Hi,
I thought this article might be interesting to some of you. Here is the project's web page.
Faster algorithm than FFT for sparse signals found
-
- KVRist
- 146 posts since 18 May, 2011
Oh please dont...I study computer sciences and the Fourier transform is the main subject of the math class...It took me quite a long time to study it and now it just pisses me off...NO MORE FT! 
Looking for a proper one.
-
- Banned
- 12367 posts since 30 Apr, 2002 from i might peeramid
math is hard. let's get ice cream after school, in my new car. with ken!
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.
-
- KVRist
- 53 posts since 5 Jul, 2010
Yeah and it's addictive.
Math: Not even once.
Math: Not even once.
-
- KVRist
- 278 posts since 2 Nov, 2009 from California Central Coast, USA
IIRC this is lossy 
"But as the researchers point out in their new paper, empirical studies show that on average, 57 of those frequencies can be discarded with minimal loss of image quality."
"But as the researchers point out in their new paper, empirical studies show that on average, 57 of those frequencies can be discarded with minimal loss of image quality."
-
- KVRAF
- 1617 posts since 11 Dec, 2008 from Minneapolis
Yeah, 'sparse' stuff ... It also looks like there's overhead such that there's a lower bound on the number of samples 'n' before it starts to outperform FFTW. It looked like that bound might be sort of high for this to be completely relevant to audio DSP, but I wouldn't really know. There was also a higher bound against another sparse algorithm for lower 'k'/#frequencies pulled out, but the authors seemed pretty excited about being optimal (provably so asymptotically) for certain ratios of k and n higher than previous state-of-the-art sparse FFT-ish stuff.robgxmn wrote:IIRC this is lossy