Voxel Quest January 2016 Update | Hacker News Reader

Voxel Quest January 2016 Update

voxelquest.com

303 pointsshawndumas128 comments

Wow, I was high sceptical that you would be able to scale this up past the short view distance. But not only have you managed to get a good view distance, but you also added in a nice animation system. Congrats to you! This is coming from someone who has spent years designing their own voxel platform. Please get this out to users to let them start building things, you will be amazed at what they can create.

Thanks! Don't tell anyone, but I was legitimately skeptical too. :)

Many many many game developers get sucked into the infinite vortex of trying to "find the fun". A computer is a very large blank canvas and the set of all possible games is infinite. It can be overwhelming trying to decide what game to make.

Often, developers wander off to "make an engine" partially as an excuse to put off making those scary decisions. But an engine isn't a game and if you want to ship something people can play (and not just make games themselves with) you do have to make that choice.

You've made an insane amount of progress, but I wonder if you're struggling with this too. If I can make two suggestions that might help:

1. Realize a negative choice has positive value.

Every idea or feature sounds cool, so the default answer to "Should I add ___?" or "Can the player do ___?" or "Can the engine support ___?" is going to be "yes" because it all sounds fun. But the end result is something that takes infinite time to produce. Your time is finite.

It takes courage, but I find I make better progress when I deliberately allow myself to choose not to do something. Remember, any feature you omit pays you back in the most important currency: time. By saying, for example, "No multiplayer", you're saying "Yes" to a lot of other features you will now have time to implement.

2. Make the game that only your engine can do.

You're in an unusual position in that you already have a very interesting, unique engine. Most developers out there just have half-baked clones of existing engines.

When you're trying to figure out what kind of game to build on top of that engine, focus relentlessly on the kind of gameplay your engine, and only your engine can support. For me, that's rich, detailed, destructible terrain. If it doesn't take advantage of voxels, cut it.

Looking at the videos over the past year, I see lots and lots of basic character animation. Every game engine can do that. It's not interesting to see yours can too, and it takes (took) a ton of your time to do it. Don't spend time catching up to other engines. Spend time building on top of what you already have.

I don't care if a character in the game is a simple billboard—I actually think billboard characters would look nice with the pixely voxel style—I'm coming to your engine for the terrain. Everything else is secondary.

Good luck! It's been inspiring watching you make progress on this.

Thanks Bob! Totally agree with you - it really is a battle of fighting against my "engineering instinct" which just wants to build stuff all day, no matter how useless it is. I have deliberated over this internally with myself and others for a long time. In fact, half way into the character animation I was kicking myself for even beginning working on it. I tend to justify things (whether rational or not) as a content problem ("Oh I can just build this system and it will make making this type of content so much easier!"). I now really need to hone down on gameplay (and stop just telling myself to do so). :)

> I tend to justify things (whether rational or not) as a content problem ("Oh I can just build this system and it will make making this type of content so much easier!").

That can also be a trap (one that I myself am particularly prone to given how much I love roguelikes and tool-building). When it comes to games, I'm just not wired right to enjoy hand-authoring content. So when I do game-like stuff, I tend to lean towards procedural generation because it lets me avoid that unpleasant task.

But, I've learned, painfully, time and time again, that that tends to be way harder than hand-authoring. It kind of makes sense. If hand-drawing a fun level is difficult for me to do, why should it be easier for me to write a program than can author a million of them? In some sense, the latter must be strictly harder since it is a superset of hand-authoring a few one-off things.

Tool-building does matter, of course. But sometimes you just gotta make some stuff.

Another option is to push it onto the player: make the gameplay itself creative. That can work (SimCity), but it can also fizzle out (Spore).

The problem there is similar, I think. With hand-authored content, you have to come up with a narrative path or maybe a tree through the possible game space that's fun to play. With player-authored content, the entire play space has to be fun. Otherwise they'll be able to make anything, but most of the stuff they make won't be fun to make.

