What's left for the Source engine?

Originally posted by StickFigs
Dynamic Wood....

Doom III engine sucks! Not enough interactivity!
Click to expand...

which you know! because you've played it!

oops, that was an alpha.

let's see... dynamic lighting is interactive. it also has a physics engine, you know :\
 
Stick Figs is kinda slow. You'll have to forgive his ignorance. There are a lot of trolls on this board...
 
Originally posted by Quotidian---
here is my dream for the future of FPS..


a realistic damage system.
my explanation: Each individual model has simulated vital organs with in it. These organs would move with the model and by at the same locations you'd find in yourself.

thus each player would have a simulated heart. major arteries (spelling) liver, lungs, brain.

then each bullet trajectory would always be determined (this can have otehr possitives but i'm only describing the effects of damage on players)
So when each bullet entered the body, it wouldnt' be via hit points, it would simply pass through the MASS of the player, and would continue on its way, thats basically how interactions would happen with all bullets.. then the damage this foriegn object would do would be determined based on
A) what just hit your mass (FMJ hollow points, etc.)
B) blunt damage (the intial hit of the bullet, such as if a vest stopped the bullet from actually penetrating your mass, blunt damage would still be cuased, maybe this caused your leg to fly backwords, you to lose balance, get shookin up, loose breathe)
C) all the physics of the bullet before it entered.. as in trajectory angle, speed, etc.. so that if the bullet could penetrate you would be determined.. and the angle of which it would go through you would be determined..
D) then what if any major organs where hit.. since they are simulated it would be easy to tell what part of them where hit and if they were rendered completely useless or just damaged.
E) based on the damage to organs and body mass.. the 'damage' to yourself would be applied..

so say your liver is torn to shreds, and your unable to walk, you would be put to the ground, allowd to continue to aim/shoot/look around.. but not get up.. and you'd have probably 20 seconds to live before your screen would fade black.

say your heart gets taken out.. and say the force wasn't enough to effect you blunt wise, but your heart is deffiantly gone.. you would have one or two steps before you hit the ground unable to move/shoot/look.. but a few seconds of laying there before screen goes black.

and on and on..

this would also make for some incredible medical use.

etc. etc.
Click to expand...

The original Deus Ex had something along those lines.
 
kelisis, I dont know if its just being tired thats gettin to me but i laughed when i read your post, out loud, which i dont usually do while reading. your funny. :cheers:
 
BlumenKohl: Deus Ex didnt have those features. I mean, in Deus Ex you could lose limbs and what not, but in regards to bloodloss, i mean come on, You could play most of the game with no legs and 1 arm and you wouldnt loss an ounce of health.

Although i really love that game. One of the best, without a doubt
 
Have none of you played Red faction? it had deformable terrain/objects.

Aswell as the upcoming Soldner:
http://soldner.jowood.com

Cloth/hair can be done with today's technology too, it's just no game really wants to push it.

Blade of darkness did real time lighting and dismemberment good too.
 
fluid mechanics is so complex that I suspect that it will never acuratly be portrayed in a computer game. ever. What I hope to see is textures replaced by extreemly high polygon count enviroments.
 
FYI: The finale of gaming engines is one which supports atom creation (and splitting, just for fun).
 
I doubt engines will ever go that far. you could make something look photorealistic without all that atom stuff. such a simulation would be used for physics problems, rather than games. it would just be unnessisary
 
Heh, looking back, I should have named this thread "What's NEXT for the Source Engine" not "What's left" :D

Whoops! :)
 
Last edited:
Re: Re: What's left for the Source engine?

Originally posted by Robson
id is really *paying* for those shadows, though. I've heard that as much as 50% of their CPU time is being spent on shadows. That's enormous.

And even then, it's not *really* unified -- it's still a combination of projected (texture-based) shadows and stencil shadows. That was the part that bummed me out. When you see nice, smooth, soft shadows, those are projected; when you see unnaturally crisp shadows, those are stencil. As beautiful as Carmack's new system is, it's still not as "unified" as it could be.

When all lights are represented by emissive geometry (no point lights, no spot lights), when the shadow softness is derived from real penumbra calculations, and when we start integrating radiosity-mimicking calculations in real-time... *then* we'll be approaching a true lighting unification.

:naughty:
Click to expand...

Yeah, um. What he said. Penumbra...that's the ticket. I'm guessing you have a modelling background?
 
Couple things.

