Half-Life 3 - analysis and expectations

Which technologies have the highest chance to happen in HL3?

  • Advanced facial animation with meta-programming

    Votes: 9 90.0%
  • Semi-static area lighting with prebaked radiosity paths

    Votes: 7 70.0%
  • Physically correct material system

    Votes: 6 60.0%
  • Advanced particles / weather system / GPU particles

    Votes: 6 60.0%
  • Persistent open world MMOFPS

    Votes: 1 10.0%
  • Something else

    Votes: 3 30.0%
  • Nothing of the above

    Votes: 1 10.0%

  • Total voters
    10
Joined
Jul 10, 2013
Messages
52
Reaction score
20
Hi there!
No doubts, we all are excited waiting for the next installment of our beloved game universe, and taking wild guesses on how awesome it will turn out (or not). Grounded speculation is one of my favorite topics, so I've compiled my personal list of ideas and features that I believe will find their way into the final game. Please feel free to agree or object, if you like.


2014

The year 2014 makes a lot of sense for HL3 release date. First of all, its 10th anniversary of HL2, and this decade leap could make a solid foundation for story progression - a lot can happen in 10 years in the world dominated by oppressive extraterrestrial forces (assuming 1:1 time flow, which is rather wild speculation on my part). Secondly, according to recent Jira-related leaks, Source 2 and HL3 are deep into development, with Source 2 featuring dedicated 'Artists' and 'Characters' groups amongst many others, and 'srs2sdk_...' groups as well. And finally, 2014 is the year of massive adoption of next-gen consoles and projected year of SteamBox release.


Source 2 and HL3 killer features

2004's Source Engine was built around dozens of innovative technologies of its time, paving ground to unprecedented player's immersion via solid body physics and rich character / facial animation. As we remember, these technologies influenced HL2 game mechanics a lot - we had various physics puzzles, a physics-controlling gun, and interactive story revealed through exposition from other characters and in-world events. These are common things in games today, so the question rather is - what can we expect tomorrow?

Before proceeding, let's not forget our target specs, so any further speculations could have at least some sort of credibility. The average semi-gaming rig / next-gen console of 2014 will have: multicore x64 CPU, 8 GB of RAM and DX10+ GPU with some sort of CUDA / DirectCompute / OpenCL available. What groundbraking technologies would be able to take advantage of that level of hardware?

  1. The first thing I would bet my money on is an insanely complicated facial animation system featuring meta-programmable VM for mimic processing. Recent works of Quantic Dream in this field ('Kara', 'Hard Rain', 'Beyond Two Souls', 'The Dark Sorcerer') inspire a lot of optimism for the feature of immersive game characters, and I just can't imagine Valve missing such an opportunity to leverage emotional impact of their upcoming games. But to remain characters highly dynamic and interactive, engineers will have to incorporate some sort of hyper-language for 'emotional processing', and that will potentially lead to game characters being complex identity matrices, expressing their emotions depending on player's input and game context.
    Season_4_-_Walt.jpg
  2. The second thing, I would imagine being an integral part of Source 2, is the paradigm shift in lighting model. Valve were never crazy about fully-destructable environments, but when it comes to FPSs, strong and moody cinematic lighting, coupled with excellent environment design, was always high on their priority list. In 2004, that resulted in using static prerendered lightmaps, which were pretty much the only valid choice for quality lighting at that time. But tomorrow, hardware will be much more capable of. Out of all the technologies available for tomorrow, I would bet on semi-static lighting with prebaked radiosity paths. Given the large portion of environment is static, such a lighting could be easily recalculated on the fly when it's necessary, at relatively small computational costs required. But advantages that it brings are tremendous - area lights of any shapes, naturally looking soft shadows from dynamic objects, soft self-shadowing, color bleeding and much more. More important, in case of not-so-good hardware it gracefully defaults to deferred compositional lighting, which is largely popular in games today.
    40795_1249908545_submedium.jpg
  3. And finally, the key thing. Havok physics might be the killer feature of HL2, inspiring Gravity Gun, but HL3 can't rely on it anymore. Technology, that makes or breaks HL3, should lie in dimension of general GPU programming (CUDA, etc...), since that is the only thing that sets apart the hardware of tomorrow from the hardware of yesterday, besides the obvious GFLOPS and GB leap. What such a key thing could be? Personally, I'm still struggling with this one. One possible thing could be physically correct material system, organically blended into sci-fi universe. The other thing, I have high hopes for, is advanced particles / weather system - it could rely on GPU concurrency a lot, and may introduce some new gameplay concepts, like nanites programming, weather control, turbulence gun, solid-liquid-gas state change, etc. May be the key thing will not be hardware related at all - it may turn out to be somewhat of a much larger scope: for example, HL3 may happen to be persistent open world MMOFPS with shared singleplayer experience against god-like AI system, or even permanent large-scale multiplayer battle of Resistance and Combat. Feel free to take a wild guess and voice your opinions.
    2iusrs.jpg
 
lol'd at the Walter White picture
 
Back
Top