Jump to content

"There are no limit in size?"


Jeronimo

Recommended Posts

I think there's a idea here, but what can we do with that idea now?

 

Voxels are volumetric objects. What games that feature building do is allow us to strap a bunch of those voxels together in order to make a functioning construct.

 

Now, we know core units will cost enough to make, so that it will be a deterrent to use one unwisely.

 

However, microvoxels are another construct and could take up more server resources depending on what they are used for. Let's say someone creates a massive ship out of microvoxels. It'd be cool, but depending on things, could increase the performance impact on the server. By how much though?

 

Creating microvoxels is a good idea, but the implications of it are what needs discussed here. We also need to know how the server performs normally, then allow a decision of whether or not microvoxels would work.

 

I think the server could handle them, but we aren't sure of all the implications.

Link to comment
Share on other sites

Most people will have ships a few dozen metres long at the smallest, and I'm sure some organisations will be making enormous transportation vessels and capital ships that are kilometres long. With such enormous constructs and huge amounts of 3D space to keep track of on the servers, you want to make the server load more strenuous by subdividing the already quite small quarter of a metre blocks to something even less? Most things that one would want to build at that size is probably more suited to being an element to save on space.

 

have you played something else than Pokemon and Minecraft?

 

Link to comment
Share on other sites

I think there's a idea here, but what can we do with that idea now?

 

Voxels are volumetric objects. What games that feature building do is allow us to strap a bunch of those voxels together in order to make a functioning construct.

 

Now, we know core units will cost enough to make, so that it will be a deterrent to use one unwisely.

 

However, microvoxels are another construct and could take up more server resources depending on what they are used for. Let's say someone creates a massive ship out of microvoxels. It'd be cool, but depending on things, could increase the performance impact on the server. By how much though?

 

Creating microvoxels is a good idea, but the implications of it are what needs discussed here. We also need to know how the server performs normally, then allow a decision of whether or not microvoxels would work.

 

I think the server could handle them, but we aren't sure of all the implications.

 

They ve open to tourism space journey for 32 million dollars, and in less than 20 years you ll be able to have it for 5000 dollars

 

maybe not long after 2 years later, you ll probably be able to buy SSD drive of 1t, graphic cards with 8go, processors 16 cores, so data isnt a problem neither internet speed, there will always be a way to compress and process data better

 

its about evolutions and innovations

 

NQ have started something that some other companies have mastered 4 years ago

Link to comment
Share on other sites

have you played something else than Pokemon and Minecraft?

Yes? How is that relevant? Pokemon needs only enough servers to run the global trading system and battling. The game is mostly singleplayer and most multiplayer interactions do not require the transferring of large amounts of data because all participants have all the information about moves, Pokemon and items. Minecraft is a single player game that has multiplayer servers not set up by Mojang/Microsoft.

 

In comparison, DU is a large, complex voxel game running with thousands of players on a single server. Just because modern games are capable of nice looking graphics does not mean that the server can handle it. Please make your points more coherent next time instead of forcing me to try and interpret what you mean.

Link to comment
Share on other sites

Yes? How is that relevant? Pokemon needs only enough servers to run the global trading system and battling. The game is mostly singleplayer and most multiplayer interactions do not require the transferring of large amounts of data because all participants have all the information about moves, Pokemon and items. Minecraft is a single player game that has multiplayer servers not set up by Mojang/Microsoft.

 

In comparison, DU is a large, complex voxel game running with thousands of players on a single server. Just because modern games are capable of nice looking graphics does not mean that the server can handle it. Please make your points more coherent next time instead of forcing me to try and interpret what you mean.

Indeed. Servers aren't some magical item that can run anything, just like our computers. Depending on how each voxel construct updates, it's performance, and the tasks it performs, it will affect the server differently. So, we can't simply just state that "yeah, microvoxels would be cool". We need to add the process of figuring out the impact on the server, deciding what these should be used for, and so on.

 

Each game has a development process. For DU, it's coming up with the ideas and systems, seeing if it's possible to create them, and then figuring out if they can handle them. They come up with so much stuff, but do you know how much stuff they've had to get rid of because it wouldn't be feasible? Not just server side, but client side. Every client will be more taxed the more stuff you throw at it. What would Microvoxels add to that stress?

