Since this week begins the release of the new Sony and Microsoft game consoles, and Gabe Newell is doing his absolute best to keep his SteamOS / SteamBox products in the front of your brain, now would be a good time to let my opinions on the subject loose.
First off, describing “Generations of Gaming” in terms of console releases is stupid. Wrong mostly, but stupid too. Why? Because industries this big don’t turn on a dime, and constantly change in small but measurable ways. It would be much more accurate to describe the history of gaming as a continuum, with console releases being just some of the milestones on that line. The best example of this is the launch PS3 and Xbox360 vs the versions that ship today. Both systems have undergone extreme changes, from UI to basic capabilities from the day they shipped to today. The Xbox360 launched without service apps, a web browser or the ability to natively play any video other than Windows Media formats. They have tweaked and changed the 360 to support a couple of internet services, some extra codecs, and added in a web browser and voila! a new device entirely, much closer to the vision of the Xbox One than the original launch 360. The same could be said of the PS3 with significantly improved online services through the Playstation network and the addition of web apps as well, although Sony also reduced functionality at the same time with the very controversial removal of Yellow Dog Linux support from their consoles, as well as the revolving door of backwards compatibility strategies they used. Because of these changes, if you were simply looking at a list of specifications you would put a 2005 Xbox 360 in an entirely different generation as a 2013 Xbox 360. Because of this fact I don’t pay much attention to ‘this generation’, ‘last generation’ or ‘next generation’ and look simply to capabilities. For the purposes of this article I’m going to compare gaming hardware from a few different perspectives to prove my continuum hypothesis.
The first point of comparison is, of course, graphics capabilities. There are a few different things to consider here, although the one most widely discussed and compared is raw computing power as measured in FLOPS (although the older chips did not measure general computing performance, so we’re going to go old school and measure Polygons Per Second like we’re still in high school). This is actually the silliest of all hardware properties to compare against, but it is what it is, so let’s start there. First lets go all the way back to the first game consoles and PC Graphics cards designed specifically for polygon pushing 3d: The Sony Playstation, the Nintendo 64 and the Voodoo Graphics systems as contemporaries.
As you can see, the winner here is the Voodoo 1 by about 2 – 10x. So for the first 3D Generation, the PC started with a commanding lead which only grew throughout those years. Skip ahead to the next series of consoles, and we begin to see an interesting evolution. Let’s compare the ‘second’ generation of 3d gaming, the Playstation 2, the Nintento GameCube and the Xbox, along with the high end graphics card of the day; the GeForce 3. So, by this point 3d graphics were the expected norm with sprite-based 2d games all but gone (although they do make a reappearance a few years later on the web as flash games) and Microsoft had waded into the console arena with the Xbox. So who had the more powerful graphics? The console industry, or the PC.
There you can see that the Xbox and the GeForce3 were neck and neck, leaving the rest of the pack behind, considerably. Why is that? Well, the Xbox WAS a PC. It was a Pentium 3 with a Geforce3 (basically) and it helped Microsoft establish a foothold in the console industry.
As we head into Generation 3, its clear that, with the exception of the xbox which basically was a high end gaming PC stuffed into a little black box, the console industry has definitely lagged behind in terms of raw power. Things spread out again when we hit the current generation (assuming the PS4 and Xbox One are ‘next generation’). This time around Sony and Microsoft were the big players, and they wanted to convince everyone that their machines were space age tech, man! Faster than any PC you could buy at the time! Well, not really as we’re about to see. The Playstation 3 used the nVidia Reality Synthesizer (RSX) which was really just a rebranded 7800 GTX with less memory bandwidth and less pipelines, the Xbox 360 used the Xenos Chip and the Wii just sucked, so we won’t talk about that in the graphics portion. Basically its about as fast as an Xbox1, and slower than whatever GPU is powering your smartphone. This is the generation where we were promised PC-quality graphics? Did we get it? Not even close. The high end graphics card to compare this generation to is the nVidia Geforce 8800 GTX, which was the first Unified Shader Model video card and changed the graphics game forever. It came out a little bit after the console releases, but it’s the reason that splitting video games into console release generations is stupid. Just so we have a baseline, lets discuss what the 8800 GTX was. It was the first ‘real’ GPU, one capable of running any general purpose code on programmable shader processors that could run intense parallel calculations. In graphics-centric terms, it also allowed the GPU manufacturer to create a continuous die of graphics processor, rather than splitting the work into Vertex (Polygon) and Texture (Textures…) shaders, which was always a balancing act on what current games would require vs future games. It also kicked the crap out of anything else around for 2 years straight.