There is absolutely, positively, NO end in sight for what games will use and how complicated they can get. Photorealism is nothing compared to REALISM, which, believe it or not, has a name. It's called "floating point color" and it's the future of monitors and bit-depth as we know it. It's beyond photorealism, because it's realism.

So as far as how things look, floating point will be a definite end all, however it will require pixel (atom) precision to look real and I seriously doubt anybody will say "ok, we're done, it can't get any better" because it will get better and better, until we've reached the holodeck from star trek the next generation. THAT my friends is the ultimate end-all of everything.

And yes, we are working towards such possibilities. And yes, computers will be able to do all of this in real-time without breaking a sweat (or a circuit!). Never say never when it comes to technology man!
 
well I think that next we will see some form of free fluid dymanics. free meaning contairless. this dosn't mean perfect though just something neat. Neatness factor is big and so will probably be the ticket to getting stuff in game. In a few years this far from perfect fluid dynamics will be at teh level our solid object dynamics are at now
 
Originally posted by Stiler
Have none of you played Red faction? it had deformable terrain/objects.

Click to expand...

Yeah, in a ghetto sorta way it had deformable terrain. Wasn't the least bit realistic however.
 
I want battlefield 1942's entire game mixed with HL2's physics

*drools*
 
Originally posted by Homer
fluid mechanics is so complex that I suspect that it will never acuratly be portrayed in a computer game. ever. What I hope to see is textures replaced by extreemly high polygon count enviroments.
Click to expand...

For relief in brick walls you can better just use normal maps, unless you have rendering power to spare.

For characters, about 50K polygons combines with normal maps are enough. For really complex spacesuits (like in the Quake 4 renderings) 300K poly's.
 
That realistic AI stuff sounds cool. When a character behaves like a real person (can understand speach, and learns etc.). Wow!

Funny though, because you'll have created a real conciousness in the process (one which you cannot tell apart from a real person just by talking to it). Therefore you'll become a mass murderer just by switching the game off :frown:
 
Well everyone time you do something in the real world, like taking a step down the street you are probably killing a hundred thousand microorganisms... so who cares :)
 
actually one thing i think Doom3 Engine has that Source doesnt (maybe) which would be Per Polygon Hit Detection...

which is very cool! because you dont have hitboxs and you can shoot someone between there legs and under there arms... and miss...

now thats pretty cool... no more hit box rubbish...

means people will have to learn to aim properly!


TheRook
 
Lol, CS has given hitboxes a bad name. If hit boxes are close enough to the model, then it's as good as per-poly hit detection. The problem with CS and the lack of accuracy is all down the crap way they handle recoil and the way bullets spray rather than inaccurate hitboxes.

Anyways, the problem with per-poly (especially in multiplayer) is that if people are using different models then it's not fair because a smaller character has a lower amount of target area than a smaller player. It also takes more server time and network bandwidth to do right.
 
Originally posted by TheRook
actually one thing i think Doom3 Engine has that Source doesnt (maybe) which would be Per Polygon Hit Detection...

which is very cool! because you dont have hitboxs and you can shoot someone between there legs and under there arms... and miss...

now thats pretty cool... no more hit box rubbish...

means people will have to learn to aim properly!


TheRook
Click to expand...

Actually, HL2 does have per polygon hit detection. It's in the "valve info only" thread somewhere.
 
Originally posted by smilez
Well everyone time you do something in the real world, like taking a step down the street you are probably killing a hundred thousand microorganisms... so who cares :)
Click to expand...

Micro-organisms generally do not care if they are alive or dead.

If in game characters become believable as humans (ie. you can converse with them, hurt them etc.) to the point where you would not be able to tell that they were virtual (unless told so), then IMO they are as real as we are. It would be wrong to kill them.
 
Originally posted by MrD
If in game characters become believable as humans (ie. you can converse with them, hurt them etc.) to the point where you would not be able to tell that they were virtual (unless told so), then IMO they are as real as we are. It would be wrong to kill them.
Click to expand...
That raises some interesting ethical questions: if they act like a normal human being and are sentient, do they have rights like normal human beings have? Would they eventually realise that they're in a videogame , would they consider us as their gods since we created them?

Weird.
 
I do not see how they can get away from the modeler/map maker having to make breakable points or pre determined points.
They can make it more realistic to play by adding more points and allowing just that point to break where you shoot for strong material (wood/metal/brick) or it all to fall apart if its weak or brittle (glass).
Once it is broke, the original peace is replaced by 2 peaces and textures that are asigned. How can remove those boundries?
 
