Strange thread!
brecht has gone to solid angle, so congrats to him…
lots of people want to use arnold and blender together… fine so far…
but the implied cause and effect… that somehow solid angle “owe” the blender community something because they hired someone who coded for blender… I’m just gobsmacked at the unrelated nature of these two things…
as for the debate about cycles vs arnold and blender’s apparent inability to produce scenes that would get a benefit…
i just find that hogwash too… maya, max etc all suffer the same problem… it’s why products like katana exist… because most generalist DCC software can’t cope with the extreme complexity of summer action blockbuster scenes…
but in all seriousness if people want a plug in then i hope they can band together and get one developed… expecting Brect or solid angle to be involved seems a step to far though…
If I were solid angle I’d think it best to follow Chaos groups lead with Vray:
ie wait until a developer out there makes an integration (the true spirit of open source) and once it’s proven to be good and usable and has a following amongst the blender community recognise that it’s good for business to offer financial support to the guy so that he can maintain and bug fix and extend and they can boast good blender support out of the deal…
I really see no reason to even desire an Arnold plug-in for Blender. Blender is not being used in situations where the benefits of Arnold (extreme distributed optimization) are being drawn upon. I think that it has potential, but Blender will not supplant Maya, Houdini, or any other plug-in rich application for many years.
We have a new release of LuxRender about a year away, and very few of us can currently take full advantage of Cycles with our processing power. I am very happy with our current renderers and would prefer to see any and all work concentrate on them.
And while thinking about professional uses, I don’t think Blender needs Arnold to be attractive. It’s cheap, resource efficient, and can be used in any situation where a single desktop or small render farm will suffice. I’m thinking small additions to scenes, compositors looking to add something extra to shot, small splashes of water, particles, and on and on. Hell, Cycles can produce some great-looking metals. I could imagine adding extra robots or machines in the background on a single desktop with Blender. None of those scenarios requires Arnold, and that’s great because it would just force a studio to buy yet another license.
People who think Arnold is only useful when you are making 500-million polygon scenes have never worked with it. When I’m working on personal projects in Cycles, I regularly have moments where I think “this wouldn’t be a problem if I had Arnold…”. Mesh-grate floor or chain-link fence is a few hundred-thousand polys? Who cares, make a standin for one section and instance it. Sculpt is 4 million polygons? Meh, just bake a vector displacement map, adjust subdiv level as high as you can bear, and let auto-bump sort out the rest. I need 4k textures for all of these buildings, even though only 1 will be that close up in any given shot? Meh, just do it, texture cache will deal with it. Don’t have time to trace all those rays for 3d motion blur? Yes you do, you’re using Arnold, it traces rays at 8 gazzillion miles an hour. Alpha-mapped leaves too slow? That’s ok, just make the tree with mesh leaves, who cares if its 1 million polys. It’s a standin anyway, and it’s not like Arnold cares.
You never have to think “Ohhh…is awesome my idea too complicated to render…?” The answer is no, because Arnold doesn’t give a *$&%. It’s just always stupid fast. The idea that you need a team of people working on a summer tentpole to hit that kind of scene complexity isn’t true. You can do it with 1 person’s personal project easily. Sculpts, architecture shots, illustrations, short animations. It’s just an absolute joy to work with.
J: If you haven’t tried, I highly suggest leveraging MakeTx next time you’re working with large texture sets like that. WAYYY faster than the default tex cache
I didn’t realize that Arnold was only $1,300 per license. That’s a pretty good price for such a famous piece of kit. That’s well within the reach of hobbyists and smaller houses. I was under the impression that Arnold was in the neighborhood of 3-4 thousand.
You know what, I have changed my mind. Having a Blender-Arnold interface would be a fine idea. I still don’t think that it’s a pressing issue, though.
That discovery got made about a week into switching over. We leave nothing uncached now. Something like that is on the todo for Cycles, right?
Keep in mind, they (atm) have a minimum purchase of 5, and there’s also the ($300/seat) support contract, which includes updates and their (awesome) support team. So you’re looking at a minimum buy-in of $8k. There’s also no discount render-only licenses, which can creep up the pricing a bit compared to some other engines. Still well worth it though, IMO. At least if you can put 5 licenses to good use.
They do have really simple pricing though. There’s one type of license, and they don’t charge for exporters. So it’s just 1 key + support ($1600 total) for each computer that needs to run Arnold somehow, and that’s all there is to it.
Ahhh. Ok. That may be where my initial impression came from. I again reverse my opinion back to my original view: we have other, much more important things on which to concentrate.
The support is fine. I think an annual support fee of $300 is very reasonable, but that minimum five-license contract is a real killer.
This would imply that Cycles is in a safe spot right now, because this means Arnold would not be of much use for independent artists which comprise the majority of people in this forum (one of the types that Cycles caters to).
An 8K buy-in would mean paying the equivalent of every other competing render engine combined (that even includes the new Clarisse software and Keyshot), so I guess they would be expecting that you are part of a team then if you are to use the engine.
Don’t get me wrong, it’s an amazing piece of software where people have major trouble finding something it can’t do. It’s just that the required team license purchase is a massive catch when you find excitement about it being just 1K.
Individual licenses are coming. You just have to keep in mind that they’re still a little company without a dedicated service-only team. The number of support requests that would come in with individual users would be incredibly overwhelming, and it’s not like they’re hurting for clients as is.
I’m not blaming them. Not at all. I don’t get the impression that their choices are being driven by arrogance or greed (as opposed to Autodesk). I’m just saying, without anger or annoyance, that with their current license structure, I see no future vis-a-vis Blender. I wish them all the best and hope that they reach a point where they can support individual licenses, lower prices, and maybe even official Blender support.