These numbers only tell half the story, because during this time is when the graphics industry moved from fixed-function shaders to the unified shaders of the 8800GTX. The RSX in the PS3 used fixed function, and Xenos used a unified shader model. This is why Xbox 360 games have had to make far fewer tradeoffs on new titles and tend to look a lot closer to their PC counterparts. This brings us to today, when Sony and Microsoft are releasing their ‘next generation’ consoles, with all the glory that entails. So, in terms of numbers, has the use of standard PC parts (not really, but ok) made a difference in closing the performance gap? Well, sort of. The gap smaller than it was in the PS1/N64 days and bigger than it was in the PS2/Xbox360 days in terms of raw numbers. Both the PS4 and the Xbox One use a variant of ATI APUs based on Jaguar CPU cores (8 of them) and Graphics Core Next (GCN) GPU.
As you can see, the PC side is still kicking ass in terms of raw power, so if hardware were the determining factor, wouldn’t the generational shifts be measured based on dedicated Video card releases instead of console releases? Or on feature sets based on OpenGL or DirectX capabilities? Or, why don’t we choose an entirely new metric?
This brings us to the next obvious metric to judge video game generations on: Features. This is actually a lot more nuanced a conversation than some charts with processing power, because an Xbox1 has more in common with a modern PC than a 386 from 1992 does, and an Xbox360 from 2013 more closely resembles an Xbox One than it does its own 2005 launch version.
When the original big daddy of consoles launched (NES), it could play games. Thats it. No UI, no non-game interface at all. The same was true for the SNES, and again for the N64. But Sony had a different path in mind when they released the Playstation, likely owing a lot to its success. It could play CDs. This was a very big deal at the time, since it meant you could have a living room CD player without another box. This was the start of the console-as-entertainment center phenomena which is basically considered the norm today. When the Playstation 2 and Xbox’s launched one of the key differentiations went in Sony’s favor, in that the PS2 was a DVD and MP3 player out of the box, when DVD players were not common. The Xbox had the capability but required the Xbox Remote peripheral to make it work.
However, partway through the Xbox’s lifecycle, something amazing happened. Because it was based on generic x86 PC parts a group of hackers got together and wrote a new operating system for it, based on Linux called Xbox Media Center. You might know it better by its modern name XBMC and it was INCREDIBLE. This development forever changed the course of console development by putting full blown PC media capabilities on a console for the first time. It could play MP3s, it could play NES and SNES games (what??), it could play downloaded video files sitting on a network share. It could do it all, and would greatly influence both Sony and Microsoft in their next releases.
By the time the PS3 and Xbox 360 were rolling off the Chinese factory lines, the video game world had matured considerably. It was clear that this was not a child’s hobby, and that there was serious money to be made. Sony spent billions trying to create a machine that would outperform the 360. They had custom processors made, with a custom instruction set trying to squeeze a little more performance out of the silicon of the day. Whereas Microsoft fell back on using off-the-shelf PC hardware, just not any that was previously associated with Microsoft. They used Macs. Or the PowerPC processors found in Mac G5s to be precise.
Even though Sony tried their damnedest to make a faster machine, as we saw above, they failed miserably. Both companies advanced their media feature sets to more closely resemble the Xbox Media Center machines, with the ability to play video files off of special servers (DLNA) and with some pretty extensive MP3 support. There were issues with codecs that shipped, and a lot of work needed to be done on the back end to get a good experience, but the support was there. The Xbox360 could even act as a Windows Media Center extender and was a fantastic machine if you had that infrastructure running in your house. And that’s not even getting into the HD Disc format war, which was REALLY stupid since the internet kicked the crap out of both, but Sony had the major advantage there with native Blu-ray support.
However, the real generational shift around 2006 did not happen with the media functionality, although it continued to improve throughout the lifetime of both consoles, dramatically changing the advertised feature sets over 5 years. The real generational feature shift came from little ole Nintendo, with the WiiMote. This little bastard made gaming fun for million of people who hated sitting on a couch holding a weirdly shaped vibrating plastic toy.
This change in control methodology was so revolutionary and well received that both Sony and Microsoft scrambled to come up with their own version of it. Sony used cameras to achieve this feat, with the Playstation Eye and Playstation Move, while Microsoft took the concept a little farther with Kinect. This kind of extraneous control would play a key part in the next set of releases, with the bundled Kinect controversy.
