Source 2 Benefits?

Sidedraft

Party Escort Bot
Joined
Feb 8, 2014
Messages
263
Reaction score
67
So I've been thinking about all the great and modern games that still run off the original source engine (albeit updated and patched) and I started thinking: "What will source 2 engine bring to the table?" Source 1 seems to be (for the most part) a jack of all trades. And the only reason I'm actually making a whole thread about it is....because I literally can't think of much it could possibly have to offer and was wondering if any of you guys could. I mean, It has to have stuff to offer right?

Improved graphics is something really easy to think of, but they probably will be good, nothing jaw dropping amazing-and valve will slowly update as they need for new games and probably get a similar lifetime out of the new engine. Like they did with source 1 doing games from half life series all the way to portal 2 and csgo.

Can anyone think of some other benefits? because my brains pooped today.
 
Last edited:
The most obvious thing to say right now, given how most of these features are already available, would be more accessible content creation. Source has always had a bunch of problems with file compatibility, not to mention how it continually forces developers to create content with occasionally restrictive rule sets or criteria.

Being able to quickly and easily import a texture, model, sound, or particle effect (barring sound - all of which are possible in the currently released version of the Dota 2 Workshop Tools) will be huge for productivity not only for Valve, but for third party and indie developers.
 
Greater compatibility with different operating systems (Valve has been talking about Linux a lot.)?
Oculus/VR integration?
The Steam Dev Days mentioned better UGC support.
Some sort of built in integration with SteamOS / the steam controller?
Half-Life 3? (Too hopeful?)

And as for graphical fidelity, I'm starting to see that plateau more and more. There is only so much we can do with current technologies, and the human eye can only see so much. Besides, something could have the best graphics in the world and still look ugly because of poor/broken/stupid aesthetics.
 
Look at Dota 2 map creating process and think about all the possibilities Source 2 can provide for first person shooters and alikes.

Imagine, you have a standard HL2 asset - a wire fence block. And you want to produce its derivatives - a torn fence, a broken fence, a burned fence, etc., probably with combinations of those. Instead of remodelling this asset each time you want a new option (which is very time consuming), you could literally use something like a modifier brush and paint all this stuff on a standard asset, not only altering its texture but geometry itself, adding tearing to wires and what not.

Awesome? We need to go deeper. Now imagine that all these geometry-level modifications can be applied dynamically, which means that fences could bend when explosion happens nearby, or tear, when you launch rockets to them. Not only that, but all the static modifications during the map creation are non-destructive, which means if you don't like the way your asset modification turned out, you will just have to repaint some influence mask somewhere, without touching the original assets.

This is pure DX11 approach to world building. But we could take it even one step further. Imagine you have a bunch of clean textured assets and you want to build a heavily weathered map out from them. In this situation you would probably have to retexture all them, alter some models, bending metals, weathering concrete, breaking brick walls here and there. But what if you could just run a script, that simulates winds, humidity and overall climate changes over time and creates a few global low-res influence masks (like lightmaps) altering almost every object in the scene based on its position, exposure to winds and rains, etc. And then you add final touches by hand, retouching those masks, adding tears and bends via modifiers, etc. And remember, all this is non-destructive, which means you can revisit your map any time later and for example remove some modifiers if you don't like them, without breaking anything else.

In my opinion, this is what Gabe was talking about when mentioning a simplified content creation.

Figure 01. Alyx's jean shorts heavily torn with Source 2 modifier brush
newShading.jpg
 
Last edited:
Look at Dota 2 map creating process and think about all the possibilities Source 2 can provide for first person shooters and alikes.

Imagine, you have a standard HL2 asset - a wire fence block. And you want to produce its derivatives - a torn fence, a broken fence, a burned fence, etc., probably with combinations of those. Instead of remodelling this asset each time you want a new option (which is very time consuming), you could literally use something like a modifier brush and paint all this stuff on a standard asset, not only altering its texture but geometry itself, adding tearing to wires and what not.

Awesome? We need to go deeper. Now imagine that all these geometry-level modifications can be applied dynamically, which means that fences could bend when explosion happens nearby, or tear, when you launch rockets to them. Not only that, but all the static modifications during the map creation are non-destructive, which means if you don't like the way your asset modification turned out, you will just have to repaint some influence mask somewhere, without touching the original assets.

This is pure DX11 approach to world building. But we could take it even one step further. Imagine you have a bunch of clean textured assets and you want to build a heavily weathered map out from them. In this situation you would probably have to retexture all them, alter some models, bending metals, weathering concrete, breaking brick walls here and there. But what if you could just run a script, that simulates winds, humidity and overall climate changes over time and creates a few global low-res influence masks (like lightmaps) altering almost every object in the scene based on its position, exposure to winds and rains, etc. And then you add final touches by hand, retouching those masks, adding tears and bends via modifiers, etc. And remember, all this is non-destructive, which means you can revisit your map any time later and for example remove some modifiers if you don't like them, without breaking anything else.

