Hi Steve,
(also thanks David for replying).
On 18-04-12 04:05, Steve Underwood wrote:
Talking about learning DSP, I have just started in a new book on DSP
("Understanding Digital Signal Processing") to learn more about the
practicle approach of DSP.
The one by Richard Lyons? That's a very popular beginner's book, and
seems to get people started who have struggled badly with other books.
It's indeed an interesting book. A ham from Czechia proposed this book.
Now I must say I do am glad this is not the first book I bought on DSP.
For something who is completely new to the subject, it does go over some
subjects quite fast.
I had already read the book "DSP, A practicle guide for engineers and
scientists" before which was good to first "get your head in the correct
twist" and get a good grasp of what DSP is. (the book is also online:
http://dspguide.com/)
It's much more theory and "explaining", but -for me- I needed this to
understand what DSP is about.
These two books do complement eachother quite well.
Now, I get the impression that that to do this, I also need to learn to
work with tools like scilab, GNU octave, freemath (or other open source
alternatives to mathlab). Can somebody point me to good book or
information on how to use these tools for DSP-related applications.
The documentation that comes with these applications is all several
thousands of pages!!!
I never use anything but C for my DSP work, both for algorithm
development and final implementation. Use whatever you feel comfortable
with. If you use matlab there are lots of toolboxes to makes a
professional's life easier, but you won't learn much unless you work
through what's inside the toolbox.
I'm interested in learning these tools as they seams to be a good way to
experiment with DSP.
I just want to do some basic things like create a frequency-plot of a
FIR or IIR filter, just based on it's coefficients so I can experiment
with this a little bit.
The gmskmodem actually has two FIR filters included in it (code that
came from the "pcrepeatercontroller" software project), so I would be
interested to actually see the exact frequency graph of them.
On the net I found this (see below), but it doesn't work for me. I tried
it with a gauss FIR-filter but the frequency respondse plot I get is a
flat line, everything but the gaussian graph I expected.
taps = [ 6.455906007234699e-014 1.037067381285011e-012
1.444835156335346e-011 ... 6.455906007234699e-014 ]
(43 elementen of gauss filter)
plot(taps) <<<--- gives plot of FIR (in this case the gauss-curve), dus
dat is goed
H_fir = poly(taps,'z','coeff') / %z^ ^ 43
H_fir.dt = 48000 <<<<---- (48 Khz sampling)
bode(H_fir,1,48000,4096)
For programming, C is OK for me too.
I know this is quite of-topic in this list, so I do want to keep this
discussion as limited as possible. But questions in (e.g.) the scilab
user list remained unanswered. (I think there are not that much
DSP-oriented people on that list).
That seems unlikely. Although most college students use matlab for DSP,
a lot of professionals use scilab. The college students use matlab
because they get it cheap, and the toolboxes let them do their
coursework with very little thinking. Matlab is the opium of the DSP
classes.
Well, I did ask a pretty question in the scilab "users" list, but did
not get any reply.
The 2nd week of may, I will be giving a talk of "DSTAR and GMSK" in our
local radio-club here in Ostend. Actually, it's more or less the content
of the "what does a D-STAR signal really look like" blog-article I once
wrote (which explains the format of a D-STARE stream); plus some
information trown in about GMSK-modulation and demodulation; and also a
couple of words on audio-vocoder. (based on the talk that David have).
The goal is to show that digital communication isn't "black magic" at
all, and it actually a subject that most people DO can understand.
One of the things I like to show is the frequency spectrum of a 4800 bps
FSK-signal ("0101...") modulated with and without a gauss-filter applied
on the input.
The goal is to -as you can image- to show that a gaussian filter limits
the bandwidth of a transmitted FSK signal by "rounding off" the input
signal that goes into the FM transceiver.
But I want something visible as that is the best way to explain this.
I was hoping a tool like scilab would be ideal for this; but sofar
nobody has given any hint on how to do this. I was a bit surprised as I
would image tools like scilab are actually designed for this, no?
In the end, I must just write a small C application to just create two
audio-files (once with and one without gaussian filtering) and then use
audacity to create a frequency-plot of them.
Steve
73
Kristoff - ON1ARF
------------------------------------------------------------------------------
Better than sec? Nothing is better than sec when it comes to
monitoring Big Data applications. Try Boundary one-second
resolution app monitoring today. Free.
http://p.sf.net/sfu/Boundary-dev2dev
_______________________________________________
Freetel-codec2 mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
