Armory Engine Feedback: Armory at Blender Conference!

(Jamie B) #181

Wow, that architectural visualisation in real-time was just amazing.

Really cannot wait to start testing this when its ready, It looks as if its going to be a game changer… if you pardon the pun lol.

You have taken the BGE, and moulded and evolved it into something much more well equiped and suitable for use by modern day developers :slight_smile:
I’m sure that this engine is going to recieve a lot of attention when it lands! And rightly so, this looks like it will be an insanely usefull and fun tool :slight_smile:

Thank you for all the hard work you have put in to making this!! :smiley: :smiley: :smiley:

(cdo1955) #182

will this be multi threaded to take advantage of multiple Cores?

(rogper) #183

I wasn’t so excited about something since the day I’ve found Blender it self.

(Pivetta) #184

Actually you don’t need anything like that, depending on the style you want to use I would in any case make some 3d camera mapping:

If you would like to make it just with 2D pictures then in unity you may just use 2 cameras with different layers activated on each, but I wouldn’t suggest you to get into camera deformations by hand.

With this I would point another question to @Lubos, how will be the camera management in your engine be? Will be there the possibility to blend and overlap many different renders and point each single render path for each camera in the render pipeline editor?


(d4damage) #185

I’m getting very excited about this :slight_smile:

(quickmind) #186

Is there an eta on this project? I’m not trying to be pushy but I like to know so I can count down to the launch date… :)…

(Lsscpp) #187


First release will be hopefuly ready by the end of May, aimed purely for curious minds and testing.

(Sirgeorge) #188

I am curious and excited about the prospect of this coming in and doing for the game engine what Cycles did for the Blender internal renderer. While this engine still has a lot of work that has to be done on it, I was wondering of there was any word on integrating this as a branch into Blender already or if that is even being discussed at this time. Does the development team even know about this at this point in time? Any information related to the topic would be useful. Thank you.

(adriansnetlis) #189

I doubt that BF Development team knows. However, UPBGE fork team is well informed about this. I’m not sure about it, but, iirc, lubos allowed to use several source code elements for the devs uf UPBGE. Or am I wrong?

(BluePrintRandom) #190

yeah, he said anything the bge developers find useful they can use, and the Git is already open,

(elmeunick9) #191

I doubt there is anything of any use, the BGE codebade and Armory codebase are completely different. They aren’t even made in the same language.

(adriansnetlis) #192

OpenGL is always the same(except for versions, but I am sure this is none type of problem). All of the shader code could be easily merged in the BGE, at least I think so.

(SilentRainstar) #193

You’re doing God’s work, my friend.

(Pivetta) #194

Another question floats in my mind

@Lubos As the render path is working on separate modules, could it be possible to run something like voxel path tracing and voxel rendering?


(carmatic) #195

Blender integration is the sole point of this thing, offering a unique workflow
I hope that it will not have the kinds of limitations that the current BGE has, where constraints don’t work, modifiers don’t update, etc etc

this is using a raytracing rendering method like Cycles, right? can the number of samples be adjusted on the fly in order to meet a target frame rate, perhaps based on a rough and preliminary analysis of the scene

(lubos) #196

I love those games! Probably won’t get enough time for out of the box implementation but hopefully the system will be flexible enough to allow this with some effort.

@Jamie B
Thanks a lot for kind words! But, don’t want to ‘overhype’ anything, things will be broken, slow or ugly, especially in the beginning. :slight_smile:

Want multi-threaded, it’s currently under development for the 2D renderer in Kha, I will then start adapting it for the engine. Definitely important but it’s not in as of yet.

There is a possibility to assign a custom render path to each camera, or use the same one and set different contexts for some passes. With that said, I did not actually test this yet, and I recall some stuff was only having one camera in mind. But in the end it should be a thing that works.

Hi! I doubt that BF team knows too. Some members may know(I exchanged a sentence with brecht), and the only concern so far was the naming, which is being taken care of.

@adriansnetlis, BluePrintRandom, elmeunick9
Yah, exchanging source code would likely be terribly difficult. There are not even OpenGL calls, since Kha provides it’s own graphics API to make Vulkan/D3D/Metal and so on possible without rewriting stuff. The hardest thing for me was putting a versatile shader system together and all the render pipeline stuff - my note on not minding the exchange was meant for that, if it would come useful. :slight_smile:

I hope so, very interesting stuff. For voxel path tracing, do you have global illumination in mind or a ‘regular’ render of voxels? The problem would probably be to impose some kind of limitation in Blender scene assembling, to allow only cubes and parse that into specialized format.

When things get broken(and they certainly will), I am here to fix. Raytracing pipeline is very limited, it is a proof of concept right now. I have some more ideas to push it further, will be sure to post that. Adjusting number of samples on the fly is nice idea, writing that down, should be fairly easy.

A month already passed, I still did not manage to get a preview version out(yah) - I tried to give reasons and complete overview of the progress here:

There is some more stuff mentioned but the main new things are Hosek/Wilkie sky model support, which you can drop into world nodes the same way as in Cycles

And a transparency support, using a weighted blended order independent technique. This finally enables to simulate glass nodes and similar!

Hope to pull off another run similar to this month, and get things out. As always, everything is lots more work than anticipated. :slight_smile:


(quickmind) #197

…YA! YA YA YA…sky nodes… Transparency… I love it. :smiley:

(adriansnetlis) #198

Cool. By the way - I can setup a car paint material for demonstrations if you need one. Just contact me before you wanna make the demo;)

I am looking forward to see your Hosek-Wilkie implementation:)

(Sirgeorge) #199

Hmm, with a ray tracing system in place I could set the samples to 1 and get a very nice artistic style. I have done real-time tests with ray tracing in the Cycles Engine on both 24/30/60 fps and it managed to work out really well even though the engine wasn’t intended for it (I would suggest putting a setting in that allows you to adjust the samples in your engine). I’m already getting many ideas on the subject. Don’t worry about things getting overhyped with me. We are all bleeding edge consumers here so at least I can speak for myself when I say you don’t have to worry about “blowing it”. It’s awesome to see this sucker getting made. Also, getting ray tracing to work in a real-time renderer would allow you to, if properly done (which would probably take a while), potentially get true real-time photorealism. This is one of the holy grails of video game development and I’ve been waiting for it.

 I would say the platonic form of this engine would be the power of Unreal Engine 4 in a blender package with the intuitiveness of Blender and the other benefits that come with being in the Blender package including the community as well as the holy grail one button port and the holy grail ray tracing. For me though I'll be happy with everything just before the holy grail talk starts happening. Again, though, don't worry about me being disappointed with whatever the results are. This is an awesome project and it will do well regardless of the status of it vs it's platonic form.

Finally, what was it about the naming of the engine? What was the concern?

(HeadClot) #200

Hey Lubos,
Got a question about Armory Engine.

Will there be support for large worlds with the Armory Engine?

If so how will you go about implementing it?