A voice interview I did for the gaming site GiantBomb.com just came out:
http://www.giantbomb.com/podcasts/the-future-is-voxels/1600-624/
There is a lot of my usual crazy talk there. If you want to take on any of these issues here, please leave a comment in this post.
Great interview, while listening through an idea sparked into my head.
ReplyDeleteYour hiring of a content creator(s) is a great idea. They will add tremendous value to your project and company, in terms of attracting potential customers as well as potential investment. But, what is the next step?
At some point in the future, would you consider making an online instance library for developers to use with your engine? Larger games/companies will have the resources to custom craft large amounts of custom/unique instance objects, but a crowdsourcing effort will invariably make your engine 10x more attractive to smaller independent startup games. A video or two showcasing how easy it is to add new instance objects to your engine will go a long way in swaying traditional developers into using your product.
Yes an online library would work. It could take the form of a marketplace. I hope if the tools are good and the data formats simple and open, people will share in any way that works with them.
DeleteIn general I favor non-centralized models. That is, I would not store all this data in the cloud necessarily. I think people should be able to own their creations. If they want to store their content online, that's one particular service, but it does not have to be the only way.
A marketplace is a great idea to generate content, however you should be careful in its execution. If the only way to acquire instance objects is to pay for them or create them yourself, you will effectively alienate hobbyists from using your tools. You don't want to do that, because so many developers in the community start out as hobbyists.
DeleteIf you follow the model that the google play store has set up you should be good. But you may need to hire someone or a company to set that up for you.
You are most definitely in the forefront of voxel software XD (Atleast, of the publicly announced stuff). I cant wait for the scenario you described: HD Voxels. How long do you personally think that it will take the Hardware manufacturers to make graphics cards which are optimised with voxels in mind?
ReplyDeletePolygons still do very well for rendering. The advantage in voxels is more about storing and processing information. I do not think hardware manufactures will add any circuitry for voxels. They will continue to improve the performance of their hardware in more general purpose computing. Voxel processing is just one application of this.
DeleteEven if Moore's law holds, growth won't be that fast. 3D is really expensive. For instance, if you have voxels that measure one 1x1x1 meters today in two years you can have 1x1x0.5, in four years 1x0.5x0.5 and only in six years you will hit 0.5x0.5x0.5. So growth has to happen on each individual dimension. Voxel sizes will shrink by half every six years.
Let's say a $1000 worth of hardware runs 0.25 meter voxel sizes this year. In 12 years the same grade of hardware will run voxels at 0.06 meters.
This is still large for some applications. In general I do not see voxels replacing everything we do with polygons. They work and makes sense for a lot of them, not all.
Thanks for answering my question, it creates quite a bit of insight =D
DeletePerhaps you should give Euclideon a call? They say they can stream massive landscape scans over the Internet, so why not?
DeleteThey have not proven that yet. I don't think their tech can be used in games for the current and next hardware generation. You can see my take on their approach here: http://procworld.blogspot.ca/search/label/Euclideon
DeleteThey've proved it to several geo companies. Give them a call and they'll come round and demo it for you, they may even offer to convert your terrain for you as a demo. It will be fun.
DeleteThey have not proved you can make a game. A geo company will be interested in different things, does not really translate into games.
DeleteIf they could do game they would have done games, this was their original angle and why they got so much attention in the first place. They claimed they had something 100,000 times better than the CryEngine or Unreal Engine.
If they had been all about geo visualization since the beginning we would not be talking about Euclideon now.
I don't think they even understand what a game needs, it would be interesting just to find out what it can do though, and they say it's single threaded.
DeleteThe landscape is not changeable though, so yeah a waste of time - I wanted to see it working in a cool game environment. :)
Bruce Dell from Euclideon is just a sensationalist. I believe he is knowledgable enough to realize what his tech really is, but he presents it in such a way that it attracts headlines. Thats how he got his government grants and maybe even commercial contracts although there is no solid proof that he got any real clients other than his words.
DeleteNice interview. I was nodding in agreement with your crowdsourcing vision - my sons are minecraft addicts and I have seen a small portion of the vast array of amateur construction out there (some of which is very high quality). Even with minecraft's limited construction tools, my 8-year old has built some astonishingly complex castles and suchlike. From what I've seen of your tools so far, voxel farm custom content promises to be very exciting. I can't wait to be able to let him loose on voxel farm! (his dad is rather keen to try it out too... ;-)
ReplyDeleteOne further thought: What kind of content are you aiming bring in from crowdsourcing that can't (eventually) be generated on the computer? I'm assuming your ultimate goal for the engine would be that level/story designers could drop procedural castles/houses/bridges/features etc into specific locations in their world without having to lift them from a crowdsource database of items? Or have I misunderstood where you're going with this?
I do not have the full picture yet. The reason why is because this system has to be built from the ground up as an entertaining experience. Once the basic bones of the system generate engaging gameplay, we can start layering other features, like content that comes from players, assemblies of mods, etc.
DeleteFor me personally procedural asset generation is of interest, because I am a programmer and don't have the skill to create game assets, but for a solo, in pajamas development effort, I can't afford to pay artists.
DeleteVery interesting
ReplyDeleteI am not sure that this post is the ebst to ask, what do you think about the latest release of Mantle, amd's low level api foror as an alternative to opengl and directX
ReplyDeleteI do not know the specifics. The promise of less overhead in API calls is interesting.
DeleteIn general I feel skepticism whenever I see a new API being announced. With AMD taking over console hardware they are in good position to get a new standard out and strengthen their sales for PC. I'm not convinced a new API is the only way to advance, I rather see more consistency and performance out of the standards we already have, like OpenCL, WebGL, etc.
Also they claim the improved performance comes from a better alignment between the API and the new GPU architecture. This is a double-edged sword. What happens when the architecture becomes obsolete? If the API is coupled to it, it may also go away.