Re:long FIR filters
Re:long FIR filters
- Subject: Re:long FIR filters
- From: Ian Kemmish <email@hidden>
- Date: Fri, 2 Feb 2007 18:25:04 +0000
On 2 Feb 2007, at 5:58 pm, "Cor Jansen" <email@hidden> wrote:
I would like to implement a very long FIR filter for a room
correction application (about 10.000 taps).
I know the PowerPC chips has the Altivec module to do this kind of
tasks efficiently.
The "Developer" section of Apple's website has some documents about
porting from well-designed instruction set architectures to brain-
dead ones.
For "straightforward" usage, such as filters (straightforward in the
sense that a lot of people do it), the key recommendation is that
instead of using the vec_* built-in functions, you first of all
rewrite your code to use the higher level functions in the Accelerate
framework. Once you've done that, your vector code should
automagically become portable.
Of course, if you're not doing filters, life can be more
interesting. I cache short segments of audio in the vector register
file and have sequences of tens of thousands of vector instructions
at a stretch which have no equivalent on Intel at all. I'm not-
entirely-flippantly wondering how hard Apple's lawyers would jump on
me if I wrote an AU host for the version of Embedded Linux that runs
on Playstation 3....
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - -
Ian Kemmish 18 Durham Close, Biggleswade, Beds
SG18 8HZ
email@hidden Tel: +44 1767 601361 Mob: +44 7952
854387
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - -
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden