Source 2 engine possibly in development

[JumBie]

http://www.valvetime.net/threads/exclusive-next-gen-source-2-engine-is-in-development.215336/

Our resident investigator Barnz has done some more digging in the Source Filmmaker script files, and he's uncovered probably some of the biggest news we've ever reported on our site.



He's found many more references to a "Source 2", but not just one vague line like our previous entry. We're only one file deep, and already it's clear that the references to "Source 2" are indeed referring to a next-gen engine that Valve is currently developing. I'm only going to show one line in this post, but there are about 60+ references here, and this one line is probably the most telling.

Here's the code paste at Pastebin.com, and this is the file path on the harddrive:

".../SourceFilmmaker/game/sdktools/python/global/lib/site-packages/vproj/vproj.py"


Line 1387:
'''Return an str with the current engine version.
If key doesn't doesn't exist, assume 'Source', otherwise invalid -- assume next-gen 'Source 2'.'''


There are references to Source, Hybrid, and Source 2 in this file. We speculate that Hybrid is the base for which Source 2 is sitting on. Barnz also says it looks like Source 2 is using the same pipeline as Source, meaning that it will probably be keeping its .BSP level format. We haven't found any references to DirectX or OpenGL yet, or Episode Three for that matter, but we're still investigating.

The Ep3 reference from before was just a comment, so it's unlikely we'll find any more mentioning of it. Source Filmmaker was in development parallel to Source 2, hence why these script files have so many references to the new engine.

Valve wants you to know that they "don't like to brag, but Source is considered the most flexible, comprehensive, and powerful game development environment out there. And it's about to get even better."

 

[Gordon Freemen]

Screw another Source Engine how about some EP 3 first Valve LOL.

 

[Grooveriding]

I hope it's impressive. Source was pretty decent back at release day of Half Life 2. Hopefully when Source 2 sees the light it will be comparable/better than the likes of Frostbite 2/CryEngine3/Unreal 4 etc.

More importantly that they release EP3 on the engine, followed by Half Life 3. FFS.

 

[Gordon Freemen]

I hope it's impressive. Source was pretty decent back at release day of Half Life 2. Hopefully when Source 2 sees the light it will be comparable/better than the likes of Frostbite 2/CryEngine3/Unreal 4 etc.

More importantly that they release EP3 on the engine, followed by Half Life 3. FFS.

I would not get too worked up about an all new engine I would be more than happy/surprised if EP 3 sees the light of day this year or next LOL.

 

[Googer]

Over the years rumors have been flying around to a Source Engine 2. Not long ago, valve denied rumors of any sequel to the source engine and stated they preferred to improve upon an existing and mature platform vs starting over from scratch. It's highly likely this is just another bit of internet horse manure.

What ever is to be released will probably be just another revision to the existing code base.

 

[Stringjam]

....Source is considered the most flexible, comprehensive, and powerful game development environment out there.

Really?

 

[gorcorps]

Really?

Well when you use generic terms like "powerful" you can make it mean whatever you want. We know it's not the most graphically advanced, but it's certainly one of the most scalable engines out there. People with fast enough hardware to run other modern engines well don't care, but medium and low end users have historically had more bang for the buck with the Source engine.

That said... it's too damn old at this point.

 

[EDUSAN]

i would agree with what valve said, i also think Source to be quite flexible and powerful and easy to learn/use, i just think that the games that used the engine were not so good, except the games that valve themselves made (TF2, L4D, Portal), and they never cared much for graphics.

people see source engine as it was in hl2, but the engine, as far as i know, had big improvements in time, adding features that were shown in ep1 and ep2 if i recall correctly, like HDR and bigger open areas

but well, as i said, i never saw ANY great game being done with source engine besides valve's games

 

[Tweak155]

If every game looked like HL2 I'd be fine with it. It isn't jaw dropping but it is good enough. Let developers actually focus on a story and see where it takes us.

 

[Rakewell]

I wouldn't be surprised if S2 is set to herald HL3

 

[gorcorps]

I wouldn't be surprised if S2 is set to herald HL3

There's been many with the opinion that at this point, they better be using a new engine for HL3 or it'll just be all this waiting for an "outdated" game.

 

[Red Storm]

There's been many with the opinion that at this point, they better be using a new engine for HL3 or it'll just be all this waiting for an "outdated" game.

It's funny though, I don't consider Source outdated at all. Portal 2 was a fantastic looking game.

 

[Texashiker]

Source engine 2, or some kind of improvement in the source engine only makes sense.

Source is going on 8 - 9 years old. Even though there have been various updates, it still does not support large maps (borderlands for example) or Dx11.

 

[Olikan]

in youtube... i remember a modder was working in radius lightning system for source....yes, the same system used in crysis 2 and BF3
i post the video latter, currently at work

 

[PrincessFrosty]

They badly need to overhaul the engine, it's stuck heavily using BSP for world geometry which has some pretty significant problems for large maps, open areas and general optimisation. The developers tools are absolutely awful in comparison to something like Unreal Engine 3 and Crytek engine especially where you can just drop straight into levels and play immediately.

Quite frankly the graphics have been behind the curve for a long time, it's still all DX9 which looks extremely dated, the additions to the engine has mostly been pretty shallow cosmetic things, cut down HDR, film grain, pre-baked physics the kind of thing other engines would do in real time these days.

They wont kill the engine to make source 2 though, I'm sure it will just be a large overhaul most of the core stuff will remain, I very much doubt if we'll even see a move away from BSP at this stage.

 

[moonbogg]

Meh. Its going to be a long forever of a time before we see anything new from Valve. I guess 2015 or later before any new games come around.

 

[wuliheron]

The source engine is a heavily modified version of the original Quake engine. Modified for Havoc physics in particular and Portal 2 is an excellent demonstration of just how much basic physics can be done on even something as wimpy as an old dual core processor. If I had to guess the source 2 engine will expand that to a minimum of a quad core.

There are plenty of physics programs that can run on a gpu using 80-300 streaming processors, but the hard part is finding ones that can run on two or four cpu cores and still add something worthwhile. That's the kind of thing you need a mathematical genius for and one reason it's taken them so long to develop the source 2 engine.

 

[VulgarDisplay]

I thought it was a heavily modified version of GoldSrc which was a heavily modified version of Quake?

Either way, Valve needs a new engine simply to keep up with the times. Unreal and Cryengine now basically let you develop in real time for all platforms simultaneously. They also free up programmers by allowing almost anyone to edit the way a level behaves with simple commands. If Valve plans to keep up with these next gen engines they need to build a new engine.

I can't wait for the 6 month dev cycles of useless crap to begin. New engines like unreal 4 and cryengine 3 will allow developers to crank out games at an alarming rate in the future.

 

[Olikan]

source engine is very, very modular

...all done by a badass modder

http://www.youtube.com/watch?v=4IzCAhNNOCU&feature=player_detailpage

http://www.youtube.com/watch?v=bMjXx-KweIo

http://www.youtube.com/watch?v=82AMwQGjc8Y&feature=player_detailpage

 

[wuliheron]

I thought it was a heavily modified version of GoldSrc which was a heavily modified version of Quake?

Exactly. They took a heavily modified version of the Quake engine and combined with heavily modified Havoc physics. The basic idea was to start with a graphics engine built for speed and then see how much physics you could add without slowing it down to a crawl.

Either way, Valve needs a new engine simply to keep up with the times. Unreal and Cryengine now basically let you develop in real time for all platforms simultaneously. They also free up programmers by allowing almost anyone to edit the way a level behaves with simple commands. If Valve plans to keep up with these next gen engines they need to build a new engine.

I can't wait for the 6 month dev cycles of useless crap to begin. New engines like unreal 4 and cryengine 3 will allow developers to crank out games at an alarming rate in the future.

