David

I finally put Cubiquity 2 (my experimental micro-voxel engine) on GitHub! It consists of a small number of zero-dependency C++ files released into the public domain. It's an early version and not yet ready for use, but I hope it is of interest: github.com/DavidWilliams8…

@procfxgen @ephtracy There is indeed a limit of 256^3 per model, but you can place multiple models in a scene (each with their own transform). I use this to create a voxel chunking system. With this in place, I think the MagicaVoxel size limit is then 2000x2000x1000.

@DavidW_81 @ephtracy Thanks. MagicaVoxel is not limited to 256x256x256 voxels ?

I've just released my library for writing high-resolution #magicavoxel files. You can get it here: github.com/DavidWilliams8… #voxel #voxelart @ephtracy

@ephtracy Thanks to @quaternius for the assets, which I voxelized using Cubiquity.

I've been doing more work on a .vox writer for exporting voxelised scenes to #MagicaVoxel or Avoyd (shown). This warehouse test scene has also been expanded since I showed it last. More details here: reddit.com/r/VoxelGameDev… #voxel #voxelart

@quaternius Each object in MagicaVoxel has a size limit of 256 voxels, but next I'll be combining multiple objects to export larger scenes.

Working on a MagicaVoxel exporter for Cubiquity. Assets by @Quaternius, voxelised using #Cubiquity, rendered in #MagicaVoxel.

@miketuritzin Thanks, I do recall skim-reading that paper at some point... I might go back and have another look.

@DavidW_81 Hmm, glancing at the paper again I think it may have been the [KRB∗16] reference I was thinking of, but not 100% sure.

Enjoyed reading this paper, which is one of the few I've found that covers the topic of dynamic updates to sparse voxel octrees. (In this case, specifically sparse voxel DAGs, which are essentially deduped octrees.) github.com/Phyronnaz/Hash… 1/

@miketuritzin Hi Mike, which other paper does the idea come from? Section 3.1 only references "Kämpe et al. [KSA13]" (the original sparse voxel DAG paper), which doesn't cover dynamic updates. I'd assumed this section was a novel contribution but would be interested if it's covered elsewhere.

The basic idea for the dynamic updates is to create a new tree root that reuses all unchanged subtrees from the previous tree. This means nodes are never modified, only created. (As the paper notes, this idea comes from another paper.) 2/

The pen in this picture helps show the scale of the voxels (they are about 4mm across). That means about 10 billion voxels, compressed down to only 5Mb!

The building is mostly finished so I'm now starting to populate it with models from @thebasemesh There's about 300 meshes in the scene, each one gets voxelised to a separate volume which then get combined.

More work on the warehouse scene which I'm using as a test for the voxeliser in the Cubiquity voxel engine. #GameDev #Voxels #VoxelArt

More work on voxelisation in Cubiquity. I'm making this test scene in Blender and voxelising at 4mm per voxel. The resulting Sparse Voxel DAG takes only 500kb on disk! More information in the Reddit post here: reddit.com/r/VoxelGameDev… #gamedev #voxels

The model is from threedscans.com/uncategorized/…. I'm hoping to make a public database of voxel scenes in the future.

The algorithm is robust against cracks, etc, and gives a proper solid (filled) voxelisation. It can also handle multiple objects per scene and multiple materials (not shown). The current version is slow but I think a GPU implementation should be significantly faster.

I don't post much on Twitter these days but I'm still working on Cubiquity. I've been improving the mesh voxeliser recently. This test model is 2.5m triangles voxelised to 22 billion voxels, and takes only 20Mb of memory as an SVDAG (6Mb on disk).

@HenkPoley @GossiTheDog @taviso Wow, I've been unmasked... what a suprise!

@GossiTheDog @taviso Basically @DavidW_81 😉

Hah, reviewing commits I notice that Jia Tan cannot spell the word "guarantee" correctly -- he misspelled it multiple times in commit messages. How can I grep every repo in github for the same spelling error? 😂

@PolaronEngine Hi, I'm happy to answer questions over Twitter but unfortunately that's all. I have too many other commitments so I'm not available for collaboration, etc.

@DavidW_81 Fancy a chat?

@WebbChilton I released the source code but it's still not ready for use. I don't have much time for development so progress it slow! github.com/DavidWilliams8…

@DavidW_81 Any chance this is available for people to use yet? :) pleeeeease?

Hey everyone, I've recently been very quiet regarding the next version of the #Cubiquity voxel engine. Unfortunately I don't get as much spare time as I used to, but it *is* moving forward...

@bjornornorn Hi Björn, thanks for this! I notice the conversion functions in this post are MIT licensed while the Oklab ones were also public domain. I assume this was a deliberate decision (which is fine, of course)? I only ask because my own work is public domain, so I can't embed MIT code.

Looking for alternatives to HSV and HSL that better match color perception? I've made two called Okhsv and Okhsl. Try the interactive color picker demo here: bottosson.github.io/misc/colorpick… Or read about all the details: bottosson.github.io/posts/colorpic…

@anders_breakin @xDirtyPunkx @BrianKaris @mmalex @kmett Fantastic, thank you, it looks like I've been reinventing the wheel again! Flipcode is a gold mine.

i think this by @BrianKaris is my favourite talk of all time. must watch. his answer in q&a about average triangle size being higher than he thought (4 pixel edge ish) reminded me of a long forgotten sw rasteriser someone at bullfrog wrote in the 386 era! (around dungeon keeper..