In my opinion, this is what Gabe was talking about when mentioning a simplified content creation.
head-explode.jpg
 
Very Interesting Vincent, So you're suggesting almost like "smart objects" and "smart textures" which know when things are happening to them. Do you think source 2 is able to create these kind of things....and do you thinking an average persons desktop combine (who we can hope have increased bot intelligence [ -heck maybe thats something they could benefit])?

Also the textures on alyx jeans was pretty impressive.
 
HAHAHAHA, I see what you did there.

Seriously though, that jeans thing is amazing.


I....uhhhh........ That was a mistake..... My mind thought it out too fast for my fingers to type and went on to a new subject before i could finish......I meant to say average persons desktop While they have a bunch of rebels fighting against a bunch of combine (as in a big battle that takes up a lot of the computers resources).......

So umm yeah....
 
Very Interesting Vincent, So you're suggesting almost like "smart objects" and "smart textures" which know when things are happening to them. Do you think source 2 is able to create these kind of things....and do you thinking an average persons desktop combine (who we can hope have increased bot intelligence [ -heck maybe thats something they could benefit])?

Also the textures on alyx jeans was pretty impressive.

Yep, since they're retiring DX9 and going full DX11+ this time (and appropriate OpenGL + OpenCL for Linux/Mac), they can build the entire engine pipeline around geometry shaders and compute shaders, and those are very capable of producing procedural geometry, morphs, spacial masks, etc, as Dota 2 workshop already illustrated. That's going to be like a Minecraft 2.0 HD for modders, but with smart assets instead of cubes. :D It seems like what Snowdrop engine is doing with its powerful visual scripting, and I don't see any reason for Valve with their in-house engineering talent not to follow the cause.

In terms of AI I don't think everything will change drastically. Players don't need smart AI, they need one that makes them feel smart... and have fun.
 
Last edited:
I don't like to be a downer but the picture of alyx jeans tearing seems to have come from http://www.cgrecord.net/2013/11/download-houdini-cloth-tear-detailing.html which was published January 29th or earlier.

Of course that was just an illustration of the concept. I do not possess a full Source 2 toolset yet. :)

Unacceptable. Those....Mofo....Citizens....Were....Always....In....My....Way....

Well, denser navigation graphs and more advanced behaviour trees are expected of course.
 
Source was mind blowing in 2003-2004. Imagine Source 2 in in the near future.

Seeing the Source tech demo from 2003, I would imagine the Source 2 tech demo would bring even better things to the table. I mean, Valve really showcased how good they we're doing physics, shaders, facial animations, lighting and etc. at the time, so what do you expect from them now? Of course their going to bring the same things, but also other new and innovative mechanics.

Maybe we'll get some water physics.
Maybe we'll get some DMM physics here and there.
Maybe we'll get a nice particle system.
Maybe more advanced and dynamic AI ragdolls. (Like in GTA 4)

Not only this, but technology and ideas that will make map creation a lot more easy, and gameplay a lot more smooth. The hardest of all, was making a map that looks great. It took time and dedication to really make it look good. Hopefully, Valve will make more variations and options to geomotry, props and textures.

I mean, the amount of stuff we can do now is huge. Great technology combined with some great aesthetics; I can't even imagine what immersive world Valve is capable of pulling off. But I bet it will be delicious just to look at.

The way Valve impressed the people back in 2003, was making a consistent and dynamic system, that would clash together with the gameplay. For Half-Life 2, that was physics. One of the most fun aspects in Half-Life 2, was shooting a zombie in half with a sawblade using physics. And it worked so fluid, beacuse it was a dynamic system. So what is Source 2 going to bring instead of physics as a core gameplay mechanic? Who knows, maybe it will be the same. All I know for sure is, that Valve will make Source 2 spectacular.

tl:dr Source 2 =
cake-icon.png
 
Last edited:
Holy crap. You could make a whole game based on manipulating truly fluidic water. I'm always awestruck by what is coming round the technological bend; mainly because it's so amazing, and somewhat because my coffers could only hope to be half ready to afford it.
 
Better development tools. XSI is officially supported by valve, but it's discontinued and unsupported by the original company and is incredibly frustrating to use. Effect editing for some things is impossible without third party software that is difficult to find and use. There is no way to edit the GUI without third party software, which is again, extremely frustrating, and updated hammer for being easier to use. To goldsrc users, the hl and cs map editors, they have no problem with the current hammer editor, but to someone new to it, hammer editor is incredibly frustrating.
 
Of course the graphics updates, every engine update has to have it. Gotta love the 16K textures.
 
Back
Top