cM :: MORLEY|DIGITAL

(this article orginally appeared on Daily Tech) 

A perfect HD video experience is in sight, and it’s going to be free with that motherboard you want to buy for your HTPC

I was reminiscing with a graphics partner last week about the history of 3D acceleration and how exciting it was when the Wild West of 3D was ruled by add-in accelerators, 20! FPS in Unreal, and a three way throw down between APIs. 

Remember spending that extra $299 for a second Voodoo II to go from 800×600 to 1024×768 resolution in Half-Life?  Those were the days when you could show your technically apathetic spouse the difference between Quake II software rendering and 3D acceleration and have them genuinely be impressed.  

These days I’m surprised the art of Spousal Acceptance Factor hasn’t spawned how-to manuals.  Try explaining anti-aliasing with gamma correction to someone who doesn’t know pixie dust from a pixel and you’ll just get blank stares.  Today the graphics wars are an orgy of stream processors, technical jargon, and enough bandwidth to run internet backbones if only they could fit in their entirety on a GPU.  And money seems to be no object, considering the upward trend of pricing.  Despite a couple misfires from both camps, ATI and NVIDIA have 3D acceleration down to an art form.

But an underlying, perhaps unsexy to some, struggle is brewing at a level that reminds me of those late 90’s glory days, and unlike 3D graphics where photo-realism is the dangling carrot tied to your chariot, the finish line really is in sight.  And it’s not moving.

GPU-accelerated HD video processing and decoding is just starting to come into its own, and all sides have their guns trained on each other.  As we go forward into an era of 100% CPU offload of the entire video decode pipeline, it’s no longer an issue of dropped frames or not – it’s moving into a realm of who better handles noise reduction, per-pixel motion adaptive de-interlacing, multi-directional diagonal filtering, proper cadence, and other Spousal Acceptance Factor trivia.

To continue my parallel with 3D acceleration, let’s take a look at the two most popular benchmarks for both realms: 3D Mark 06 and HQV.  Arguments about the usefulness of 3D Mark aside, there is no “perfect score” – faster graphics processors and CPUs will get you a continuously higher score and better frame rates.  That theoretically translates in the real world to acheiving higher resolutions, more anti-aliasing, better lighting, etc.  However, Hollywood Quality Video (HQV) by Silicon Optix, the de facto benchmarking software for video processing, does have a perfect score.  You can’t get a higher score than perfect.  There is no extra credit.  Either you handle that obscure cadence from Japanese Anime flicks or you don’t.  And until the Digital Cinema Initiative hits your living room (don’t worry, 2K and 4K are a ways off), 1080p is the only resolution we have to aim for.

The advances being made today in video processing on the PC are awesome.  I recently took a look at a publicly available driver that improved a single HD HQV test component from FAIL to 25 (perfect).  A perfect score in a demanding test in one single driver release.  I can only compare that to going from 3000 in 3D Mark 06 to 20,000 with one driver release on the same hardware.  It just doesn’t happen.  There are still gains to be made in PC-side management of and compatibility with today’s HDTVs, but that’s also improving by leaps and bounds.

And the best part about this whole saga, and what I was referring to in the title, is that all of this technology is getting stuffed into mainstream video cards and is heading to motherboards.  Without the need for power-hungry and ASIC-hogging 3D acceleration transistors, technology like PureVideo HD and AVIVO HD are bringing fantastic HD viewing experiences to the masses without the need for $1200 and 450 watts worth of graphics cards.  Between the NVIDIA 8600-series and the ATI 2600-series, which are the pinnacles of consumer video processing on the PC today, our options are fantastic. 

While the overall war isn’t over, a new battle is beginning to take place in the integrated realm.  The white elephant in the room is Intel, which rules this roost in market share, and I believe the next 6-9 months are going to be awesome.  Current integrated GPUs can’t even handle 100% DirectX Video Acceleration without CPU assist, much less 100% CPU offload of bit stream processing for H.264 or VC1, but you can bet money that it’s coming.

The best part is that it’s not going to cost you an arm and a leg, the finish line is in sight, and it’s an exciting time for GPUs again.  (Yes, DX10 is very cool, too.)

I’ve seen so many pie charts on who is winning the high definition media format race that I haven’t needed food in a month. But, hey, let’s face it, the best way to drive traffic is to talk about 1) Apple 2) HD-DVD vs Blu-ray or 3) Politics.  The thing is, none of us really knows for certain who is going to win. Sure, for some of us it’s our job to expound upon our earthly wisdom and crown a winner. As journalists, technologists, visionaries, yadda, yadda, yadda, it’s what we get paid for. Then we hold on to our butts and hope we’re right.

It’s really no different with Blu-ray and HD-DVD.

If I knew right now exactly who was going to be left standing I’d be able to make more money than the guy who invented the pet rock (like, a million dollars!) It would save on R&D time, focus marketing efforts, and speed up acceptance of the next generation media to come, as consumers never want to bet on a Betamax or Laserdisc again. Not to mention that people would think I was a Really Smart Guy™. The HD-DVD conglomerate and the Blu-ray mafia would love to convince you that they are both pounding the snot out of each other. And they have droves of marketeers with advanced certifications in PowerPoint to prove it. As the saying goes, there are lies, damn lies, statistics, and synthetic benchmarks.

The good news for you is that I’m going to tell you to skip all this data, don’t believe anything you hear about who is going to win, and just buy both formats.

