Looks like I’m gonna be getting this very soon.

Alright, here’s the first release of SCP-087-BHD. Enjoy it guys!

EDIT: Note that this is a complete game download, you do not need anything else to make this work. This is the complete package.

Well I got the complete Half Life package on Steam today. Helped that I saved a 50% off coupon from the winter sale. It was only 20 bucks for the package with that hefty discount. I even had just barely enough left to buy Geometry Wars: Retro Evolved off of steam, which is one of my favorite all time Xbox Live Arcade Series. Altogether that adds a ridiculous 11 games to my steam library, and adds some of the most widely praised video games of all time to my extensive personal library.

Fun!

Only, there’s already a hitch. Just 57 minutes into the original Half Life, and I already have a glaring problem. My graphics card.

Now to be clear, the game runs PERFECTLY on my computer. I can run Crysis on High settings perfectly with my card too. This problem isn’t something related to my near-snobbish need for graphical perfection, it’s related to a subtle trend I’ve picked up on in the last few years. The trend that my graphics card just isn’t good enough anymore. Metro 2033 really drove this home with its near inability to run in directx 10 mode, and any game with significant use of shaders magnified the problem 100 fold. Just Cause 2 took the low directx10 performance of my card and made it painful to deal with, making it nearly impossible for me to play the game at times, and difficult everywhere else. So as you can see, there’s a clear pattern of “not good enough” that’s been on the rise for the last 12 months or so.

So what’s the problem with Half Life? I don’t have enough video memory to max Anti aliasing settings in OpenGL. My 9600GT holding a pathetic 512MB of GDDR3 VRAM, it can only support a maximum AA level of x16. Which, for the enthusiasts out there, is 2x2 SSAA combined with 4x MSAA. The highest possible setting is 32x, which is 2x2 SSAA combined with 8x MSAA. That might not sound like a lot over the setting I have now, and to be honest the difference isn’t all that extreme. I would rather use 4x4 SSAA by itself, but that requires running the game in DirectX mode (D3D) which has a lot of problems with properly displaying the in game text elements on my computer.

I just find the inability to turn up this setting because of VRAM limitations inherently maddening. Nvidia cards have always been VRAM anemic, but it hasn’t been a real problem for me before now. I remember way back when I still had my 7600 GT, hitting walls like this with simple settings on Command and Conquer Generals was just maddening. It prompted me spending the 200 dollars to upgrade from my old PNY card to my EVGA 9600 GT. I guess I’ve finally hit that same type of roadblock with my 9600 GT.

The only question remains to be, when will the next generation of cards finally release?

Kepler is supposed to be a massive performance improvement on Fermi, it seems stupid to drop $300 on a soon to be out of date Graphics Card with a major release imminent. But if Kepler doesn’t meet performance expectation, or flops like AMD’s Bulldozer CPU, graphics cards like the GTX 560 Ti 448 core are almost guaranteed to disappear in a heartbeat.

Heh, it’s certainly a nerve racking time to be an enthusiast looking for graphics cards.

This started as a rumor that no one believed, but has gradually expanded to the point where every bleeding edge tech site is reporting it like it’s fact. Being as Tom’s is the most recent example, I’m linking to their article here, although a number of other sites are showing this as well. According to what’s been released the initial reports of Nvidia changing their naming scheme and naming these cards the 700 series seem to be utterly debunked. Interestingly though, the more ancient and fascinating rumor that Kepler had doubled the shader count from 512 to 1024 on the flagship single core models appears spot on. Still, here’s a question, why is this being treated like it’s fact when it’s just a rumor?

The answer is simply that it makes perfect logical sense. There isn’t anything in the sheer technical data here that doesn’t line up with what we all expected Nvidia to do, the performance data on the other hand is a little on the shocking side.

Kepler is 45% faster than a card that’s 20% faster than a card that’s 2~3% faster than the graphics core the new Xbox is rumored to based on. Apparently Nvidia learned its lesson from the last console release and hopes to massively out pace the new consoles before they release in 2013 or 2014. Cutting the floor out from consoles like this is a brilliant and dirty move, and should lead to extremely inexpensive versions of these current high end cards being available in future generations to compete directly with the future xbox consoles. In graphical performance it should be a clean sweep in the PC market’s favor. The only question is whether developers will jump on board or will go with consoles despite the extreme technical disadvantage they will have out of the gate.

Personally I’ve been looking for a new graphics card and the 660 Ti looks like it’s along the lines of what I want in a card. What about you guys, what do you think about all this?

I can’t decide on something guys.

In the price range, which one of these is a better video card?

The GTX 560 Ti 2GB?

or

The GTX 560 Ti 448 Cores FTW?

The 448 Cores has 3/4ths of a gigabyte less dedicated memory, but that memory runs faster than the 2GB 560’s does. Still having a 512MB 9600GT I’ve kinda had enough of not having enough vram, but is 2GB too much?

On the other side the 448 Cores also has 64 more CUDA cores that are more powerful than the 560 Ti 2GB’s cores. My 9600GT actually has a total of 64 CUDA cores, so I imagine that the increase in power there will be quite a lot.

