Jump to content

  • Log In with Google      Sign In   
  • Create Account

Creating open world map


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
14 replies to this topic

#1 EthanRev6   GDNet+   -  Reputation: 204

Like
0Likes
Like

Posted 24 February 2014 - 05:27 AM

Hey People that I know not but refer to you for ultimate knowledge and understanding,

 

          I'm curios to know what the best way to create a 3d high detailed open world Island would be. Or if importing it into different programs for variation of detail would be easier. I plan on using Unreal s Game Engine to actually make the game, but are there easier faster ways to create a world. I want the level to be about the same size of the new gta5. I know that is a huge task just didn't know what everyone's view was on the subject.


"And I looked, and behold a pale horse: and his name that sat on him was Death, and Hell followed with him"


Sponsor:

#2 Ashaman73   Crossbones+   -  Reputation: 7492

Like
2Likes
Like

Posted 24 February 2014 - 07:50 AM


Unreal s Game Engine

If you want to use a large island like world, I would take a look at the crytek engine (Crysis/Farcry) which might be better suitable to this task then the more level based UDK engine.



#3 creatures-of-gaia.com   Members   -  Reputation: 377

Like
2Likes
Like

Posted 24 February 2014 - 11:03 AM

Interesting, I thought it was the opposite. I thought that UDK was better for very large worlds and crytek better for "normal" worlds because the engine looses precision when the player is very far away from the origin. I may be wrong though, I read it quite some time ago and it should be verified. That said, I think crytek engine has tools to build island terrain and such.



#4 EthanRev6   GDNet+   -  Reputation: 204

Like
0Likes
Like

Posted 24 February 2014 - 05:59 PM

 


Unreal s Game Engine

If you want to use a large island like world, I would take a look at the crytek engine (Crysis/Farcry) which might be better suitable to this task then the more level based UDK engine.

 

 

 

Interesting, I thought it was the opposite. I thought that UDK was better for very large worlds and crytek better for "normal" worlds because the engine looses precision when the player is very far away from the origin. I may be wrong though, I read it quite some time ago and it should be verified. That said, I think crytek engine has tools to build island terrain and such.

 

 

 

0_o Well now I am confused haha 


"And I looked, and behold a pale horse: and his name that sat on him was Death, and Hell followed with him"


#5 Ashaman73   Crossbones+   -  Reputation: 7492

Like
2Likes
Like

Posted 25 February 2014 - 01:01 AM


I thought that UDK was better for very large worlds and crytek better for "normal" worlds because the engine looses precision when the player is very far away from the origin. I may be wrong though,

To evaluate an engine, always take a look at its reference games. GTA is a continuous world, much like Farcry/Crisis. Whereas I do not know of any unreal-based game with a large continous world (thought it could be out there , I dont really know all UDK games wink.png ). All udk based games I played so far have a very defined level structure, with quite small levels.

 

Thought, there are always very specialised engines out there, which will do a much better job for this task, GTA,Just Cause 2 or Frostbite (BF4), but they are not really as available as cryengine, udk or unity.



#6 creatures-of-gaia.com   Members   -  Reputation: 377

Like
1Likes
Like

Posted 25 February 2014 - 02:31 AM

...well, I was curious, so I googled a bit (btw, what the OP should have done!)

 

Here are some excerpt from forum posts I found:

 

CryEngine

 

The practical maximum map size for a Cryengine 3 game is 4096m x 4096m - technically you can make terrains bigger than that but you'll start experiencing memory issues at larger sizes. There is a way to work around this limitation by building the actual terrain of the map from externally modeled meshes instead of the built-in terrain tools (a user called cwright has managed to build a working flight sim game spanning hundreds of kilometers across) but you'll start to experience precision problems such as player camera vibration, fog flickering etc at large distances from the map origin.

 

UDK

 

At 1uu == 1cm the physical limits of actor movement, collision and physics are 27 square kilometers. The physics of a map max out at 524,288 uu, beyond that only PHYS_None will work. Additionally you can encounter visual artifacts due to float imprecision as you approach that boundary.

 

 

...it's likely both engine have their limits about open worlds. Such maps are already huge. If you need to grow bigger, perhaps there are some workarounds, perhaps not.



#7 EthanRev6   GDNet+   -  Reputation: 204

Like
0Likes
Like

Posted 26 February 2014 - 12:07 AM

...well, I was curious, so I googled a bit (btw, what the OP should have done!)

 

Here are some excerpt from forum posts I found:

 

CryEngine

 

The practical maximum map size for a Cryengine 3 game is 4096m x 4096m - technically you can make terrains bigger than that but you'll start experiencing memory issues at larger sizes. There is a way to work around this limitation by building the actual terrain of the map from externally modeled meshes instead of the built-in terrain tools (a user called cwright has managed to build a working flight sim game spanning hundreds of kilometers across) but you'll start to experience precision problems such as player camera vibration, fog flickering etc at large distances from the map origin.

 

UDK

 

At 1uu == 1cm the physical limits of actor movement, collision and physics are 27 square kilometers. The physics of a map max out at 524,288 uu, beyond that only PHYS_None will work. Additionally you can encounter visual artifacts due to float imprecision as you approach that boundary.