I'm sorry, but that last one is completely faked photoshop trickery. A masterful trick at that, but I do see inconsistencies between the two images. If you look cloe enough, you'll see them too. I already know how he faked it all.

As for AI, what if the AI got so smart, it escapes to the internet to avoid being turned off. Would we end up with a Terminator or Matrix like situation for real?
 
Originally posted by MrD
Micro-organisms generally do not care if they are alive or dead.
Click to expand...

(One might argue that micro-organisms "care" in a very low-level way, in that all living creatures will fight for their survival. But then we'd be getting really off-topic.)

:rolling:
 
Originally posted by MadsMan
As for AI, what if the AI got so smart, it escapes to the internet to avoid being turned off. Would we end up with a Terminator or Matrix like situation for real?
Click to expand...
Unless you were retarded you would limit the game to only allow the character to able to interact with the game world... unless you made the game world able to interact with the real world they would not be able to do any harm.

The difference when compared to SkyNet in the Terminator movies is that SkyNet was designed to be able to directly interact with other computers and gain control of them.

The difference when compared to the Matrix would be that the character does not exist in the real world; thus it could not "wake up" and escape the game world... and in the Matrix the clients can "bend" the rules (like the old speed hack).

Even if the character somehow realized that the game world was not the real world (you could try explaining it to them, but they probably wouldn't believe you) he/she/it could do nothing about it.

Though it would be interesting, if/when games get this advanced, to let a character take control of a real robot by giving them a virtual computer... but you would need to make sure the robot isn't strong enough to do any harm.
 
Hey, there are a lot of retarded people out there. But seriously, what's to stop someone from ULing a sentient program to a server, or creating one of thier own. There are just infinite possibilities. And my point about "The matrix" was more of "in the beginning" ya know when programs became sentient, and began to spread like a virus, taking over all manner of computers and electronics. Little ones and zeros that had purpose of its own. And as someone else mentioned, all lifeforms pretty much strive to survive, hence evolution. So why wouldn't a sentient program strive to survive as well, possibly evolving as it learned? It's not crazy to assume that a sentient program could break out of its coded confines. Any program can be hacked, afterall. Makes you wonder....
 
Lets make a real skynet mod for HL2, it will really escape the game and communicate and take over other computers. It will be true AI which learns to operate machinery, and overtake the human race. Who wants in?
 
I know people always chant "Never say never!" when talking about technology, but I do not think we will ever be able to create self-aware, sentient AI. Self-awareness just is, I don't think it's something that can be created or taught. While artificial intelligence will undoubtedly become very advanced in the future, it will never become self-aware. And if it were possible to create, it would be counter-productive (as a self-aware rescue-bot may be unwilling to risk its life to fulfill its duties) or, at worse, dangerous.

As for those renderings, woah! Maybe I'm just twisted in some way, but every time I see something like that, I always wonder who will be the first to use such technology to create an, ahem, adult oriented game.
 
For AI to be that complex enough to bring up ethical questions, it has to be intelligence with the ability to learn, not just be programmed. Otherwise it would be like a mosquito. You don't see them planning world domination do you?

I know of AI programs that people can actually teach things to, but their learning skills are severely weak and limited. It has to be able to draw conclusions and adapt to situtations, without being programmed. Skynet is far far away.

If it can't do that, then it's no threat, because if it can't adapt or draw conclusions with limited information, it could be eliminated pretty easily. Where areas, something with the ability to adapt and draw conclusions could see danger (made by us to stop it) coming and adapt to it asap.
 
Brando holy poop those are awesome images, if you can find any more, please please post them! Hilarious that they even fooled MadsMan, but all you have to do is look at the hair to automatically tell it's really CG and not a picture of someone.

Man just imagine games with those graphics. *uber-drool*
 
Originally posted by Mountain Man
I know people always chant "Never say never!" when talking about technology, but I do not think we will ever be able to create self-aware, sentient AI. Self-awareness just is, I don't think it's something that can be created or taught. While artificial intelligence will undoubtedly become very advanced in the future, it will never become self-aware. And if it were possible to create, it would be counter-productive (as a self-aware rescue-bot may be unwilling to risk its life to fulfill its duties) or, at worse, dangerous.

As for those renderings, woah! Maybe I'm just twisted in some way, but every time I see something like that, I always wonder who will be the first to use such technology to create an, ahem, adult oriented game.
Click to expand...

We are machines in a way, we use chemical reactions instead of electric, but still we gain self awareness with just physical processes. Then why can't a computer program have self awareness?
 
Originally posted by DaveKap
Brando holy poop those are awesome images, if you can find any more, please please post them! Hilarious that they even fooled MadsMan, but all you have to do is look at the hair to automatically tell it's really CG and not a picture of someone.

Man just imagine games with those graphics. *uber-drool*
Click to expand...

Uhh, no I noticed the hair was 3d. That's not my point at all. Ever heard of compositing?
 
Originally posted by BlumenKohl
For AI to be that complex enough to bring up ethical questions, it has to be intelligence with the ability to learn, not just be programmed. Otherwise it would be like a mosquito. You don't see them planning world domination do you?
Click to expand...

A mosquito can most definitely learn, just not on the level you appear to be referencing. An article I read in Scientific American gave a projected timeline for AI in the world, and it's end was 50+ years off. The relevance? The article suggested that creating an AI with the capabilities of a fruit fly is around 5-10 years off. Just because something is a tiny, minute, organizm, does not mean it is not without computational complicities.

Originally posted by BlumenKohl
I know of AI programs that people can actually teach things to, but their learning skills are severely weak and limited. It has to be able to draw conclusions and adapt to situtations, without being programmed. Skynet is far far away.
Click to expand...

I really hope your not talking about programs like Dragon Speak. Yes, there are several programs, in several fields and applications, that can "learn". But this is a scripted, programmed, learn. In other words, the code would be a scripted, albiet monstrous, sting of variables:

  • Food Preference

    1. If user asks for "Chocolate", write preference, time, and amount to database and average with previous results.

    2. Next time user mentions food keyword, corrolate data and suggest common preference.

The key to "real" artificial intellegence is reason. This is the difference between a robot saying, "Master prefers to eat chocolate in the evening." and "Master likes to eat chocolate in the evening, it seems to make him happy." (Without scripted events)

The current group of top scientists and contributors project we may see this level of intellegence within 50 years, but it may be longer.
 
I have no doubts that we will get there. Like someone earlier posted, we are nothing more than chemical machines. Our thoughts, and impulses are all electrical signals processed by the brain. Humans are sentient, most likely because we have bigger brains than any other animals, and we use a much higher percentage of our brains than other animals do(relatively speaking). But alas, humans still use a very small percentage of thier brains, and the potential there is enormous.
 
Originally posted by MadsMan
Uhh, no I noticed the hair was 3d. That's not my point at all. Ever heard of compositing?
Click to expand...

Lol, dude. Did you even look at my reply...you know, the one with the tutorial? It's six pages of a detailed texture tutorial for believable skin, here's the end result:

http://www.3dm-mc.com/tutorials/maya/texturing/069.jpg

This guy is well known in the 3d community, it's all legit. Perhaps you should do a little research before calling someone a fraud.

Note: He does use Photoshop to adjust hue, balance, etc. (It refers to that at the bottom of the picture in question.)
 
Yes, I saw that. And the results on that image are obviously CG in origin. The other image(of the young boy), has too many faults that don't line up. I have seen this kind of trick before, and it's most likely a model that he cannot turn 360, otherwise the illusion is lost. I do believe the eyes and hair are 3d, but the rest would not stand up to scrutiny. And why does his shoulders look blocky in the render, and nice and smoothed out in the viewport. just doesn't make sense. Little things like that. Also the shaded version of the model has some inconsistencies in lighting compared to the full render. As if the full render has a photo-map projected over the geometry, thereby inheriting it's lighting. And if you look in the shaded viewport, you will also see that the backgroundis pre-lit, which could also support what I am saying.... But hell, I could be wrong, but my senses tell me that I'm not. sorry if his upsets you. :/
 
You must log in or register to reply here.

Similar threads

CombineAdvisor3
Source 2 expectations?
Replies
1
Views
3K
Clan Wolf 2
Clan Wolf 2
MFL
Fascinating look at the ongoing psychological evolution of flat hieirarchies.
Replies
2
Views
2K
Tollbooth Willie
Tollbooth Willie
B
From Warren Spector to Arkane Studios: Half-Life 2's cancelled Ravenholm episode
Replies
2
Views
10K
Clan Wolf 2
Clan Wolf 2
B
Who's left or joined Valve in recent history
Replies
4
Views
32K
MFL
MFL
B
The other cancelled ports of Half-Life
Replies
3
Views
7K
Tollbooth Willie
Tollbooth Willie
Back
Top