 

The answer is, they are possible to implement. They would probably have to be restricted, but you could implement them. However, I'd rather see the game be here sooner and in a better playing state. Not to mention, I feel that already, we'll have the tools we need to create what we want. It's cool to have extra, but how many till we've reached a point where it's just not worth it.

Link to comment
Share on other sites

NQ have started something that some other companies have mastered 4 years ago

Tell that to the people who created No Man's Sky and Space Engineers.

 

If people mastered it 4 years ago, then why has it only emerged now? Why has this idea of a game such as DU only been possible now? The only people who tried to develop this kind of game so far as since then failed. 

Link to comment
Share on other sites

uhm.... people? ever heard of dual contouring (for the ones arguing over the enormounsness of the 25cm voxels)? they use that in DU, this allows for some degree of detail withing the 25cm volume, like slopes.

 

the point is a data point and a voxel are not exclusive, in the contrary. a datapoint is not a voxel, but a voxel has at least one datapoint to be able to be rendered. Euclidion has never demonstrated physics or real time modification, thus cannot be considered a voxel engine, but rather a datapoint visualizer.

 

if you want higher fine tuning of your structures i have already said in another post we could have custom parallax/bumpmap/custom shader textures and that would be feasible as option for people with monster rigs. Opengl is quite versatile in that manner. it would even be possible to use tesselation and a 3D texture to determine custom voxel interiors, or even raytraced voxel interiors for fractal effects.

 

all that is feasible. depending on their rendering engine, if they have enough control it could be implemented in 1-2 sprints.

Link to comment
Share on other sites

Most people will have ships a few dozen metres long at the smallest, and I'm sure some organisations will be making enormous transportation vessels and capital ships that are kilometres long. With such enormous constructs and huge amounts of 3D space to keep track of on the servers, you want to make the server load more strenuous by subdividing the already quite small quarter of a metre blocks to something even less? Most things that one would want to build at that size is probably more suited to being an element to save on space.

 

they said they bake the meshes continuously, thus only modifications impact performance.

Link to comment
Share on other sites

Indeed. Servers aren't some magical item that can run anything, just like our computers. Depending on how each voxel construct updates, it's performance, and the tasks it performs, it will affect the server differently. So, we can't simply just state that "yeah, microvoxels would be cool". We need to add the process of figuring out the impact on the server, deciding what these should be used for, and so on.

 

Each game has a development process. For DU, it's coming up with the ideas and systems, seeing if it's possible to create them, and then figuring out if they can handle them. They come up with so much stuff, but do you know how much stuff they've had to get rid of because it wouldn't be feasible? Not just server side, but client side. Every client will be more taxed the more stuff you throw at it. What would Microvoxels add to that stress?

 

The answer is, they are possible to implement. They would probably have to be restricted, but you could implement them. However, I'd rather see the game be here sooner and in a better playing state. Not to mention, I feel that already, we'll have the tools we need to create what we want. It's cool to have extra, but how many till we've reached a point where it's just not worth it.

 

dude, you have everything in the wrong way. First you worry about the idea, then you worry about the implementation and then you worry about the hardware cost. Servers ARE these magical boxes that can do everything and ARE like our PCs.

 

If things are not done in that order you cannot know if the idea is worth the costs in the first place and thus may miss on opportunities and even decide on implementing a stupid feature. you need to know why you do something before you know how you do it and then know what you are doing on the servers. If you start in the servers, you will use a raspberry pi to print hello world! instead of having a good DU experience.

I did the calculations for the cost/win ratio of a cuda computed space battle for 2000 people on the server side, and came to the conclusion there is no reason to limit the development cost as GPUs are cheap as hell nowdays, and cuda is extremely well implemented/documented and allows for reusability of cpp code which allows for an easy implemenation of cuda computed physics and other things.

Link to comment
Share on other sites

they said they bake the meshes continuously, thus only modifications impact performance.

I know that the modifications to the procedurally generated content is what's actually stored (it's a clever idea to keep used server space lowered) but having more voxels to track per unit volume still increases the amount of data that needs to be transferred. As you pointed out, there is some workarounds even within the 25cm space but I think that needlessly decreasing the voxel size isn't worth it at this stage when NQ isn't even sure how well their tech will handle all the players.

Link to comment
Share on other sites

I know that the modifications to the procedurally generated content is what's actually stored (it's a clever idea to keep used server space lowered) but having more voxels to track per unit volume still increases the amount of data that needs to be transferred. As you pointed out, there is some workarounds even within the 25cm space but I think that needlessly decreasing the voxel size isn't worth it at this stage when NQ isn't even sure how well their tech will handle all the players.