 

 

...it's likely both engine have their limits about open worlds. Such maps are already huge. If you need to grow bigger, perhaps there are some workarounds, perhaps not.

Thank you. I did some googling as well and I have researched quite a bit about it, but im not engaged in an actual discussion about it. That is why I created this. I didn;t know if anyone had any particular success with either system. Thanks again.


"And I looked, and behold a pale horse: and his name that sat on him was Death, and Hell followed with him"


#8 Ashaman73   Crossbones+   -  Reputation: 7492

Like
0Likes
Like

Posted 26 February 2014 - 01:04 AM



Like
1Likes 2Likes
Like

Posted Yesterday, 03:31 AM

...well, I was curious, so I googled a bit (btw, what the OP should have done!)



Here are some excerpt from forum posts I found:



CryEngine



The practical maximum map size for a Cryengine 3 game is 4096m x 4096m - technically you can make terrains bigger than that but you'll start experiencing memory issues at larger sizes. There is a way to work around this limitation by building the actual terrain of the map from externally modeled meshes instead of the built-in terrain tools (a user called cwright has managed to build a working flight sim game spanning hundreds of kilometers across) but you'll start to experience precision problems such as player camera vibration, fog flickering etc at large distances from the map origin.



UDK



At 1uu == 1cm the physical limits of actor movement, collision and physics are 27 square kilometers. The physics of a map max out at 524,288 uu, beyond that only PHYS_None will work. Additionally you can encounter visual artifacts due to float imprecision as you approach that boundary.





...it's likely both engine have their limits about open worlds. Such maps are already huge. If you need to grow bigger, perhaps there are some workarounds, perhaps not.

Scale is really, really very vague in games:

  1. Scale is relative. The perception of scale in games comes from the visual aspect. The brain compares known objects (eg human character, mountain) with other objects in the scene and derives a kind of sense of scale from it. It really doesnt matter at all, if the human is 2m, 1cm or 3.14 km tall !
  2. The physics of a game world need to be much smaller. If you model a human movement speed of 5km/h you would need almost one hour to walk on a 4096m x 4096m map from one side to the other. Now make your map 4096km x 4096km and you would need 1000 hours game time. Games are not about perfect reality and even less about long traveling times.

 

Therefor, the scale of a world is more an illusion and there are other more important factors for a game world. One of the most important features is:

Is the engine capable to render and simulate a game world which delivers the illusion of a large world?

 

In this sense and in my opinion, cryengine is better suited for this task than udk, even if udk have larger levels on the paper ! Take always a look at referenc games to evaluate a game.



#9 EthanRev6   GDNet+   -  Reputation: 204

Like
0Likes
Like

Posted 26 February 2014 - 01:16 AM

 



Like
1Likes 2Likes
Like

Posted Yesterday, 03:31 AM

...well, I was curious, so I googled a bit (btw, what the OP should have done!)



Here are some excerpt from forum posts I found:



CryEngine



The practical maximum map size for a Cryengine 3 game is 4096m x 4096m - technically you can make terrains bigger than that but you'll start experiencing memory issues at larger sizes. There is a way to work around this limitation by building the actual terrain of the map from externally modeled meshes instead of the built-in terrain tools (a user called cwright has managed to build a working flight sim game spanning hundreds of kilometers across) but you'll start to experience precision problems such as player camera vibration, fog flickering etc at large distances from the map origin.



UDK



At 1uu == 1cm the physical limits of actor movement, collision and physics are 27 square kilometers. The physics of a map max out at 524,288 uu, beyond that only PHYS_None will work. Additionally you can encounter visual artifacts due to float imprecision as you approach that boundary.





...it's likely both engine have their limits about open worlds. Such maps are already huge. If you need to grow bigger, perhaps there are some workarounds, perhaps not.
 

Scale is really, really very vague in games:

  1. Scale is relative. The perception of scale in games comes from the visual aspect. The brain compares known objects (eg human character, mountain) with other objects in the scene and derives a kind of sense of scale from it. It really doesnt matter at all, if the human is 2m, 1cm or 3.14 km tall !
  2. The physics of a game world need to be much smaller. If you model a human movement speed of 5km/h you would need almost one hour to walk on a 4096m x 4096m map from one side to the other. Now make your map 4096km x 4096km and you would need 1000 hours game time. Games are not about perfect reality and even less about long traveling times.

 

Therefor, the scale of a world is more an illusion and there are other more important factors for a game world. One of the most important features is:

Is the engine capable to render and simulate a game world which delivers the illusion of a large world?

 

In this sense and in my opinion, cryengine is better suited for this task than udk, even if udk have larger levels on the paper ! Take always a look at referenc games to evaluate a game.

 

The only problem I have with that and I could be wrong in saying this but if you rely so much on scaling wouldn't you lose detail in doing so. The closer to accurate scale the more realistic detail you can bring forth. Not exactly with the level but with character modeling etc? Again I am not sure if this is the case just seems like it would be true.