"One more system will solve my problem" is the siren's lure isn't it. But I have learned "It's turtles all the way down". Each problem solved reveals more opportunities. So I just try to enjoy the process of learning, if I am lucky it will end up being fun too. If not I enjoyed the journey.

I never knew the source of that expression until just now: https://en.wikipedia.org/wiki/Turtles_all_the_way_down

I agree, especially if you are just making one game, there really is no point to procedural generation - other tools will make all that much easier (Maya, Blender, etc). I'm trying to think outside the context of just making one game, but rather making something that people can broadly use that it is simpler than Maya and Blender, and even disguised as a game to lure people to it. This could be a foolish move - the odds are already against me making a game, let alone a platform/engine. But I feel like somebody has to shake things up a bit, we are getting too comfortable with our AAA Unreal games. :)

I also struggle with adding extra features to a game I'm working on. I've added "You're not going to need it" to the top of my to-do list as a way to force myself to think everytime I add something on there. ;) https://en.wikipedia.org/wiki/You_aren%27t_gonna_need_it

Have you ever heard of the game Magic Carpet? See http://www.gog.com/game/magic_carpet . It had insanely good destructible terrain, base building and exploration. I think something like that might work extremely well with your engine and also there has never been another game quite like it that combined exploration and resource gathering with the almost RTS like base elements (and of course the fun of creating volcanoes and dropping asteroids on your opponent's base).

Yep I know about Magic Carpet - its one of those games (along with Outcast, Commanche, etc) that drew me to alternative engines over polygon-based ones. Initial ideas for a prototype did revolve around base building, mini RTS / tower defense style stuff.

Magic carpet, what great memories! A modern take on the theme would be great in VR.

As someone who has commented on your engine threads in the past(and pointed out this challenge) it's good to see that you're getting towards the content stage of dev.

Here is my tip this time: try to find a lot of anchors outside of the game engine. You can do design incrementally by defining what the "rules", "logic", and "lore" of the world are, in order to clarify it. For example, use a mind map tool like a writer to plan scenario ideas. Pick a genre of music and a visual design language and incorporate them into the game's identity. Consider the limitations of the camera and control scheme and allow those elements to shape the scenarios. Borrow plots and characters from stories you like. Compose an essay response to something you want to critique, and then figure out ways to turn that essay into game form. If you don't have ideas for game mechanics, you can develop the story more instead and you'll eventually find some story elements that convert into mechanics.

The more of these things you can piece together in a coherent way, the more it will start to look like a complete design where "the work is cut out for you", instead of being a mass of bad content to crank out. There are a lot of ways to go about defining the design, but you will have to find some anchors if you want good content without a lot of meandering and second-guessing. The meandering should happen in notes and sketches, well before you try to write final code or assets.

Oh, and all of this can be done with an eye towards marketing. You can opt to choose and highlight elements specifically to capture an audience. The hard part tends to be finding concepts that you're happy enough with to want to think about for months on end.

I agree - ironically I am not particularly interested in a fantasy setting and many other people are tired of fantasy as well (although I could still probably get away with such a setting). Fortunately if I do decide to do some other design there is not really much work towards content yet - all of the stuff in there is broadly applicable. I have always been interested in a Fallout-style universe. As much as those things are important (and they are) above all I think I really need to laser in concept-wise as well; that may just mean mean going with something tried and true in the early stage.

You could always go the route of many 80's and 90's RPG's and mash up some fantasy and science fiction themes - time travel, alien worlds, other dimensions, etc. Nearly every popular franchise of that era pulled in those tropes and they give you tons of breathing room to pull in whatever elements you want.

Yep, that era of RPGs is my primary inspiration. My favorite series were Ultima, Fallout (1/2), various DnD games like Eye of the Beholder, Dungeon Master, and quite a few more.

