WOAH!
cough* give me the source code *cough
Man, simcerely EPIC, this looks like MC but better.
Sorry for bad english :b
Downloaded and tried out, and while my system is far from a powerhouse I do well "
Phenom II x4 3.2ghz, 8gb ram HD 7750 2gb ( prob. weakest part of system oh well) .
I ran your example but not only was the menu rather laggy, but moving mouse around had severe lag to point of needing to exit. I was surprised given statements about it running well regardless of size of world, etc.
What should this take to run smoothly ?
Thx
nl
It runs wonderfully on the Oculus. Amazing work Fella!
Thanks!
In the video, it was all dynamic lighting without any GI.
In the video, I used a chunk size of 20x20x70 if I remember correctly.
It was a nice challenge 
In that video, the performance is roughly similar or slightly worse than minecraft on a good GPU. In the current state of the project, the performance is, well, let’s say at least 10 times better
For the packaged game there is a free download link in the video description.
Maybe you had other stuff run in the background? How’s your GPU doing in other UE4 games?
Awesome! My goal with this is actually to make it an awesome VR game. Just having some problems with getting a devkit.
If anyone knows how to talk to valve to get a vive pre, I would love to hear about it 
I sent them a mail two weeks ago, but I don’t even know if they read it.
Can we get a project download because i would like to look at it and learn from. By the way people were telling that it is impossible to voxels using blueprint only but i see that infact it is :D. You are awesome.
One of the more expensive operations in a Minecraft-like engine is lighting propagation calculation when you’re not in direct sunlight. None of the screenshots show underground environments with torches/other light sources. Is this handled? I would imagine Blueprint would have trouble doing the required calculations every time a block is created/destroyed. Moreover the existing lighting in the screenshots with the nice shadows and sun-rays might even get in the way of underground lighting, I suspect. I’d love to know about your plan for underground lighting, if you’ve given it any thought yet.
Lighting also greatly affects mesh chunk creation because once you start propagating lighting information you have to regenerate neighbouring chunks every time a voxel is added/removed to make sure the lighting is updated, this is a very large performance cost (comparatively)
The other performance cost in Minecraft-like engines is that once you start adding caves you all of the sudden have a lot more exposed voxels faces to render, and performance suffers. Minecraft uses a “cave culling” system for that, you can read more about it here:
I also noticed some of your trees are cut off, they’re probably on the edge of chunks and the tree generation doesn’t flow smoothly into neighbouring chunks, this is yet another tricky problem to solve, the the bigger the structures (eg. villages) the harder the problem, at least computationally.
At this point I don’t think it’s fair to make performance comparisons as the hard problems don’t appear to be solved yet. But it is looking very impressive, and I have no doubts it can look better than Minecraft in the end, but I’m not at all sold on doing all the hard problems in Blueprints, you’d likely have to migrate to C++.
I would love to see water with physics in this! :DD do what Mojang only dreamed of!
why do this?
Uhh… he said this isn’t based on Minecraft. It only uses its textures because they were easy to get.
Beautiful work shame you cant profit from it due to copyrights on the textures
If someone can somehow bridge Minecraft’s multiplayer API into a more mature version of this, all shader mods will have problems. Also the guy to port the multiplayer will have problems with Microsoft…
This looks pretty amazing A. Although I’m not interested in duplicating Minecraft’s texture style, I am interested in the Voxel Technique you’re using. I speculate you’re using 1 Group of Instanced Static Mesh Quads per Material to build the visible faces of the ‘Cube’ exterior, removing non-visible interior quads with some fancy algo. Hot or Cold on my guess here?
Yeah, because of the dynamic nature of voxels, one couldn’t use pre-computed data for GI. Well, until now. Nvidia’s VXGI (Voxel Based Global Illumination) is full realtime GI. You probably heard of that already. It does come with its own voxel based ambient collusion as well.
If you have a subscription on GitHub, you could easily grab the Nvidia Gameworks version of UE4, and use VXGI in no time. Yes, this tech kills FPS. But I wonder what a “Minecraft” with GI would look like.
Mind, though. I haven’t tried it with instanced static meshes, but I read once that VXGI does have problems with that.
So I guess there is a resounding yes for marketplace
So [MENTION=24522][/MENTION]; I think it should go on the Marketplace and there seems to be a few people who agree 
While you are working on VR project at the moment, I believe the initial sales would easily pay for your new HTC Vive
@ Shooter is at 1000+ sales now…
I know how you could get one, apply for Unreal Dev See my idea below:
[MENTION=8] [/MENTION]; on Twitter Retweeted [MENTION=14973] [/MENTION]; tweet: /status/713161055158013952
Also, we’ve partnered with Valve to offer 500 Vives via Unreal Dev Grants. Developing for VR? Apply here! Epic MegaGrants | Bewerben Sie sich für eine Finanzierung zur Förderung der UE-Open-Source-Community - Unreal Engine
So how about applying for the and say you want to make a voxel based open source VR solution in Blueprint which will be released in the Epic marketplace for download like Shooter, but it will also become its own game (like Shooter on Steam) or just a learning resource game etc… I bet you’d win
Its still open source even if you have to pay for access to the source. Just like UE4 till GDC15.
If you do win, then I want something for the idea 
hehe
Its not already on the market place? Come on man, this is gold!
Bump for marketplace.
Absolutely amazing!! I’m currently working a on a Climate Change short for 10-13 year olds, with UE4 and lowpoly assets. As a SOLO direc/anim/prod myself I think you just broke the uncanny valley of BP + Procedural = super-smoothness. This is also due me researching procedural during this week, ending up with houdini plugin for my Cinema 4d designs. I’m still a novice when it comes to UE4, but since I got my 1st character for the Climate Change short to run around in UE4 and seeing your project I am definitely in it to win it. If you don’t mind I might sent some questions via mail later on, for some input.
Yeah. I did it in blueprint for that video just to test and show how powerful blueprint is, but a lot changed since I recorded that video and since I want to get every bit of performance for VR, it’s all C++ now
In the video, I always used whole cubes. In the current version I no longer use ISMCs and handle the polygons/vertices manually.
I also thought about trying that, but with a 90 fps target on a GTX 970 I guess I don’t need to spend time on getting VGXI to run
It’s not that much about the money, it’s mostly about getting one soon, and if you preorder one now you have to wait many months. But I can’t release the game before I got a Vive :rolleyes: And it’s also quite hard to work on a VR game without being able to play it in VR.
I saw you switch bp to c++ now, can you release a bp version too? Not need to be as good perform as the c++ one, but it will help a lot on bp learning
Yeah I get that hard to work on VR without having a VR device. I must had misread somewhere in the thread, but I thought you already had a device and had already started, but was wanting a Vive.
What I was alluding to was my second statement which you didn’t quote in the response.:
Since you have the BP Voxel, apply for the Vive , win it, get it sooner than if you pre-ordered 
Simple