Subject: Re: [linux-audio-dev] float a/d (was Linux support for IEEE1394 mLAN?)
From: Karl JH Millar (kmillar_AT_MIT.EDU)
Date: Thu Sep 28 2000 - 05:38:13 EEST
Tom asked:
>3) Are the currently superior low pass, band pass, and high pass
>filters using float or integer math?
I haven't looked at the code of the filters, but from my understanding of the
numerics of these types of filter, I'd strongly expect them to be done using
floating point, and probably use double precision internally, even if the
input and output are single precision. (Incidently, this incurs almost no
overhead at all on x86 cpus.) The numerics of low frequency filters are one
of the few cases where precision is really important. (If you've done numerical
analysis think of solving stiff differential equations - it's a similar
problem.)
Karl.
This archive was generated by hypermail 2b28 : Thu Sep 28 2000 - 06:29:36 EEST