Have you tried asking for help with how to invent/build contents & fun-factor on the TIGsource forums? There seems to be a section dedicated to "the non-programming aspects of game-making":

https://forums.tigsource.com/index.php?board=5.0

for example a "sticky" thread titled "Game design resources" from one of the subforums:

https://forums.tigsource.com/index.php?topic=31350.0

Also, I'd think people like to help with generating/brainstorming some ideas when provoked, like you're doing here with throwing around some thoughts. I can barely resist adding my $0.02 of random suggestions already :)

I browse TIG a lot but had not thought to do that - not a bad idea :)

I think this is great advice. Is it accurate to summarize it as "artificial constraints can create unpredictable strengths"? (In this case, the strength might be a strong sense of identity)

There is probably a name for this concept, but I don't know what it is.

In my opinion, that's what has been missing with Starbound, which has been one of the punchingbags for crowdfunding/pre-order games. The engine looks great and it has a lot of unique content - but its just an engine right now, with no real content for people to play through while they task and tinker with changes (they are on, what, the 4th full combat overhaul?). In contrast, during development of Terraria or Minecraft you could get enjoy playing it.

The major difference is that Starbound made heaps of money (not to mention that Terraria sold 7 million + units!) I'm still looking for change under the couch. :)

Oh come on, that's just... not right, given the amount of work and time you put into it. I don't see any "Donate" button on your page, so I just bought the pre-order. Don't lose it all under the couch ;).

Thanks for your support! :)

Heaps of money do not make you happy! They just change what stresses you. Don't discount just giving users the ability to build, I spent 10 times more time building in MC then I did playing "the game".

NovaLogic's Delta Force (1998) and Delta Force 2 (1999) had a "voxel game engine [that allowed for a] nearly unlimited draw distance" [1]. Though the other objects in the game were rendered with polygons, has a far view distance been difficult with modern voxel engines?

[1]: https://en.wikipedia.org/wiki/Delta_Force_2#Game_engine

It's not a true voxel engine, just a 2D heightfield renderer. Note the lack of holes or undercut walls.

Delta Force and Commanche were among games that drew me to alternative engines. Doing a CPU-based renderer gives you lots of new tricks and abilities, things you can't do with simplified massively parallel architecture like GPUs. Many games have a far view distance now (see Uncharted or Just Cause) - but doing it in a dynamic world that is procedurally generated in realtime with solid geometry is a whole other bag of cats. :)

Hear ye! I lost interest in games right around time time 3D graphics cards started coming out. I figured, that scene is pats its prime now; it's going to just be generations of the same thing over again, but rendered slightly better. Programmers no longer in control over the details, just pushing some triangle meshes into some hardware blackbox.

All software Doom, Quake, Wolf 3D, Rise of The Triad and such: that was still cool stuff.

The game programmer should control every pixel going into the frame buffer.

well you COULD use OpenCL to "control every pixel".

Here's an example of path tracing in "realtime" (ok, it's not quite there yet) on the GPU.

https://www.youtube.com/watch?v=BpT6MkCeP7Y

check out this blog too: http://raytracey.blogspot.com.au/

That used to be true, but right now you not only can, but have to control every pixel. The fixed pipeline is dead, now you have to write your own shaders and those can do literally anything. Or use CUDA/OpenCL for even more power.

Shaders still have plenty of limitations relative to a CPU (you can do virtually anything with them, but some tasks are too expensive to do efficiently). In particular, taking advantage of temporal and spacial coherence is much more tricky on a GPU - on a CPU the early 3D renderers could effectively use this stuff with scanline and fill tricks.

I love the Voxel Quest posts because they're demonstrating some seriously cool tech. However, they feel a lot like the Wolfire development blogs [0] - great tech demos without a cohesive vision for a game. From an engineering point of view, the end goal doesn't matter, but I would also love a 'bigger picture' roadmap of where it is and where it's going. Why would I use this engine? What do I 'get' over the competition in the current state of the engine? What should I 'get' by the end?

