• Advertisement
Sign in to follow this  

How video game level architecture fundamentally differs from reality

This topic is 1263 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Advertisement


I want instead to take the chance and discuss these design decisions in the context of immersion. Eventually this might help us creating more immersive themes and levels for video games.

 

for me, level boundaries and level linearity give me a "rat in a maze feeling" which breaks immersion. almost all shooters seem to have this feel to them.

 

performance oriented level architecture is only an issue when the resulting level design seems "out of place" in the context of the game.   IE the level design limitations required for high performance have a noticeable effect on gameplay, look and feel, and/or believability.  A number of studios have successfully designed a game around a given performance oriented level design methodology - IE they start with the performance method and then figure out what type of game it can be used for with a high degree of believe-ability.   "out of place" levels most likely occur when the engine is attempting to render some environment other than that for which it is designed - or must render some less believe-able scene in lieu of what ought to be shown, due to engine performance related level design limitations.

 

"within walking distance".     oblivion and skyrim both suffer from this to some extent. very full worlds, but not very large. not too much of an immersion breaker. your "where's the windmill" example is a definite immersion breaker.   i've had to deal with this issue in Caveman 3.0 as well: is the frequency distribution of everything (resources, encounters, etc) correct?  in some cases i can extrapolate from real world data (location and frequency of caves and rockshelters). In other cases i must "guess-timate".

 

a related issue that's somewhat of an immersion breaker for me is "towns" with a population of about half a dozen NPCs, or "cities" with <100 population. IE unrealistically low numbers of npc's for the supposed size of a settlement. its is usually caused by limitations of the graphics engine (cant draw too many NPCs onscreen at once) or lack of content (it would take us too long to create the content for a skyrim city with 1000 NPC's, so we'll just do 50 or so).

 

this trend can also be seen in things like two dozen trees in close proximity representing a forest. by contrast, in Caveman 3.0, the minimum size forest is 5 miles across and contains 1,415,700 tree models.  needless to say, outdoor scenes in Caveman 3.0 are not level based.

Share this post


Link to post
Share on other sites

i don't see user created content as the solution to the content creation bottleneck.

 

whats required is a turn key integrated solution that encompasses integrated graphics (including animation), audio, physics, damage, and behavior modeling. such a system would include libraries of stock graphics, animations, and audio, and would support multiple physics, damage, and behavior models. content could be customized and shared. the system would also be customizable and extensible to add custom behaviors, etc.

 

then we could get away from building objects and get on with building worlds.  or at least spend less time on objects and more on worlds.

Share this post


Link to post
Share on other sites

i don't see user created content as the solution to the content creation bottleneck.

 

whats required is a turn key integrated solution that encompasses integrated graphics (including animation), audio, physics, damage, and behavior modeling. such a system would include libraries of stock graphics, animations, and audio, and would support multiple physics, damage, and behavior models. content could be customized and shared. the system would also be customizable and extensible to add custom behaviors, etc.

 

then we could get away from building objects and get on with building worlds.  or at least spend less time on objects and more on worlds.

 

 

You name a number of things that would have to be achieved anyway WITH Player Created Content,  so why not add the ability to tap into players creativity/effort as well ?   Idiot-proofing of tools would be good, but isnt required (just fewer people would be able to effectively use them)

 

If you are going to get the game industry to cooperate in that way, then why not go all the way?  (Players who will do alot and not for moneys sake)

 

You also missed what I said.   Its not just 'objects', its EVERYTHING that makes up a game ('worlds'). that would be accessible to create.  The Game Engine Components likewise would have to be built up to support this entire system.

 

"Missions" are too few in most games ? 

 - Decide what you want (and how it fits in to 'the game').  Specifications may come from a larger 'game project'.

 - Build a scene up (templated styles of terrain for the particular game genre/setting) out of available building blocks (customize a few critical ones and integrate them)

 - Place lots of available (largely useable) props to customize that space further (an Important Tool is the search system to just find/review the ones you want)

 - Include 'stock' NPCs who have generic behaviors ( flavor THEM by adjusting their templated attributes, setting their goals/activities)

 - Select and craft 'special' plot objects -- key-objects, key-NPCs, effects   (mod alot of existing stuff)

 - Create Plot - triggers/goals/special dialogs with contingencies/action sequences/exit-conditions/etc..  (flowcharty stuff)

 - Cutscenes choreographed by Tool that is directed like a Theater Play Script (action behavior scripts preexisting...)

 - Test it (actually run it on local simulator).

 - Fix things, adjust things, tweak things. Repeat.

 - Submit it to the game's community for critique (they run/ 'test it'), commentary, rejection/acceptance

 - Adjust (if needed) and resubmit

 - Publish to players.

 

Omitted above is opportunities to collaborate bewteen multiple 'developers' (a whole slew of posting/request/checklist tools/process) to allow people to do piecewise the elements they do best/easiest and pass it on.

 

---

 

The 'value-added' thing that game companies would then  be selling will be their ability to skillfully/effectively Tell the Story / Stage the Action / Build their Vision --- less alot of the mundane grunt work that limits them so much now.

 

 

Unfortunately game companies are invested in the way they do business now, and wont want to share ANYTHING they produced.  SO it may take an independant effort to get this all started and once it is better proven, then the companies will come begging and THEN they should only be allowed to make use of this new system IF they throw in ALL their assets they previously have built as seed data into the common community.

 

.

Edited by wodinoneeye

Share this post


Link to post
Share on other sites

We have everything you want already.

 

It's called VISUAL STUDIO.

 

What you are asking for is not a game, it's not even a game engine. To do what you want FOR EACH GAME, boot up visual studio and start from scratch.

 

 

I would like to start up an application, set a few flags, hit run and go home.

 

Next morning I would into the office and have a fully functioning game.

 

Technically it's possible, using procedural generation and genetic algorithms, but I'm not holding my breath. Writing a visual system for evaluating a generated generic 3d mesh is probably measured in man lifetimes rather than man days.

 

Have you ever written a game? 

Share this post


Link to post
Share on other sites

We have everything you want already.

It's called VISUAL STUDIO.

What you are asking for is not a game, it's not even a game engine. To do what you want FOR EACH GAME, boot up visual studio and start from scratch.

I would like to start up an application, set a few flags, hit run and go home.

Next morning I would into the office and have a fully functioning game.

Technically it's possible, using procedural generation and genetic algorithms, but I'm not holding my breath. Writing a visual system for evaluating a generated generic 3d mesh is probably measured in man lifetimes rather than man days.

Have you ever written a game? 

 

 

Wrong thread or forum ?

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement