Wednesday, February 3, 2010

Video Silicon in the Laptop?

As video chat and videoconferencing becomes democratized, the question comes up: will computers now sprout dedicated silicon to perform these video-specific tasks?  As in most things, the answer is yes and no; in large part, it depends on what you call a computer.  Here's what we're likely to see in the traditional laptop.

Computers and laptops muddy the question because they can already do pretty good video processing with what they have.  The mainstream has brought prodigious speed advances, multi-core architectures, partitionable 64-bit data processing, and integral pixel processing functions that really accelerate the kind of number crunching required for VC.  So one answer lies here: yes we'll see specialized silicon to do video processing in laptops because we added it years ago, we just didn't do it as an encapsulated hardware codec.  

The laptop exists in a continual tension between cost and performance.  A computer maker is always looking at the tradeoffs between adding abilities and improving existing ones, versus adding the cost to make these additions.  This is why we don't see embedded cameras on every laptop, and why every every laptop's camera we do see doesn't have forty megapixels and a 15:1 zoom lens.  What happens is that when they're added, these cameras get the minimum performance that will do a general-purpose job.  It's the same reason that computers make lousy speakerphones - they have mics and speakers, but they're compromised to fit the price available. And the space, of course; nobody's advertising a laptop computer that's "New!  Improved!  A Half-Inch Thicker for Better Sound!"

Now, sure - not every user is average.  Those who want to stretch the envelope with very high-power requirements, like high pixel counts and frame rates, may need special silicon.  But if they're doing this, they'll need more than just 
the processing - they need the better camera to feed it, and better sound, and a better lens.  And the baseline performance level is continuing to improve as well, so that off-the-shelf computers converge with an ever-increasing subset of user requirements (does anyone besides me find that their cellphone camera takes care of a surprising fraction of their snapshot needs?) 

So this all boils down to one conclusion: computer users are either satisfied with the increasingly good video they can get with mainstream laptops, or they'll invest in outboard stuff to enhance it and that's where high-end specialized processing will wind up.  We won't be seeing dedicated H.264 processors in laptop computers anytime soon, at least as we currently think of the laptop. 

A tip of the hat to Michael Graves,, for asking the question that sparked this train of thought.