[0] http://blog.wolfire.com/

This is a really good question actually. This engine isn't really meant to compete with major engines, but moreso satisfy a little niche (and maybe that could be a big niche depending on how it gets used). I think the best way to explain it is "Shadertoy for large environments" - being able to quickly dynamically create stuff with pure code. To be honest a lot of this is still unmapped territory and I'm not even one hundred percent certain that these things are all that useful. My main goal is to empower individuals by creating a relatively inflexible engine - that sounds weird but if you have a totally flexible content pipeline you are giving users enough rope to hang themselves. I guess the proof of usefulness in limitation can be found in things like Minecraft - it appeals to a broad audience of not just people who want a game, but people who like to tinker. Minecraft, by many people's accounts, is not really a fun game, but a great sandbox. I'm not saying I can get away with making an unfun game, but I think there is some sort of value in empowering users in the guise of a game. I don't know if this makes sense, let me know if I can explain it better :)

I have a notion that it's possible to have a "somewhat directed sandbox" game. I'm working on a combat/survival open world game with a procedurally generated tech tree to go along with the procedurally generated world. There is a "difficulty direction" (down) and the farther you go that way, the environment becomes exponentially more difficult, while the items/powerups become exponentially more powerful. (With the environment gaining power faster than the player, of course.)

This gives just enough structure for players to escape the Second Life trap. (What happens to people who have no needs/problems in a virtual world? They go right up the Maslow Need Hierarchy -- to Sex.)

Hopefully, this will provide the right kind of structure to let players game-balance the game themselves. (And give them enough scope to invent their way into and out of corners.)

Interesting. Welp, looks like I'm going to have to resort to adding sex to the game. :) But seriously, the idea of a "direction" for difficulty had only vaguely occurred in my mind (for whatever reason, I am drawn to games where you ascend or descend to progress, most notably Spelunky).

Welp, looks like I'm going to have to resort to adding sex to the game. :)

Hey, it works in Minecraft. (But in Minecraft, you have to earn it.)

sex in Minecraft

I didn't expect I'll be googling that phrase ever in my life. :o.

I was referring simply to having children, but apparently, people have taken things a bit farther. (I hope in jest.)

I'd be interested in seeing a voxel game involving defensive architecture (preparing for disasters). When I started playing Minecraft, I was interested in making buildings, lighted streets, etc that would make it easier to survive at night.

It turned out to be somewhat pointless because once you know how, you can easily skip the night phase altogether. A game where you can't do that might be interesting.

Also: preparing for floods. What happens when the river overflows or the storm surge hits?

Yep, a survival type game would be well suited to the engine. It is definitely part of the long term plans. In the long run I would like to have a world with some simple towns and people can move about and raid towns and so forth and quests kind of emerge from those circumstances. Maybe a bit like Mount and Blade, but in a more seamless world.

There are lots of fighting games and I don't think that's what makes voxel games unique. The cool thing about voxels is the environment. A player-vs-environment game puts the environment front and center, instead of making it the backdrop.

Another idea: build a sand castle and see how long you can hold out against the rising tide. No bad guys involved, just waves. Almost a literal sandbox game.

Skyrim and Fallout 3 and 4 are basically somewhat directed sandbox games. There is a main questline you can follow, but most players just immediately take off to go exploring that cool thing in the distance. Then die because the mobs are too high level :) . And that gives you incentive to level up, which frames your interaction with the world.

> My main goal is to empower individuals by creating a relatively inflexible engine

I'm a big fan of fostering creativity though restrictions, so this makes sense to me. I would be interested in any notes on how you try to realize that goal in the engine. It would be interesting to see the difference between limitations imposed by performance v.s. limitations imposed in the name of a focused creation experience.

In general, any time you're not "simply" (not that the problems are simple) creating a system that exists elsewhere (physics, liquid, etc), it would be interesting to hear your thoughts on where you choose to deviate from the 'norm' and why.

