This post is part of our February blog series. We asked several thought leaders to respond to this question: "What technology do you love most?" Stay tuned for more posts in this series, and for the upcoming March series on UX/design. This post is by Tom Anderson, President and CEO of Weymouth Design. Tom is interested in how advances in mobile and digital technology, like FFTs, can transform healthcare and interactive design. Find him on Twitter.
A few weeks ago, Nick, one of our video editors, was waiting for the server to compress and crunch his brilliant work into a viewable, transferable, and integrated file. Time was wasting. I had to remind him of the good old days in the late 70s and early 80s, when I was his age, and the long nights were spent lining up tapes and editing in a painful, laborious, and repetitive all-nighter fashion.
Are things ten times faster now? At least! And thanks to a breakthrough at MIT, things could get another 10 times faster pretty quickly. Video compression. Image compression. Audio compression. Certain types of analytical tools and mathematical analysis. My view: The tools we use in our industry will work better. The leap from idea to execution will be faster. Users will have a richer, more immersive experience. Think: Better 3-D. Think: A quantum leap in mobile delivery. Think: a high-IQ Siri.
The breakthrough – Faster Fourier Transforms. I read about it on Fast Company, jumped to MIT news, contacted people at Adobe, kibitzed with my technically adept and quick posse of innovators, one of whom is a math historian, (“oh, yes, we read that and were pretty excited too”). And the higher math in the research paper looked good to me! Siri, I made that up, but if the mathematics holds, it is only good news for our favorite digital assistant’s brain and voice.
Let’s back-up. Jean Baptiste Joseph Fourier was a French mathematician and physicist who, among other things, is credited with first noticing the greenhouse effect. His transforms are a way of breaking data down into basic elements – kind of like data atoms – so that we can see what it’s made of. Any piece of music, for example, can be broken down into and, by the way, reconstructed from, Fourier’s basic building blocks. The reason that Fourier tranforms are good for compessing music and video is that when you break this data into basic blocks you discover there are blocks you don’t need – overtones of a violen above what human beings can hear, for example – so you just throw them away. The trouble is that cracking data à la Fourier takes a lot of computing, so if you want to do it before closing time you have to do it faster.
Two statisticians, Mr. Cooley and Mr. Tukey, figured out a better way in the mid 1960’s and created the Fast Fourier Transform. Do something fast and somebody will want to do it faster. If the data you’re tearing apart is of a particular kind, a kind that’s not all that unusual, then the folks over at MIT have figured out a way to go Cooley and Tukey one better. Voilá, the Faster Fourier Transform. Siri is getting delirious with glee! Nick may still have all nighters – but it will be for adding a couple of art nouveau flourishes rather than grinding out needlessly obese and attenuated video files.
For the mathematics, history, and Fast Company aficionados:
The faster-than-fast Fourier transform
Larry Hardesty, MIT News Office, January 18, 2012
How Speeding The "Most Important Algorithm Of Our Lifetime" Could Change This Modern World
By KIT EATON Mon Jan 23, 2012
With grateful acknowledgement to a math historian and teacher, Scott Guthrey, Publisher, Docent Press. http://docentpress.com/
*Graphic image by Christine Daniloff, courtesy of MIT News
Graphic image by Christine Daniloff, courtesy of MIT News.