
N W F I I R audio tools                                               Release 0s
================================================================================

This is a development release of a room equalisation system, released under
the GNU General Public License.

You will need the GNU Scientific Library (0.4.1, 0.5 and 0.6 tested) installed 
to be able to compile. ALSA (Advanced Linux Sound Architecture) is supported 
(sort of). The code was written for the pre-0.6 found in the CVS, but since the 
API was not frozen at the time of coding, it probably does not work. Contact 
me if you get any problems. If you do not want to use ALSA at all or it does 
not compile with it, simply remove -DALSA from the Makefile.

The source contains x86 assembler, so you will need an Intel processor or 
compatible.

Currently, all documentation is found on the web,
http://www.ludd.luth.se/~torger/filter.html

When this is written (October 1, 2000) it is a bit dated though. If you 
have any questions, just mail me.

** NEWS **
I am currently developing a fast convolution engine, for extremely high
throughput FIR filters. The FIR filter found in nwfiir is a direct-convolution
implementation, and can therefore not reach very high throughput, despite that
it is well optimised. The drawback of fast convolution filters is that it
induces some extra input/output delay, but low delay has never been a goal
for nwfiir anyway (much delay is in it already due to buffering). This new
filter engine will not be an extension of nwfiir, instead it will be an 
entirely new program. This means that this may be the last release of nwfiir 
audio tools, with nwfiir being the central program.

Note: the fconv hack needs fftw (www.fftw.org) compiled with single precision
floats.

Anders Torger, torger@ludd.luth.se

October 2000