The primary idea is that it should be easy to create, modify, and revert something in a pretty standard process. In a 3D modeling program, a user usually has to go in and create some outlines and select vertices and edges and extrude them and further push and pull things from there and deal with polygon issues and so forth. They have to do all this while navigating an interface filled with one million buttons and menu options. I am aiming for something a bit more like Fallout 4's face editor - its still quite limited, but any user can understand it and produce a diverse variety of content (of course, I am extending that to all shapes, not just faces).

For other things, if I introduce some system I'd like it to be unique and have purpose. Many games do GPU-based particle physics for water, and the ability to have this produce any significant interactions isn't that great. My system is CPU-driven, and contains things that are rarely done in games like simulating pressure and a constant fluid mass, which makes for more interesting interactions (you could pour water down the end of one pipe to raise it at the other end, as just one example). Usually people find ways to modify the system that I would not think of (someone made a roller coaster in Minecraft using water canals IIRC).

Skyrim was amazing, open world with lots of quests. Minecraft kinda sucked a bit, huge world but I spent more time on the internet trying to figure out what was possible rather than play the game. The Lego games have a nice open world with quests and stuff.

But a game needs to be fun and easy to understand. I hope VQ adds lots more fun things. Bombs. Bombs would be fun ;)

There are bombs: https://twitter.com/gavanw/status/627118562629386240 :)

I spent more time on the internet trying to figure out what was possible rather than play the game

I think it's possible to make spending time on the internet trying to figure out what's possible an integral part of playing the game.

Overgrowth has been painful to follow.

Years ago, I loved to read their blog. Lots of interesting technical articles. But development has pretty much stalled for years.

IIRC, these guys are also behind Humble Indie Bundle, so money should not be a problem for them but it would be nice to finally see them ship the damn game.

I tried out the alpha or beta and I thought it was quite good for the type of game they are trying to make. I assure you they don't even need the Humble Bundle money - David has been selling copies for a long time and sold over 100k copies on Steam alone (that is around $3m USD), and I'm pretty sure they sold the majority of their copies off of Steam. His other mini game, Receiver, sold +600k copies on Steam @$5 each, which is another $3m (of course, you have to factor in Steam's ~30 percent cut and taxes, but still....)

They definitely do some interesting things but last time I tried it, it played a lot more like a prototype than a game in development for nearly a decade.

Is gavanwoolery made of magic and unicorns? This is amazing and his progress is nothing short of astounding.

If he's reading this: how do you fund work like this? You said you're raising a kid, I've got two! Is this a couple-hours-a-night kind of thing?

> Is this a couple-hours-a-night kind of thing?

Phew, I don't even have kids, and I can't maintain a couple-hours-a-night hobby between maintaining relations, dinner, and a fairly late bedtime.

Thank you! I have two small-time investors currently, although I am still living relatively minimally (I also put in a good deal of money from personal savings, and ~$35k came from Kickstarter). Its pretty much a every spare second of your life type thing, especially with a baby. :)

It looks like a lot of fun. I've been following the development for some time and really enjoy your posts and updates (especially the ray-cast renderer, so cool).

You're amazing.

Thank you, it means a lot :)

Hey, can I say I love your work! I'm also trying to get a platform off the ground to help creatives like you and me build more stuff. It's sort of a patreon for makers/inventors/developers, but instead of simple crowdfunding, I wanted to build-in all the tools I needed to grow my project and community simultaneously. I'd love if you and anyone else reading this would check it out and give me some feedback: Http://baqqer.com

I'd really love to help more people fund and follow their passions full-time.

Cool, I love tools like this. Just curious, how do you plan to differentiate yourself from Patreon and Kickstarter? (it is ok if you are just doing their stuff better, there is room for competition)

I discovered you on HN several years ago and have been following your progress ever since. You are an inspiration to do-it-yourselfers everywhere. Rooting for you man!

Thanks!

From a technical point of view it's amazing.