Every engine in existence does not have to be like every other engine in existence anymore than every game has to be shooter. CPU physics are just plain difficult to do and I'm glad someone has taken the time and effort to do them when in general physics are the bastard child of the industry. Being able to do them on 4-8 cores is about ideal because 8 cores allow for full blown matrices, but if they're going to continue beyond that they'll have to start incorporating gpu physics so my guess is the source 2 will be the last engine of it's kind.

 

[JAG87]

What Valve claims is 100% true. Source is by far the best engine ever made with respect to quality vs. performance ratio, with only Codemaster's EGO 2 trailing behind, and Frostbite 2.0 far far behind.

Engines like Unreal 4 and CryEngine 3 (never mind engines like 4A used in Metro 2033), aren't even competitive, almost nobody besides hardcore gamers have the hardware required to run them properly at native resolution, and they have to be turned down so much that they end up looking worse than Source at high settings.

And that doesn't even take into account anti aliasing. MSAA is practically free in Source, whereas in most newer engines it takes what is already a less than ideal FPS into a totally unplayable FPS.

An engine released 8 years ago is still able to produce wonderfully looking games (Portal 2), which 90% of gamers can run at 60+ FPS, and those who are hardcore can even run it with SGSSAA and have flawless image quality.

That's what the industry should be aiming for, instead of putting out games that require $1000 worth of GPU to reach 40 FPS, and then releasing garbage post AA filters to try and smear the awful graphics. But I guess NVIDIA's and AMD's money is as green as our money.

It's just too bad for us.

 

[VulgarDisplay]

Exactly. They took a heavily modified version of the Quake engine and combined with heavily modified Havoc physics. The basic idea was to start with a graphics engine built for speed and then see how much physics you could add without slowing it down to a crawl.



Every engine in existence does not have to be like every other engine in existence anymore than every game has to be shooter. CPU physics are just plain difficult to do and I'm glad someone has taken the time and effort to do them when in general physics are the bastard child of the industry. Being able to do them on 4-8 cores is about ideal because 8 cores allow for full blown matrices, but if they're going to continue beyond that they'll have to start incorporating gpu physics so my guess is the source 2 will be the last engine of it's kind.

I was more referring to the fact that the new Cryengine and Unreal engines will allow developers to put games in in probably less than half the time it used to take.

 

[tcG]

I've noticed that Source engine games have a certain feel that makes them seem more involved and direct than other engines, which seem more like you're playing a movie.

 

[wuliheron]

I was more referring to the fact that the new Cryengine and Unreal engines will allow developers to put games in in probably less than half the time it used to take.

The entire industry is moving in this direction because of the skyrocketing costs of developing AAA games and it has nothing to do with the underlying issue that Valve has to first develop the physics for the new engine. If they merely took the existing source engine and modified it to pump out crappy games as fast as they could they'd be shooting themselves in the foot.

 

[Rakewell]

What Valve claims is 100% true. Source is by far the best engine ever made with respect to quality vs. performance ratio, with only Codemaster's EGO 2 trailing behind, and Frostbite 2.0 far far behind.

Engines like Unreal 4 and CryEngine 3 (never mind engines like 4A used in Metro 2033), aren't even competitive, almost nobody besides hardcore gamers have the hardware required to run them properly at native resolution, and they have to be turned down so much that they end up looking worse than Source at high settings.

And that doesn't even take into account anti aliasing. MSAA is practically free in Source, whereas in most newer engines it takes what is already a less than ideal FPS into a totally unplayable FPS.

An engine released 8 years ago is still able to produce wonderfully looking games (Portal 2), which 90% of gamers can run at 60+ FPS, and those who are hardcore can even run it with SGSSAA and have flawless image quality.

That's what the industry should be aiming for, instead of putting out games that require $1000 worth of GPU to reach 40 FPS, and then releasing garbage post AA filters to try and smear the awful graphics. But I guess NVIDIA's and AMD's money is as green as our money.

It's just too bad for us.

QFT