glass in seconds using those displayed parameters.
I apologize for the quality of the screen capture:
[QUOTE=theoldghost;2491881. In the picture below the guy has picked ‘Clear Reflective’ but I saw a guy use a default grey and turn it into excellent glass in seconds using those displayed parameters.
[/QUOTE]
He must be amazingly fast if it was quicker than hitting new in the materials, and then choosing glass shader!! That’s it for excellent basic glass in cycles.
[QUOTE][Actually what I was talking about was a way to set the parameters of any given material right in the panel. Reflection - Refraction - Translucency , etc./QUOTE]
You totally can do this in Cycles already. If you look in the materials panel there’s a little + sign to the left of each element. Clicking that reveals the parameters for that node, but you don’t have to know that it’s a node at all. The interface is already so flexible that IMO A USER NEVER HAS TO LOOK AT A NODE TREE IN CYCLES. The node paradigm is hugely useful to do complex branchings and some things can’t be done without using nodes. But those things are way beyond beginner requirements.
If we don’t make a modest request that’s easily doable it’s less likely to happen, in my experience anyway. The best way to go is find the solution, then present it.