Artistically, the voxels are too big to look impressive to anyone used to polygon graphics, and too small to give off the "pixel-art" aesthetic of Minecraft or Terarria.

I want to play the game evoked by the banner at the top of the page, but not the one I see in the video.

I agree. The pixels are small enough that it falls into this sort of uncanny valley of the terrain almost looking realistic, but not really.

It would be better almost to go further in either direction: either make the pixels smaller and therefore the terrain and characters look much more realistic (probably technical challenges in doing this), or make them larger so it gives off the "pixel-art" look.

Yep - many things that need to be tuned visually but I don't have the time budget for it yet :) I do agree though - it actually does look better sometimes at either much lower or much higher resolution.

It can be more like the top banner (higher res, and isometric) if set to isometric mode. It is resolution constrained for realtime / perspective camera due to all the heavy computing done in in ray marching.

Want to register some support for what I see in the video - the banner looks cool, but I'm a lot more excited about the more detailed procedurally generated terrain in the video.

There are portions that look like that[1], he's just been showing off the terrain in the last few updates. Check his videos page[2], about 2/3's the way down.

1: https://youtu.be/gW04jTU5FQw

2: http://www.voxelquest.com/videos.html

I don't have anything to add to the discussion, I am just really inspired by your work, you can see the trajectory of this project with each update. The environments look simply beautiful with the huge view distance.

The big unproven thing is building something that is fun. I look at Spore as this great unbelievable game engine that accomplished exactly what it set out to do. But they forgot to make the game fun, don't be Spore.

I agree. Funny enough I came into making this thinking the "fun" aspect of it would be the easy part. Just assemble all the pieces and suddenly it is fun, right? Concept is really key - doesn't matter how many features you have, if your core concept is no good you have no game.

I will go against the crowd and suggest that you just keep doing what you're doing. Forget about "making it fun". Even in these rough demos, the various parameters and behaviors you land on clearly shows a natural instinct for it that you can trust. Forget about "making an actual game", what you're doing is magic and the more you do it, the better. You've got a gem here. It's already insanely valuable. Keep polishing it, and the world is yours. Keep this up, as long as you can. The "last mile" is going to be easy for you.

Thanks! It goes against "common sense" but you might be right. Sometimes the formula for success is that there is no formula. To quote myself from a prior tweet, "Success is what happens when hard work meets a stupid idea. Every idea looks smart in hindsight, but the good ones look stupid in foresight."

Are the game engine and game coupled together?

Meaning if I purchase the Alpha game key, will that give me access to the game engine during Alpha to start looking at that side of things?

Yes, in fact the full source is going up on github, likely right around or during release of the first alpha build.

Oh, wow! How did I miss this? If you are pushing the source on github (regardless of license, I want to avoid that rathole) just so others can learn from what you did, then where do I go to fund you?

Source will be public but I have to attach a license at least for the first couple years. If you are interested in supporting me, you can either preorder a copy or become a patron:

http://www.voxelquest.com/patrons.html :)

Well Gavin, enjoy updating your active patrons list. This bit of love from Hacker News should do you good, in a variety of ways!

Watching the demo today woke up, in me, some long forgotten excitement. My son and I enjoy the hell out of Minecraft. It's a great game and it arrived in its space at just the right time.

What you're doing, though, is going to push things in a new, even more exciting direction, in ways that you can't even imagine yet.

Thanks! I hope so :)

For the uninitiated why is this exciting? The graphics look really washed out and jittery. Thanks!

Good question! The entire scene you are seeing is being procedurally generated in realtime over 60 times per second (nothing is cached or pregenerated). It does not necessarily compete with polygons in terms of visuals, but it is great for interactivity and procedural generation and so forth. Everything you see is solid (vs being "paper thin" as most polygon games are). The entire application is only a few megabytes large, so this make its easy for users to share their creations with each other (which could be just a few kilobytes in size). It is able to represent an area of several square kilometers with virtually no memory footprint (something of this detail would require several gigabytes of memory to store all the vertices for polygon data, not to mention the texture data). This entire scene can change instantly based on user parameters. I don't do a great job of demoing all of this but I should in the future. :)

That sounds really interesting, so compared to traditional polygon based graphics voxel is absolutely tiny?