So, should I focus on power over vram? I’m not gaming in full HD, my monitor is 1680x1050 and I don’t plan on upgrading it ever any time soon.

What do you guys think?

In one of the best reviews I’ve seen of a graphics card, the AMD 7970 destroys everything with at least 15% more power than a GTX580. Now this is good news for AMD fans who were seriously struck down by the lackluster FX release earlier this year, and means AMD is definitely still a serious player in the graphics industry. The only immediate problem? Some odd results in Skyrim looks like AMD is STILL slipping on driver support. Buy one on newegg now.

In my posts for Image Based Lighting I’ve mentioned that Criterion Games’ Need For Speed Hot Pursuit is one of very few game which uses this system. Considering that the only other game known to use IBL isn’t even out yet, it stands to reason that this may have even been the first video game out there to do this, and I actually have the game. So, I felt it would be cool to go into the game, take some screen shots, and analyze how Image Based Lighting looks and performs on the Xbox 360 version of this game. 56K death machine of images after the break!

Read More

Yesterday I went on a rant about the silent hero that is Image Based Lighting. A random new lighting implementation that provides Global Illumination to complex scenes at high frame rates suitable for video games. Earlier I found and posted a quick little video that explained part of what Image Based Lighting does. Now, I finally know what Image Based Lighting is, if not specifically how it works, just the basics of the concept.

First let me point you to an article on this in Nvidia’s Developer resource site, should you be interested in the details of actually implementing this.

Secondly, let me say the Wikipedia stub is misleading and far too short.

That out of the way, here’s my understanding of Image Based Lighting. First off, in a 3D game the world around a 3D object is converted into a cube map. A cube map essentially being an unfolded cube of 2D images. The cube map we make in IBL being the walls floor and ceiling of your environment. This step is ONLY for environments where a cube map is not already the background. Then the other objects in a scene around the single 3D object {but apart from the background we just made a cube map} are copied by a shader, and the copies are thrown into the cube map and properly layered over each other in 2D representations of how a reflection should look. All of this, is then shown over the texture of the object in question, and appears as a near perfect set of reflections of the objects and lighting around it.

From this, better lighting methods along the lines of Bloom and HDR lighting models can be applied within the effect to give global illumination to an object based off of distance and direction to light sources. However, this all just provides nice reflections and glowing areas, shadows don’t work here. Well actually I take that back, shadows work, but they require something along the lines of the video I linked before. Now, in a real game, you can’t store 10,000 images of light positions and their respective shadows for each object in the game. So there are two solutions provided in the article by Nvidia. Here they are as I understand them.

1.) You fix your light sources in one direction, then compute and store only one shadow per object. This somewhat defeats the purpose of adding something like HDR or Bloom lighting, but keeps the nice reflections with shadows.

2.) You make a few fixed light source directions and blend between them {similar to the video, but with only 8 or so images}. If the shadows are made well enough, they will fool people, regardless of obvious lack of correlation between the light source and the shadow direction.

That’s pretty much all of it. I’m sure there are other subtle things that can be done to this to make it look cooler, and I’m sure my explanation messed up something. Overall though, I think it’s right.

So now that we know what Image Based Lighting actually is… my thoughts on it? I feel it’s like SSAO in a lot of ways; it’s a good effect that needs a lot more exposure. However I think this is a fare amount more significant considering what it accomplishes. Realistic reflections, and Global Illumination are really cool. It’s nice to see them added to games at long last.

EDIT: I need to stop making epic posts in a hurry! I had to revise a lot of this, I hope it’s easier to understand now!

AND THEN THE MYSTERY WAS SOLVED!

More or less…

This is one method used to calculate Image Based Lighting. From What I’m seeing, this has nothing to do with actual lighting simulations. Instead it appears to be infinitely massive amounts of pre-baked textures thrown on an object. Sort of like light mapping, but with pictures for every possible direction light can come from.

This doesn’t seem like what Criterion or Turn 10 have been doing though, so I’m going to keep digging. However, for now, this is more to go on than the Wikipedia stub from before. In any case, have a really bad quality video!

Ray Tracey just posted this up on his blog. It’s a video of a high detail, Path Tracing, Global Illumination, 3D renderer. It’s running on Nvidia’s GTX 580 Fermi GPU, it starts on just one, and then goes up to 8 at about 5 minutes in. The result is extremely close to real time at very high sampling rates with incredibly complex 3D scenes.

What does this mean? Well, considering the GTX 600 series {or whatever Nvidia ends up calling their new Kepler GPU} is supposed to be twice as fast as the 500 series at ray tracing and similar expensive computational programs. It means we can expect 4-way SLI to get these results on the latest hardware next year. Beyond that? 2-way SLI in the following year. Then finally on a single discrete graphics card in 2014.

At that point, 2-way SLI, a very popular choice among enthusiasts, will likely hit the magical real-time 30 FPS target. That’s 3 years from now, and only if more significant improvements don’t occur in the time frame.

That said, the video itself provides some very beautiful indirectly lit scenes. Turn it up to 720p and marvel at it for a while, it’s truly amazing to think we’re so close to this technology being put into place in interactive ways.