"And I looked, and behold a pale horse: and his name that sat on him was Death, and Hell followed with him"


#10 Ashaman73   Crossbones+   -  Reputation: 7492

Like
0Likes
Like

Posted 26 February 2014 - 02:55 AM


The closer to accurate scale the more realistic detail you can bring forth.

Even realistic games are often not really realistic at all blink.png

There are so many factory you need to consider, like visibility (too small), perspective (camera FOV of ~90 vs human FOV of almost 180 degree), thirdperson view (the angle of the camera let details/chacters appear smaller/larger), performance (render a football stadion with 20.000 seats or just 2.000 ). Best to take a 'realistic' game which would came close to your imagined game and analyse it. Walk through the level, take a really close look at the details. Compare it to the character.

 

Most often a single game do one thing really good (GTA: sense of large,living city) and this little thing is the reason, that they use a specialized game engine. And you will have a really hard time to copy this feature with an other engine. Engine are  often tailored around two or three main features and seldomly are omni-potent.. Therefor, choose your engine by analysing comparable games. Watch trailers, ingame video, play demos etc. to help you decide, which game engine would be more suitable to your vision. smile.png



#11 EthanRev6   GDNet+   -  Reputation: 204

Like
0Likes
Like

Posted 26 February 2014 - 04:09 AM

 


The closer to accurate scale the more realistic detail you can bring forth.

Even realistic games are often not really realistic at all blink.png

There are so many factory you need to consider, like visibility (too small), perspective (camera FOV of ~90 vs human FOV of almost 180 degree), thirdperson view (the angle of the camera let details/chacters appear smaller/larger), performance (render a football stadion with 20.000 seats or just 2.000 ). Best to take a 'realistic' game which would came close to your imagined game and analyse it. Walk through the level, take a really close look at the details. Compare it to the character.

 

Most often a single game do one thing really good (GTA: sense of large,living city) and this little thing is the reason, that they use a specialized game engine. And you will have a really hard time to copy this feature with an other engine. Engine are  often tailored around two or three main features and seldomly are omni-potent.. Therefor, choose your engine by analysing comparable games. Watch trailers, ingame video, play demos etc. to help you decide, which game engine would be more suitable to your vision. smile.png

 

So what you are saying.... Is We need to make our own game engine???? blink.png hahaha Challenge acceptedbiggrin.png


"And I looked, and behold a pale horse: and his name that sat on him was Death, and Hell followed with him"


#12 Ashaman73   Crossbones+   -  Reputation: 7492

Like
1Likes
Like

Posted 27 February 2014 - 12:42 AM


Is We need to make our own game engine???? blink.png hahaha Challenge acceptedbiggrin.png

No, I would only sugguest to evaluate different engines first, even mod-able games. There are even open source engines around (eg Ogre) which you could use as base. I would almost never sugguest to build your own engine nowadays, especially if you are just starting with game development.

 

Yes, I've rolled out my own engine rolleyes.gif , because I need a specialized engine too, but this engine is already 15 years in development and by far not as capable as the modern ones, especially if you consider tools too. If you target your own engine, you need to invest lot of time.



#13 creatures-of-gaia.com   Members   -  Reputation: 377

Like
1Likes
Like

Posted 27 February 2014 - 02:29 AM

btw, just my 2 cents, but designing/filling these huge areas and making them interesting is already a monstruous task. Exploring them would also already deliver days of gameplay. So before wanting something "bigger", perhaps you should start with it. ;) I'm sure the player wouldn't mind a map loading after running several hours accross the map. ;)


Edited by creatures-of-gaia.com, 27 February 2014 - 02:30 AM.


#14 EthanRev6   GDNet+   -  Reputation: 204

Like
0Likes
Like

Posted 27 February 2014 - 01:49 PM

 


Is We need to make our own game engine???? blink.png hahaha Challenge acceptedbiggrin.png

No, I would only sugguest to evaluate different engines first, even mod-able games. There are even open source engines around (eg Ogre) which you could use as base. I would almost never sugguest to build your own engine nowadays, especially if you are just starting with game development.

 

Yes, I've rolled out my own engine rolleyes.gif , because I need a specialized engine too, but this engine is already 15 years in development and by far not as capable as the modern ones, especially if you consider tools too. If you target your own engine, you need to invest lot of time.

 

Have you been working on your engine by yourself or do you have a team?


"And I looked, and behold a pale horse: and his name that sat on him was Death, and Hell followed with him"


#15 Ashaman73   Crossbones+   -  Reputation: 7492

Like
1Likes
Like

Posted 28 February 2014 - 12:49 AM


Have you been working on your engine by yourself or do you have a team?

The engine is written by myself and to be honest, it is stable for some years now, so you dont need 15 years. But nevertheless, a modern engine is very complicated (I'm talking about multicore, physics integration, shader, scripting, AI etc) and it will costs a few years to make it stable (I'm not talking about the first visual showoff after 2 weeks of work, I'm talking about the month and years of refining your engine). This is time lost to make actually a game. When I started the engine, there were really not a lot of options to use an existing engine, but nowadays there are hi-professional engines almost for free for indies out there.

 

If your goal is to invest only 2-3 years for a 'non-casual' PC game, then your decision comes down to either make a game using an existing engine or to make a engine with a simple demo.






Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS