W3C home > Mailing lists > Public > public-audio@w3.org > January to March 2012

Re: fftSize requirement for RealtimeAnalyserNode

From: Raymond Toy <rtoy@google.com>
Date: Fri, 23 Mar 2012 15:27:13 -0700
Message-ID: <CAE3TgXFcdJhGrfwca3Hrb=qhHFLaRku6qHBfvRCPW8K_H+edxg@mail.gmail.com>
To: Chuck Adams <cadams@fender.com>
Cc: Jussi Kalliokoski <jussi.kalliokoski@gmail.com>, "Wei, James" <james.wei@intel.com>, public-audio@w3.org
On Fri, Mar 23, 2012 at 1:18 PM, Chuck Adams <cadams@fender.com> wrote:

> Power of two isnít an arbitrary limit in the FFT, itís a fundamental piece
> of what makes it the Fast Fourier Transform.     If you want an arbitrary
> number then you canít use the  FFT algorithm, you need the discrete Fourier
> transform or the like.****
> **

No, there are fast algorithms that are still O(n*log(n)) for lengths that
are not a power of two.  See, among others, www.fftw.org.

Received on Friday, 23 March 2012 22:27:42 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:49:58 UTC