Jump to content

  • Log In with Google      Sign In   
  • Create Account


Member Since 17 Apr 2006
Offline Last Active Jul 02 2013 11:04 AM

Topics I've Started

Distance from Origin Precision Issues & Localized Coordinates

27 September 2011 - 09:24 PM

Hey guys, this is my first time posting in any of the programming forums as I'm an environment artist by profession.

However, I'm extremely interested in tools development, content creation systems/methodologies, and making environment/level creation more effecient, which is why I hope to start my own project with the intent of utilizing a novel modular approach for creating high level of detail content on a large scale.

The biggest issue I know I'll need to address first is one of scale and handling floating point precision issues with having playable areas of the map at extreme distances from the origin. I imagine this issue is faced all the time with flight simulators so I thought anyone who has insight regarding how flight sims or other massive open-world games approach this problem would be worth seeking advice from. My project isn't a flight sim, though flight from point to point is a feature I'd like to support.

The problem is that I want the scale of what you would expect in a flight sim (say 40x40 miles) while still maintaining reasonable tolerances for character level collision, physics, lighting, etc. I fully expect having to subdivide the map into sections that can be pre-cached and streamed while maintaining surrounding sections as macro LOD's or meta imposters, but I'd like to load sections without having to break line of sight. Basically I don't want to have to force the player through a dog-leg, L-bend, or other contrived transition intended to break LOS for loading purposes. Additionally, even if I only have a small section of the level loaded where the player is currently active while the remaining sections are LOD'd if this section is 18 miles from the origin I still believe I'm going to have problems.

So it seems what I need is a way of localizing the coordinate system to the currently loaded or active section of the level. Does anyone know of any methods in general terms for accomplishing this or for dealing with extremely large open-world environments? For my project I was hoping to mod Cry Engine or Unreal either myself or with the aid of a project partner more versed in programming but I'm not sure if either of these engines would even provide the flexibility I need to accomplish this.

Any ideas or suggestions are greatly appreciated.

The Fate of Modern FPS Campaign Length

26 November 2009 - 12:23 PM

The Fate of Modern FPS Campaign Length As the expectations for improved graphics, audio, effects, AI, physics, etc. for FPS games continues to rise with each generation it becomes increasingly more difficult for developers to produce games at the scale and scope that we've seen in years past when pertaining to single player mode. If developers want to acquire positive marks from reviewers and players alike this often requires dedicating most of their resources on a smaller chunk of content compared to previous generations in order to reach the desired quality bar. Some would say that the emphasis is on quality over quantity. However, as an environment artist and a gamer who loves exploring, part of having a quality experience for me is having a large quantity of content to explore over a fairly large but manageable duration of time. It's difficult to think of a blockbuster FPS such as KillZone, Halo, or CoD that has had anything near a 20 hour campaign in the last 5 years. Half-Life 2, which came out just a little over 5 years ago was the last big FPS game to have a single player mode any where near this figure. There have been FPS games such as Dead Space with slightly longer campaigns but this is accomplished primarily through limiting the variety of the environments and content, and overly repeating gameplay elements. I'm not opposed to re-use of assets (see below) but I think it can be taken too far. ________________ WHAT ACTION, IF ANY, SHOULD DEVELOPERS TAKE TO IMPROVE THE DIMINISHING LENGTH AND SCALE OF SINGLE PLAYER CAMPAIGNS FOR FPS GAMES? A- Developers don't need to change their strategy. Modern FPS games have achieved an ideal balance between campaign length and overall quality. B- Developers should be willing to sacrifice a little bit on the graphics, sound, physics, and AI front in favor of more content and longer campaigns. I want a longer and larger campaign experience even if it means less focus on fine tuning content to the cutting edge. C- Developers need to focus on pushing the envelope for graphics, sound, physics, and AI even if it means even shorter campaigns and less overall content. I want the most engaging and realistic experience no matter how short and small it is. D- Developers should consider longer development cycles so that they can maintain a high level of content quality and increase campaign length and the amount of content/features available. I'm willing to wait another six months to a year for the game's release if it means I get a large and long campaign in addition to having excellent visual, audio, AI, and gameplay standards. E- Developers should adopt new development strategies such as relying more on asset re-use and modular systems. This allows developers to increase quality, quantity, and campaign length without over extending the development cycle, but at the cost of reducing the number of unique environments and assets. I don't mind playing through two or more levels that share similar visual styles and content if it adds more length and scale to my campaign experience. F- Other. I have another solution I'd like to suggest (please reply with explanation if voting for this). I couldn't seem to find a poll ability in the forums so if every visitor could please reply with their vote (A,B,C,D,E,or F) and perhaps a little justification I'll try to update the tally as often as possible. ____________________ I tend to lean toward solutions D and E. I believe developers need to be willing to accept the fact that games of this generation and next are going to take longer to develop if they desire to push the quality bar in visuals, audio, AI, physics, and gameplay while still maintaining the scale and amount of content delivered in previous generations. I also think that there are strategies for creating content, and environments specifically that can increase the amount of content and length experienced in an FPS campaign without drastically lengthening the development cycle. Rather than creating a custom palatte of shaders, bitmaps, effects, objects, architectural, and landscaping concepts to make each level completely unique it seems that establishing a few unique "environment schemes" or themes that can be used on multiple levels would be an effective approach. This strategy has been used in the past and it seems negligent that so many FSP developers have completely abandoned it. I think FPS games could adopt this strategy as exemplified in games from other genres such as adventure games, action games, and platformers. Take SMB3 for example. It has eight worlds, each themed (an environment scheme) so that the levels within each world are unified by similar elements though feel unique as a result of their custom compositions and gameplay challenges. I could imagine in an FPS game that you could establish 4-6 different environment schemes or themes that could be used 2-4 times each giving a range of 8-24 different levels. If each level is about an hour this seems like an ideal range in terms of campaign length. Having 16 levels based on 5 different themes would seemingly be easier from an art production standpoint than 10 levels, each having there own unique environment scheme. Thoughts?

