Good-bye FFT, heloo HHT

LoopyByteloose

Fast Fourier Transforms have been in use since the early days of computing to make difficult math fall in line with computing. But I am looking at a local newspaper article touting HHT or Hilbert-Huang Transforms as the newest leading edge in such maths.

I have to admit that I have never gotten the hang of using FFT in computing, but this seems an interesting devlopment in making computing more powerful with existing hardware rather than just trying to build a better supercomputer.

I suspect we will be hearing a lot more about this.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
PLEASE CONSIDER the following:

Do you want a quickly operational black box solution or the knowledge included therein?······

···················· Tropically,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

Leon

There are also wavelets, which are better than the FFT for non-repetitive signals.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Dave Hein

Kramer,

Do you have an online link to the news article?· I would be interested in getting more information about it.· I read the Wikipedia entry on the Hilbert-Huang transform (and a paper linked from Wikipedia) and it appears to work well with seismic data, and other non-repeatative data.· Fourier analysis tends to blur out spectral lines, and it introduces false frequencies related to the window size.· It appears that the Hilbert transform doesn't have this problem.

Dave

Agent420

Have a look at these links...

http://www.hitech.technion.ac.il/~feldman/hilbert2.html

LoopyByteloose

I believe the article was in the May 9th, Taipei Times. It does have an online edition, but don't expect the article to explain the maths. It seemed to be more about introducing Mr. Hwang's HHT as a useful tool in regards to global warming.

The Huang-Hilbert Transform seems to be a recent morph of the Hilbert Transform. So you should be using both terms in search engines.

http://dpruessner.info/wiki/images/7/7d/Huang-Hilbert.pdf Try this for starters.

I was impressed because so much of our technology today still exploits Newton's physics and calculus. FFT has extended things further. I don't think most of the human population of the world really understands how much of the last 300 years have been determined by breakthroughs in math leading to more profound technological vision.

FFT was actually heavily depended upon during WWII, but both the British and American government had secret teams of people doing calculations essentially by hand. Years ago, I met one woman in a doughnut shop that worked with sonar design during WWII for sub warfare. She had noticed me trying to read a book on the subject of FFT.

Since is is good for seeing when moving trends are having breakouts, one might use it for techincal analysis of stock market prices and do quite well

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
PLEASE CONSIDER the following:

Do you want a quickly operational black box solution or the knowledge included therein?······

···················· Tropically,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

Post Edited (Kramer) : 5/10/2008 3:30:42 PM GMT

Dave Hein

Agent420 and Kramer,

Thanks for the HHT links.· I read through some of them, but I didn't spend a lot of time trying to understand the details.· I'm not interested in investing that amount of time.·· I was hoping to find a "Readers Digest" version that would provide the basics.· I have worked with various transforms in the past -- Fourier, Hadamard, Cosine, Haar, etc.· Each transform has its own special niche.· The HHT seems to have a special niche in time varying signals.· The FFT can also be used on time varying signals with the appropriate window size and window function.

Dave