Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

FIR filter coefficents WITHOUT FUCKING MATLAB

Name: Anonymous 2005-07-16 23:40

Having recently bought a radio scanner (a UBC72XLT), I'm having a go at writing a morse transmission decoder. I've tested it all working in theory by booting into Windows and using a shareware CW decoder, but the preference would be for longer-term decoding with the machine's normal OS (NetBSD).

Now to start off, I need a tunable digital bandpass filter, settable from anywhere from 600Hz to 1200Hz (the nearest Morse practice station is 916Hz). A FIR filter would do the trick... except EVERYTHING I GOOGLE SAYS CALCULATE THE COEFFICIENTS USING MATLAB'S REMEZ FUNCTION AND HARDCODE THOSE VALUES IN LOL.

As a last resort there's remez libraries around, but I'm not about to bolt 60kB of code I don't understand onto the project. How on earth did they calculate FIR coefficients before fucking MathsWorks came along? (and yes, I Google for that and a hundred variations to the question too. comp.dsp is a shitheap of unanswered questions...)

Name: Anonymous 2005-07-17 2:40

Your searching skills suck: http//www.google.com/...

There are piles of programs that do this.

Name: Anonymous 2005-07-17 6:20

>>2
Gee. Programs to give coefficents to manually hardcode into my program. Exactly what I didn't want. (Lol, many of the links on the first few pages I've already checked out.)

I have varying bandpass filter parameters; I need to calculate the coeffs on-the-fly.

Name: Anonymous 2005-07-17 16:41

I added one term to the above search and got this: http://www.watri.org.au/~ajh/libfilth/

If that's still not good enough, go take apart one of the open-source implementations. If you're writing your own program, this shouldn't be a problem.

Name: Anonymous 2005-07-17 23:46

>>4
It's very good, but unfortunately the library has several large dependencies (GSL, FFTW3, lp-solve (not very portable), and BLAS.) :/
As for picking apart the FOSS hamradio software, all the packages I've examined available to Debian either use FFT instead, or have hardcoded coeffs they themselves lifted from other programs.. The way-too-long search continues...

Name: Anonymous 2009-03-04 13:26

I'm
Gonna
Cry [/o][/o]
;_; [/o][/o][/o][/o]

Name: Anonymous 2009-03-04 13:27

I'm [/spoiler]
Gonna
Cry [/o]

;_; [/o][/o][/o][/o]

Name: Anonymous 2009-03-04 13:27

I'm
Gonna

Name: Anonymous 2009-03-04 13:28

I'm
Gonna[o]

Name: Anonymous 2009-03-04 13:28

I'm
Gonna[o]

df

Name: Anonymous 2009-03-04 13:29

I'm [o][/spoiler]
Gonna

Name: Anonymous 2009-03-04 13:29

I'm [o][/spoiler]
Gonna
Cry

Name: Anonymous 2009-03-04 13:30

I'm [/spoiler]
Gonna
Cry
;_; [/o][/o][/o]

Name: Anonymous 2009-03-04 13:31

I'm
Gonna
Cry
;_; [/o][/o][/o][/o]

Name: Anonymous 2009-03-04 13:31

I'm
Gonna
Cry
;_;
[/o]

Name: Anonymous 2009-03-04 13:32

I'm
Gonna
Cry
;_;

Name: Anonymous 2009-08-03 8:04

newfags cant triforce

Name: Anonymous 2011-01-31 20:09

<-- check em dubz

Name: Anonymous 2013-07-25 18:52

Name: Anonymous 2013-07-25 19:19

I can't into the sage field.

Name: Anonymous 2013-07-25 19:44

FUCKING MATLAB
Too bad Luke wasn't around when this thread was made.

Don't change these.
Name: Email:
Entire Thread Thread List