Secret Areas: How much is too much?

12 November 2009 - 11:27 AM

As an environment artist and aspiring level designer one of my favorite aspects of games is exploration and discovering areas of a level that are off the beaten path. I was curious about how many players share my enthusiasm for environment exploration in games and discovering "secret" areas that are not integral to the core game experience. In a level-based FPS or action game for example would it seem overkill to have as much as 40% of the environment hidden from the main sequence of the level? Personally, I think it'd be rewarding to discover such a space and really add to the replayability of the level. It seems silly to complain about getting something for free that you weren't expecting but I can see some people thinking, "gee, why wasn't this actually part of the main level experience?" Thoughts?

How to combine two game types

29 January 2007 - 09:13 AM

Lately I’ve been brainstorming a game concept that allows players to essentially create and control life on a planet(s) in a small region of a space from the birth of the stars in the player’s galaxy all the way to the rising of technologically advanced life capable of colonizing other planets. Initially, the player is responsible for the macro management of their creation but as the game progresses the player’s world becomes more complex requiring a micromanagement approach for controlling their creation and its inhabitants. Basically, the game I have in mind is a sandbox sim with RTS elements. My concern is this. Will it seem unnatural to the player to start a game that is played out very similar to a sim, but as events occur and the game progresses they begin to acquire more abilities that would be expected of an RTS game? Fortunately, this transition would be gradual and allow the player time to learn how to cope with new situations and skills, but it seems that it might discourage some players who expect the game to play out entirely as a sim or entirely as an RTS. One solution I was thinking of is that the player could choose what phase or time period of the game they start in, and the game will randomly produce or provide a preset scenario for players who don’t want to start from the very beginning of their creation. The game is divided into seven epochs, each with their own passage of time. Each epoch or period has its own set of objectives and grants the player more abilities with every transition. Any thoughts or idea?

Realistic Evolutionary Sim Concept

21 January 2007 - 08:00 AM

Hey everyone, I’ve been reading a lot of astrobiology books lately such as Rare Earth and The Privileged Planet and in the process have been inspired to design a game built upon the concept of controlling the destiny of a habitable planet in a Sim/RTS game. I already know that Spore plans on doing something similar. However, instead of only concentrating on the evolutionary development of a single species at a micromanagement level I thought it’d be interesting to direct the growth of an entire biosphere and its inhabitants on a macro level. Additionally, I’m thinking of a game that is more scientifically accurate than previous god-mode games. So far I’ve envisioned the game to be divided into two primary roles or phases of gameplay. At first I want the player to focus on the requirements for making their planet inhabitable and building up their biosphere, but as soon as an intelligent species arises on their planet I want gameplay to begin focusing on the civilization growth of the intelligent species(s) occupying the planet. I’m not quite sure how this transition can be made yet at this point, so I was wondering if the collective creative minds here might have some ideas. The other inherent problem with this transition is the passage of time. In the beginning of the game the player is going to be acting on geological timescales, such that a single second of gameplay might represent a hundred or a thousand years, whereas when the player takes control over their intelligent specie(s) the time scale is going to need to change drastically so that a second of gameplay might represent a month or a week instead. I think that if the first transition can be figured out that coming up with a solution for the timescale difference shouldn’t be too difficult. I'll try to provide some examples of player decisions in a future post. This is already long enough as it is and I don't want to deter readers.