Voxel Farm is coming to Unreal Engine. In fact, this is not so new (see here), but there are some news:
The last version on Voxel Studio can now build package to be used inside unreal, through a provided warper (in early stage of developpment, for evaluation purpose). Voxel Farm team is working on a true plugin.
Voxel Farm (the engine) and the IDE to edit Voxel data (Voxel Studio) can be seen as a content creator suite (like 3DS max or so) with a intimate integration in engines (like Speetree or Substance). What it brings is the voxel representation of the world with LOD and Physics, in game. That means total freedom to landscape AND building representation, including caves, overhang, etc, and with a great ease of modification (inside Voxel Studio). On the generation side of the product, the building grammar used inside voxel studio is a very powerfull approach to create procedural content (using a CGA-shape like approach). You can also have realtime modification of voxel (destruction, at least, and maybe creation).
The typical workflow would be to feed standard Heighmap (generated in WorldMachine for example) to Voxel Studio, where you carve and sclupt, add caves, building, etc, then use the Unreal package to go on with the level generation in Unreal Engine4.
I think this is a solid way of inviting Voxels to Unreal Engine !
That looks nice, thanks for sharing! When they made their plugin for easy and improved export i´ll buy it! And the word “destruction” makes everything better anyway. Btw does anyone know, if you can use the “Creator” license also for commercial purpose? Cuz it´s very cheap and i can´t see something like “we want x% of your income if you make more than y money in time z”. I could give them 100% of my income, as it´s 0.00€ But it also doesn´t say, that it´s only for educational purposes or non-commercial.
This is perhaps a bit of a rant, and it’s not directly related to Voxel Farm, but seeing the revenue terms kind of made me want to voice this.
So, for an ‘indie’ plan, you have a triple-combo of an upfront license-fee, a recurring monthly subscription fee, and a five percent cut of revenue. That’s not an insignificant bit of outlay, especially for a tool so niche.
Maybe I’m just doing the old-man-shouting-at-clouds routine, but I’m really growing weary of the subscription/service model. Tools like these are not services, they’re products. Back when 3DS Max was offering upfront fees, it would have been feasible to buy it on credit and pay maybe £100 a month over a few years to own a fully-featured product with relatively reasonable upgrade pricing if you needed functionality from a newer version. Now Autodesk are charging £145 per month in perpetuity for the same product and phasing out any option to buy a license upfront.
With almost every tool in a regular 3D production pipeline now moving to SAAS, ‘cloud-based services’, or whatever other marketing buzzwords the developers employ for it, it becomes a pretty poor do for a lot of indie devs. That said, I think companies like Allegorithmic (pay a modest monthly fee for the Substance suite and own the product once your payment balance exceeds the value of the software) and Pixologic (just charge a reasonable flat price for Zbrush and updates within the current major version) get the balance right concerning affordability and features; but this kind of presumption that upstart developers will be happy to pay through the nose monthly, and give up a cut of their budget for a product of only marginal usefulness to most production pipelines is just something I don’t see panning out all that well.
You’re appealing to developers who have already figured monthly subscriptions for, potentially, Adobe Creative Cloud and Autodesk 3DSMax/Maya, as well as revenue cuts from Unreal Engine, and upfront licenses for Wwise, Zbrush and potentially other tools into their budget. Why the heck would they consider shelling out even more money for a tool that in maybe 90% of applications, vanilla UE4 can handle just fine?
Well, Voxel Farm is not the holy grail, but it solves 2 problems :
terrain modeling not constraints to heigthmap
procedural generation that feels less repetitiv (L-system based)
If you dont need to solves one of this problem (or you have alternate solution), well, you don’t need it ! Criticism (about the licencing and prices) were voiced on the Miguel Cepero Blog. I am a big fan of its work, but the pricing scheme was like a cold shower to me when he goes the Voxel Farm way.
In the end, I choose a creator licence: don’t expect more from this licence than a content creation tool, with plugin for Unity, and, if I get right, for Unreal 4 soon for the same price.
For me, the Creator licence was very sufficient, to play with it and create specific parts, while keeping the standard UE4 way of doing landscape and level design (WM for landscape/UE4).
This tool have a procedural approach to generate content based on Voxel, with a high level of control.
On my point of view, the strength of the creator suit :
the power of voxel, usefull in some situation (especially landscape). Convincing caves are easy to do.
the dynamic alteration of the voxel world (through plugin)
the very powerfull L-system-like grammar approach to build procedural content (building especially, but not only).
a reasonnably well-made UI to allow for quick geometry modelling. You can even voxelize pre-made model.
If you need a deeper integration in your product (for example, a full in-engine procedural generation and world manipulation without the voxel studio), you’ll need an Indie, or more, and I personnaly think that you’d better have a strong need for it, because it can be expensiv.
But some AAA studios took this direction, and the result is very impressive (see the Landmark test for example, we are talking about Voxel-based MMO worlds).
And the same way sketchup and blender and many other tools are now getting more and more share and getting accepted even for professional work, the voxel crusher will come, either with a reasonable price or just for free and we will rather donate money to them than to scams like Voxel.
Voxel seems nice only for AAA studios that can cut some decent deal, but if anybody expects me to pay for that tools such money is very wrong.
Voxel farm is the best current stand alone dual contour solution on the market and the creator has put years into it, there is only like one competing library out there and it is locked to their own engine. The prices are understandable considering the overhead and potential of the tools involved, it is targeting AAA games to put it out there more and honestly the tool suite is complicated enough that I’m not sure they want too many indies trying to use it with low experience.
It isn’t for everyone but it is **** impressive and respectable, I won’t be buying it for myself but I would love to see some people make use of it with UE4.
Also that creator license is going to be great for terrain, since they can handle sharp edges it will be a really high quality base mesh.
It’s very impressive technically and I’ve followed his blog in regards to landscape generation.
Unfortunately certain things look rather strange, maybe it’s just issue with the presentation and not the technology behind. One of things that catches my eye is that interior and exterior of the voxels uses the same material. Like cut out of the wood beam have exactly the same texture as it’s side, just different angle of UV mapping. Regarding that, isometric rendering in Voxel Quest created much more appealing visuals and perception that objects are really made out of volumetric materials:
Perhaps it’s something that can be easily added to Voxel Farm and I’m just spoiled to expect it to be already there as a basic feature
The other thing was, two projects that I’ve seen so far using Voxel Farm, which are TUG and StarForge have much worse quality of the landscape comparing to what is shown in VoxelFarm videos. There were a lot of glitches in geometry and surface itself looked rather unfinished. It just doesn’t have a level of visual polish which you would expect from an AAA middleware.
Perhaps pricing terms will change when other competitive solutions appear on market.
Voxel quest uses I believe ray marching or at least something like it to render its objects, the reason voxel farms fracturing has the same texture inside of its models after fracturing is because the entire model is composed of the same voxel data. It compresses better, generally not an issue, but as you noted is glaring during fracturing, could be fixed but any real fix would come at an extraction and/or compression cost. Voxel farm converts the underlying voxel data directly to polygonal meshes, voxel quest does not and just renders the color data to screen.
Because of this Voxel farm plays well with 3rd party engines, something like voxel quest would not be an easy integration to another pipeline (although it truly looks fantastic).
As far as use cases and lack of quality, the SDK isn’t simple from what I understand and isn’t easy to use really well, obviously the creator of the system himself will have the best results in a case like that. Another reason why I think it may be targeted towards more experience teams instead of smaller inexperienced ones.