 

their tests seem promising, i wouldn't worry about that. my argument was mostly that there is no point to further reduce voxel size, and then proposed alternatives that would be cheaper.

Link to comment
Share on other sites

their tests seem promising, i wouldn't worry about that. my argument was mostly that there is no point to further reduce voxel size, and then proposed alternatives that would be cheaper.

give alternative example plz

Link to comment
Share on other sites

Tell that to the people who created No Man's Sky and Space Engineers.

 

If people mastered it 4 years ago, then why has it only emerged now? Why has this idea of a game such as DU only been possible now? The only people who tried to develop this kind of game so far as since then failed.

this topic is about voxels

 

at least watch the video about landmark few posts before

Link to comment
Share on other sites

dude, you have everything in the wrong way. First you worry about the idea, then you worry about the implementation and then you worry about the hardware cost. Servers ARE these magical boxes that can do everything and ARE like our PCs.

 

If things are not done in that order you cannot know if the idea is worth the costs in the first place and thus may miss on opportunities and even decide on implementing a stupid feature. you need to know why you do something before you know how you do it and then know what you are doing on the servers. If you start in the servers, you will use a raspberry pi to print hello world! instead of having a good DU experience.

I did the calculations for the cost/win ratio of a cuda computed space battle for 2000 people on the server side, and came to the conclusion there is no reason to limit the development cost as GPUs are cheap as hell nowdays, and cuda is extremely well implemented/documented and allows for reusability of cpp code which allows for an easy implemenation of cuda computed physics and other things.

 

They maybe magical but they don't have infinite computational power and memory. Each voxel takes memory, so more voxels means more memory, more bandwidth, more servers in the cluster and ultimately more money spent each month hosting the game. If they do introduce a "microblock" core unit, they should limit the size and functionality so its just for fine detail and decorative items other wise some one will build a kilometer long cruiser out of micro blocks and load down the server.

 

Honestly though 25cm voxels with the dual contorting stuff is going to be good enough for 99% of all ship designs, you can make fine details, landmark is based on the same tech as dual universe.

Link to comment
Share on other sites

They maybe magical but they don't have infinite computational power and memory. Each voxel takes memory, so more voxels means more memory, more bandwidth, more servers in the cluster and ultimately more money spent each month hosting the game. If they do introduce a "microblock" core unit, they should limit the size and functionality so its just for fine detail and decorative items other wise some one will build a kilometer long cruiser out of micro blocks and load down the server.

 

Honestly though 25cm voxels with the dual contorting stuff is going to be good enough for 99% of all ship designs, you can make fine details, landmark is based on the same tech as dual universe.

Imagine a big cube of 1000000000 meters by side, with one material, made by 1 mm cubes, how much space does ti take in memory?

1000000000000 bytes? Nope. Actually only one.

the point is voxels are structured in an octree and will be optimized, a bigger cube will be considered one voxel. Having smaller voxels only increases the data usage on the edges and increases baking time, but for a server is not too problematic as it only has to manage the memory, and isn't a lot. The problem would come in the client side as baking would take exponentially more time, as a recursive algorythm is used (in general they may have optimized it to avoid stacking the stack) to transform each step in a mesh.

 

In these kind of applications, servers do physics and have to do those to avoid hackers, and provide sync between clients. And physics are the thing to worry about.

 

But i am the devils advocate. In DUs application 25cm is enough, unless you wanna do microdrones but that can be added separately since they use voxel farm that supports any size.

Its for the argument's sake.

Link to comment
Share on other sites

Imagine a big cube of 1000000000 meters by side, with one material, made by 1 mm cubes, how much space does ti take in memory?

1000000000000 bytes? Nope. Actually only one.

the point is voxels are structured in an octree and will be optimized, a bigger cube will be considered one voxel. Having smaller voxels only increases the data usage on the edges and increases baking time, but for a server is not too problematic as it only has to manage the memory, and isn't a lot. The problem would come in the client side as baking would take exponentially more time, as a recursive algorythm is used (in general they may have optimized it to avoid stacking the stack) to transform each step in a mesh.

 

In these kind of applications, servers do physics and have to do those to avoid hackers, and provide sync between clients. And physics are the thing to worry about.