Both the Xbox 360 and the Playstation 3 changed so completely in every way throughout their lifespans, that if you look at anything other than specific hardware selections, the change in generations clearly happened around the release of the Kinect, not of the Xbox One. There were web apps (Netflix, YouTube, Internet Explorer and the PS3 Browser), far expanded local media options and even alternate control techniques (Kinect and Move), as well as the digital distribution of install-able games.
So, now we come to today with the release of the Playstation 4, and the Xbox One next week, asking ourselves “what is new this generation”. Well, the answer is all of the above, greatly improved Media capabilities, excellent control schemes for our devices and a consistent social media experience. Except these things were all added to last generation hardware. Yes, games will look better, and the additional RAM will allow for new experiences that the current crop of consoles simply cannot properly do, but the real feature adds are all software at this point, and speak to the evolution of the console as a concept.
So, with all that in mind, I am going to propose the following new terms of ‘generations’ of gaming, that take the entire wide world of gaming into account. It will be fuzzy, and much less absolute that the media using console releases, but it will tell a much fuller story.
First Generation of gaming: Lets call this the Sprite generation. This ranged from Asteroids to Super Mario Bros and spanned the mid 1970’s until the mid 1990’s. Games were created using sprite based 2d images that moved around a screen. Smaller images led to more lifelike animations but the basic principles stayed the same.
Second Generation of gaming: This one would be called 3D Generation 1, or Fixed Function 3D. This started with PC games like Quake, but moved to consoles with the N64 and the Playstation. There was a transition period in the console world, where some Neo Geo games used a combination of sprites and rendered models to get more 3D detail before the hardware could really support it. The end of this generation began with the release of the GeForce256‘s hardware Transform and Lighting engine and is just now completing with the release of the Playstation4. However, with the rise of mobile games on fixed function GPUs, this generation might have a few more years left in it.
Third Generation of Gaming: This is the current graphical generation, and it would be called General Purpose 3D, where the GPUs inside our hardware can do more than just 3D graphics, but can run many alternate parallel operations, allowing for greater interaction with particles and a closer adherence to the laws of physics. This allows us to not only leave bullet holes in walls, but actually blow holes in them. This generation started with in part with the GeForce256 all the way back in 2000, but has really taken over since the GeForce8800 and Xenos chip in the Xbox360 showed how much more flexibility this kind of graphics hardware gives developers. This will end when the processing power of Ray Tracing becomes a reality and our images become generated by calculating the path that light would take in the image rather than building them from triangles.
Forth Generation of Gaming: This would be the Internet Generation, and has less to do with graphics and more to do with functionality. This is the era of internet connectivity with gaming, that again began with Quake and Half-life, moved to the console’s with the Dreamcast‘s web browser pack and 56k modem and was solidified with the Xbox and Xbox Live. Even today this generation continues with gaming hardware being largely a client for internet services.
Fifth Generation of Gaming: I would call this the Media Extender generation, and is going full swing with Microsoft’s focus on TV and video with the Xbox One. This began with the Playstation playing CDs and has been a major influence on the rest of the industry. This generation will continue for long after the PS4 and Xbox One have lost their luster, but it wont go forever.
Sixth Generation of Gaming: This is the other ‘generation’ that we find ourselves in now, and I refer to it as the Advanced Control generation. There were some attempts in the past to get other forms of control going beyond the basic 2-axis control + buttons that both consoles and PCs have enjoyed for decades, but the real shift was the release of the Nintendo Wii. It allowed for some very interesting and unique game experiences by use of the free-flow controller wand they created. This was so successful that even though it didnt make a huge dent in the AAA gaming sphere, Wii games are among some of the highest selling games of all time. Because it didnt feel the same as an NES or an Xbox. This has been taken and run with by Microsoft and the Kinect, and the new Xbox One will be almost completely controllable this way, with voice and gestures. This generation is here to stay, and never again will go back to a single method of control.
So there you have it. Other than marketing, there is no actual reason to call the new consoles a new generation. Their graphics were almost matched in the previous generation by PCs, and all the software advents they bring to the table were worked into the last models through software during their lifetimes. New control schemes, new graphics techniques and the media functions all existed to some degree or another for the better part of a decade now. The main point is, that advances in gaming are evolutionary not revolutionary and usually occur mid-cycle, not between them.
Please feel free to tell my why I’m wrong in the comments.