Depends on how you are representing stuff. The name is a slight misnomer that hails to the early days of this project when everything was strictly voxel based. Now the only thing voxels are used for is to determine the basic terrain addition and destruction. Voxels tend to be more storage efficient because their location is inherent to their index and vice versa.

> Now the only thing voxels are used for is to determine the basic terrain addition and destruction.

Do you find that to be pretty efficient compared to other methods of terrain deformation (ROAM,etc)? Pretty curious, since I did my big undergrad research paper on deformable terrain in videogames before the huge voxel craze.

Depends on what you classify as efficient: mem usage, gpu workload, etc. Voxels are easy to work with but often represent more data than your really need.

I was mainly asking in general terms of "gamedev" efficient: Whatever deforms and renders at a playable rate and looks just as good as the alternatives.

Voxels would probably be the simplest thing to program with (they were easier to use in my first iteration than my current (3rd) iteration of the engine)

Voxels are discrete, so not absolutely tiny. Voxel = vox (volume) + el (element), similar to pixel = pix (picture) + el. Do you know Minecraft? All those cubes in its world are voxels.

The amazing thing here is how tiny this guy's voxels are while still keeping the game running at a decent framerate.

Compared to traditional polygon graphics: voxels are discrete elements in space (sort of like "giant atoms") while polygons are the surface of objects (cuts of planes in 3d space). Being merely polygons they have no depth (as GP said, they're paper-thin) while voxels are inherently volumes in your space. You could compare it to Lego vs. papercraft.

Why are the voxels (I guess the jagged edges come from the voxels?) so big?

Because every voxel needs processing, and you've only got 16.7ms to process every voxel every frame.

Is it possible to scale them like pixels? Like a "voxel resolution" that can be adjusted if you got a better PC etc.

From viewing earlier demos I believe it can be, possibly in real time. That said, reducing a voxel by half in each dimension yields eight times as many voxels, reduced by however efficient his algorithm for determining what needs to be visible and drawn is, so I think it might need to be quite an upgrade to see a big difference.

especially when you compare it to something like this https://www.youtube.com/user/AtomontageEngine/videos

This is really amazing. How on earth does it all fit in memory? If the render distance is 16km, then that's somewhere around 16000 * 16000 * 2000 = 512 billion voxels, which would take somewhere in the order of 512GB of RAM. Obviously this is not the case, so what clever tricks are being used?

The trick is that there is an algorithm that describes the world so nothing needs to be stored until it is modified, and even then those changes are limited to a certain view distance. On the cpu side, a much lower resolution voxel data set is used for collision, pathfinding, etc (Also local to the camera)

I love seeing these videos. For a long time I appreciate Gavan's work. I hope he succeeds in creating an RPG game engine.

Thanks! Glad you like them. I apologize in advance that I kind of tend to drone on (both a combination of being tired (I shot this video several hours ago at around 5 am) and my already monotone voice.

Don't worry: your passion comes through in your work, if not in your voice.

EDIT. To give some unsolicited advice: From my experience giving presentations, if you want to give a demo that sounds good, I recommend around a dozen practice runs.

Or separate video and audio capture.

Both of which require spending a lot more time on it.

Wow, it's amazing how it evolved from an 2D style isometric low res voxel engine to a full 3D voxel engine.

And probably it will come full circle. I have an isometric mode I am going to use for low end machines (iso mode can be rendered once every couple seconds, instead of 60 FPS). With the exception of animated stuff, but that is relatively cheap to compute.

So glad I kickstarted this. Very excited about the progress. Keep on with it. One of the projects I'm okay waiting a long time for because I know it's going to be worth the wait!

Thanks for backing :) I hope it is!