 

But i am the devils advocate. In DUs application 25cm is enough, unless you wanna do microdrones but that can be added separately since they use voxel farm that supports any size.

Its for the argument's sake.

 

If they stick to uniform shapes sure. But even a borg cube wont be a solid cube. And what if I were to do a net around my ship with as thin of a web as I can to catch missiles. That would be hell with smaller voxels. 

 

Added:  I do think it would be nice if later on we could create decorative components, that dont have any function, on a smaller scale. So you would be given a specific volume, maybe 1m3, where you can work with much smaller voxels to make decorative elements. Then that could be placed in the world as a solid component. 

Link to comment
Share on other sites

If they stick to uniform shapes sure. But even a borg cube wont be a solid cube. And what if I were to do a net around my ship with as thin of a web as I can to catch missiles. That would be hell with smaller voxels. 

 

Added:  I do think it would be nice if later on we could create decorative components, that dont have any function, on a smaller scale. So you would be given a specific volume, maybe 1m3, where you can work with much smaller voxels to make decorative elements. Then that could be placed in the world as a solid component.

 

Absolutely but if your watch the demo voxels are placed with 3D brushes and not one by one so your point is quite irrelevant in that aspect. For what concerns the borg cube it will barely be heavier than 25cm voxels as most components are meter-wide, but yes you are right in absolute but the gain in size will be about 5% max for a 25cm to 1cm swap (guesstimation)
Link to comment
Share on other sites

They maybe magical but they don't have infinite computational power and memory. Each voxel takes memory, so more voxels means more memory, more bandwidth, more servers in the cluster and ultimately more money spent each month hosting the game. If they do introduce a "microblock" core unit, they should limit the size and functionality so its just for fine detail and decorative items other wise some one will build a kilometer long cruiser out of micro blocks and load down the server.

 

Honestly though 25cm voxels with the dual contorting stuff is going to be good enough for 99% of all ship designs, you can make fine details, landmark is based on the same tech as dual universe.

the micro block from landmark is still a 25cm voxel

 

the scale of 25cm was for a technology from before 2014. DU will launch in 2019, thecnology 5 years later will have improved a lot, enough to allow voxel of 15cm for sure

Link to comment
Share on other sites

the micro block from landmark is still a 25cm voxel

 

the scale of 25cm was for a technology from before 2014. DU will launch in 2019, thecnology 5 years later will have improved a lot, enough to allow voxel of 15cm for sure

Can they make smaller scale voxels, sure. But remember DU is doing a lot more than Landmark.

NQ is being co servative for anything that takes computing power. There are a lot of features that could be implimented, but you can only do so many. They have to set limits to have a versitile functional game.

Link to comment
Share on other sites

Can they make smaller scale voxels, sure. But remember DU is doing a lot more than Landmark.

NQ is being co servative for anything that takes computing power. There are a lot of features that could be implimented, but you can only do so many. They have to set limits to have a versitile functional game.

 

 

whats the point in doing a lot more but no doing better?

 

NQ has developed their own single shard universe, very good!

 

But DU :

- has voxel build from landmark, and as demonstrated, the way is still long

- has scripting part from Second Life, simplified for better use for all of us

- has scifi infinite universe theme from EveO

 

Its like making a salad, you mixing all the good stuff you like, and put the right sauce on it

 

NQ single shard concept is that sauce, where everyone is like wow, so exited about, because it covers well the rest.

 

Do you really think will cover the after-taste of the rotten vegetables you put in your salad?

Link to comment
Share on other sites

In the kickstart promotion video, at 0.50min, JC mentioning that "you can build anything you want, space ships, cities, orbital stations, there is no limit in size"

 

Jc mentioning here the infinity big, from voxels of 25cm

 

What about the infinity small? or at least smaller than 25cm?

 

 

NQ is missing something in the building part, 25cm voxels will not allow to build a chair, a table, a lamp, a frame, or any decoration objects, to put in (and in fact the only things you can build) empty space ships, empty cities, and empty orbital stations

 

Its like, think big, look at it from far, and forget about details

 

 

The content of DU is pretty well thought in global, at the scale of civilisations organisations, but what about in details, at the scale of one player?

 

What will one player see ingame?

Very detailed textured mesh based elements stuck on very contrasting flat voxel builds?