Yup, jump in with both feet and your entire wallet. You really can’t lose. You do not have to choose sides. You do not have to pick the winner in the middle of the race.

Let’s break it down. You’ve got a two to $3,000 1080p HDTV, and probably spent at least as much on the supporting cast members – TV stands, receivers, speakers, etc. C’mon, admit it, you’re a gadget junky. I am too.

Now, if you’re a gamer, this is easy. Either buy a PS3 and a nice Toshiba HD-DVD player (or wait for the 3rd gen stuff that’s coming), or an XBOX Elite with HD-DVD add-on and a nice Samsung Blu-ray player. Or buy both consoles and get out for about $1200 and get some free movies along the way. And you’ll have a wicked time being able to play all the hottest 1080p video games as well. Seriously, is there a down side here, people? ‘Cuz I just don’t see it.

Or, if you’re an A/V purist and want nothing to do with “kids’ toys” – buy the upcoming Samsung combo player that is fully featured (unlike LG’s offering – but God bless ‘em for be the first to throw their hands up in the air and also refuse to take sides.)

There, you’re completely future-proof. You can quit reading all those analysts throwing numbers around or pontificating on which consumer wasteland vendor (Wal Mart or Target) has picked which format out of their ass to declare it a winner.

And once you stop reading what those analysts have to say, the ad click-throughs will fall, and their editors will tell them to get some new f’ing material.

(this article originally appeared on Digital Trends)

Once upon a time PCs were terrible gaming platforms. *gasp* Why, you ask? Well, beside the obvious lack of horsepower or gaming-centric graphics accelerators, the CPUs were simply not capable of processing in hardware the necessary floating-point math required for today’s games.

Starting with the 80386, consumer desktop PCs were the first to benefit from a symbiotic, pseudo dual-core technology that included a second processor for handling these math-intensive calculations. While this processor could not stand on its own or even allow the system to perform multi-threaded instructions, it did intercept floating point instructions and could decode them in hardware.

And boy-oh-boy did video games get a boost. Eventually the x87 floating point co-processor went the way of the dodo when its function was integrated in the 486 processor, although stripped down versions of the CPU did not include the FP unit. Anyone who owned an Intel processor with the SX designation knows it was the Celeron of the day.

Now it’s all about the multi-threaded, multi-core architecture. Both Intel and AMD have made tremendous efforts in delivering unheard of performance on the desktop by incorporating multiple CPUs into a single processor package at a very affordable price point.

This really opens up the possibilities to software designers both in and out of the gaming realm to expand physics calculations, artificial intelligence, and keep games running smoothly without hitches or lag. Of course, you need to make sure your system is running lean and mean with no extra software chugging your system down and wasting the time of the second processor. 

But let me tell you why the multinational Tier 1s benefit from dual-core processing technology more than their customers.

It’s quite simple. They load down their systems with extraneous software that you do not need and only chew up clock cycles. Microsoft has dubbed this bloatware “craplets,” and they’re worried it’s going to ruin Vista’s reception by the public. These craplets of course are revenue streams for the Tier 1s. In fact, with the razor thin margins in this industry, the kickbacks they get may mean the only margin they get! So who’s the real customer at this point? Back to my point…

Bloatware and craplets bog down system performance and make your system less responsive. Dual core technology makes this less obtrusive to the end user as a second processing core can be occupied with the all the pop-ups, advertisements, and “free” offers while you are able to get your work done on the other. And the faster PCs get, the more bloatware they can install.

You didn’t buy a bloatware co-processor, you bought a dual-core processor that is supposed to speed up video encoding, digital imaging, and make for a smoother gaming experience. Right?

Not according to those other guys.

I just wanted to take a moment to comment on my esteemed colleague’s declaration that benchmarks are “Wiggedy Wiggedy Whack.”

Couldn’t agree more, Mr. Sood.

I’ve been saying for years that, when it comes to purchasing a custom computer, fast is easily bought, but ownership satisfaction can remain elusive. It doesn’t matter if that Editor’s Choice winner you paid $8,000 for edged out another $8,000 PC by 1% in some worthless synthetic benchmark. If you can’t get help when you need it, or if you can’t even get someone on the phone, who cares what the benchmarks say? And did you even get it delivered to you within the same fiscal quarter that you paid for it?

Or what about that out of box experience? I don’t know about you, but when I spend a lot of money on an item, many of my initial impressions come from the way the product is packaged and presented to me when I first open the box. To me it means a lot that a company puts as much time into presentation as they do the actual product. It’s the little touches that tell you that you got what you paid for.

I wish that features, usability, out of box experience, and total cost of ownership (which encompasses support costs if they must be burdened by the customer) were easier metrics to communicate in a magazine than who won 3D Mark. I do not think those reviews would mirror what we see today. I think that there are companies out there doing a great job that are not always recognized for it.

It’s good to see that others are now publicly speaking out about this issue like Rahul is. There are mainstream publications that are beginning to do this, and that’s great. I don’t want to name names, but Rich Brown at CNET has done a great job of balancing the benchmarking numbers with other factors about the product or company. There are others that are beginning to as well. Hopefully we can enact a call to arms for reviewers to re-evaluate their own methods and take more consumer-oriented concerns into consideration when they go about reviewing systems.

The end result would be that customers would get a better picture of the companies they are looking to spend their hard earned money with, and those of us on the other side of the fence could focus more on the whole customer experience rather than tweaking builds for certain synthetic benchmarks that will have no impact on the day to day computing experiences of our customers.