I was curiously drawn to back this even though I haven't paid for any other video game of any kind in several years. It is exciting to see how far things have come. Can't wait to try it.

I'm flattered - can't wait to ship it :)

Link is dead right now, the video is here: https://www.youtube.com/watch?v=n0GPIvXFL0w

Thank you for making a "real" voxel engine/game. I get so tired of getting my hopes up when I see "a new voxel game", and it turns out to be polygonal based Minecraft rip off. That is not what I grew up with calling voxels. Ken Silverman's Voxel Engine and Delta Force are what I remember. It's a shame that the technology suffered for so long. Please keep up the good work and continue to innovate.

Thanks, will do :)

Will this game in the future be able to run, ported or otherwise, on the proposed Voxatron Game Console[0] by Lexaloffle[1]?

[0] http://www.lexaloffle.com/games.php?page=consoles

[1] http://www.lexaloffle.com/

I love Voxatron, but not likely - Voxatron's rendering is quite different. It is one of the few "pure" voxel engines in existence.

Wow, it's like looking into the viewport of the future of video games.

This looks amazing. I imagine this could create amazing games based on at least 2 titles:

- ultima - Great customisation / open world, but that's already mentioned on the kickstarter page.

- LBA - I would love for someone to create next version in this engine. Can't explain why, but it seems to match so well in my mind. Especially after LBA2 went 3d.

LBA has been brought up many times :) And yes, I am a huge fan of Ultima

The success of Minecraft comes from the ability to create. If you can capture that magic into Voxelquest I can see this being huge. I would love a game where I could create my terrain, set-up my army (like I would my toys) and then let them fight and be able to record and share the videos of the action. Wish you the best!

I find the achievement and persistence of Gaven absolutely astounding. You have my utmost respect and I wish you good luck and success with the project. I always love to see the updates on progress. If there were a game to buy I'd just buy it as a hat tip to show the respect.

Thanks! THere is a preorder page, but I will be no less honored if you just buy it when the alpha ships (which should be "soon")

Did not know preorder was possible. Just preordered now. Keep up the good work and good luck with the business!

why not open source the engine and develop a game around it? It'll not only gain exposure, take a lot of the optimizations off your belt, but let you focus on the long-term development of a marketable game.

I am going to open source it :) There will be a license for the first few years at least, but the source will be public. I'd like to figure out a way to do it with no license at all.

Don't worry too much about the actual open sourcing of it. Just push to github and forget. There will basically be two responses to the source code: either tinkerers or project managers.

Tinkerers will just want to poke around the code, and maybe make a youtube video.

Project managers will try to "fix" issues, "suggest" new directions, and basically try to get you to do things they want. :-P

Saw this in an earlier stage when you were showing off lighting/shading. Very impressive! I know the style isn't for everyone, but personally I love it. Looking forward to future progress :)

Looks great! Will you add anti-aliasing at some point in the future?

Yes, even going to do upscaling most likely (low res prepass with a high res detail pass)

How well does the voxel rendering parallelize based on number of cores? Related question: how much of the rendering is handled within the GPU?

Everything is rendered on the GPU. This is more based on ray tracing and marching than voxels alone (although voxels do represent the underlying terrain and destruction data). Ray tracing is highly parallelizable - each pixel can be computed completely independent of the others.

man watching your video makes me think of pilotwings 64, it would be a lot of fun to fly around your islands.

Go Schwa!

Ahh. Sweet, sweet validation. Love seeing you on here, buddy!

20 minutes of aimless muttering does not a good video presentation make. Impressive progress though.

I am not being at all sarcastic when I say that I agree. The combination of me being tired and rushing to get out the update so that I could focus on work were a recipe for for a pretty bad video. Marketing is definitely not my strong point, and moreover I can't afford the time to do it properly. But glad you like the progress :)

Reply on news.ycombinator.com

Voxel Quest January 2016 Update﻿

Voxel Quest January 2016 Update