 

 

NQ shouldnt forget who will pay and play, who they are targeting when they talk to builders and coders, to lot of people who spend their days on 3d modeling softwares pushing the latest technologies, day by day, making scenes more and more realistic that you cant make anymore the difference.

 

I m talking about gamers who are also architects, interior designers, industrial designers, graphic designers, CG artists, animators, developpers, coders etc...

 

Is DU going to be the game for gamers who grew up with 1m3 blocks in minecraft, who got maturity now to play with 0.25m3 voxels?

 

NO right?!

DU isnt based on java and still have a bit more than 2 years of development

 

 

And i think the community should keep in mind to push further and further the possibilties of the building tool of DU to make it the greatest sandbox of all times

Going below 25 cm requires some special code that three other companies already own the rights to and in all three cases it requires a lot more data to save the results. Patents matter. NQ is not Landmark/ Planet Explorers: voxel farm, Always geekie's Vox or Voxelnauts. 

Reinventing the wheel is hard and patent battles is harder. 

 

We can make lots of basic chairs etc as as normal meshes and pitch them to NQ but memory limits matter. We could do voxel sit elements that hide in the top of some voxel's but 25 cm smoothed to perhaps 5 cm in some cases would be the limit. And a voxel chair is immovable. mmm I can think of a fix for that last problem. 

Link to comment
Share on other sites

Going below 25 cm requires some special code that three other companies already own the rights to and in all three cases it requires a lot more data to save the results. Patents matter. NQ is not Landmark/ Planet Explorers: voxel farm, Always geekie's Vox or Voxelnauts. 

Reinventing the wheel is hard and patent battles is harder. 

 

We can make lots of basic chairs etc as as normal meshes and pitch them to NQ but memory limits matter. We could do voxel sit elements that hide in the top of some voxel's but 25 cm smoothed to perhaps 5 cm in some cases would be the limit. And a voxel chair is immovable. mmm I can think of a fix for that last problem. 

 

wow. just.... wow.... do you realize what you said? patents for voxel sizes are a problem? no. NQ is based in Europe, and in European jurisdiction you cannot patent a concept you need a working prototype (that's why you cannot patent pseudo scientific idiocies). you can patent a specific text (code) but is invalidated if you change the names of variables so they will not be sueable outside the US.... Additionally, they use voxel farm, a commercial voxel engine which demonstrated tiny voxels. And even if it was the case you could have scale up the rest of the world. In a program, size doesn't matter, relative scale does. how can you be sure the voxels are 25cm, what if they told you 1, you wouldn't have noticed because there is no absolute reference...

 

Additionally, dual contouring can be as precise as you want, it's only the detail (minimum thickness) that will be limited.

Link to comment
Share on other sites

How about just allowing voxel build smaller than 25cm for decorative builds, but limiting the number of voxels that can be attached to each others. Limiting drasticaly the scale of tiny voxels builds

 

What is better for servers or our computers?

100 people sharing a milion voxel ship or single dude with a folowing bot of 300 tiny voxels?

 

(well it can also be 100 dudes with tiny voxels following bots, in a million voxel ship, will be just 1 300 000 voxels ;-) )

 

And tiny voxels builds can be limited to decoration only aswell, so the physic problem is solved?

Link to comment
Share on other sites

wow. just.... wow.... do you realize what you said? patents for voxel sizes are a problem? no. NQ is based in Europe, and in European jurisdiction you cannot patent a concept you need a working prototype (that's why you cannot patent pseudo scientific idiocies). you can patent a specific text (code) but is invalidated if you change the names of variables so they will not be sueable outside the US.... Additionally, they use voxel farm, a commercial voxel engine which demonstrated tiny voxels. And even if it was the case you could have scale up the rest of the world. In a program, size doesn't matter, relative scale does. how can you be sure the voxels are 25cm, what if they told you 1, you wouldn't have noticed because there is no absolute reference...

 

Additionally, dual contouring can be as precise as you want, it's only the detail (minimum thickness) that will be limited.

 

no need to be so rude, all this (voxels, mmo, space, politic) is a big discovery for most members

you sometimes, even for me, hard to follow because very pointy, that is very precious during this pre alpha periode for dev team (if they read all what you posted)

dont expect all members have same knowledge as you, and sometimes you should find better words for your explainations

 

there are tons of absurdities and misjudgement in the forum, but its like that

please take more time to teach what you know and can bring us

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...