discuss-gnuradio
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Newbie Question: Linking with GnuRadio Libraries


From: Martin Dvh
Subject: Re: [Discuss-gnuradio] Newbie Question: Linking with GnuRadio Libraries
Date: Sat, 29 Jan 2005 16:59:02 +0100
User-agent: Mozilla Thunderbird 0.9 (X11/20041124)

Bill Peter wrote:
 > (1) Can I use the GnuRadio libraries to link in my own "driver" c++
code to do the same thing I do with Matlab--but now with GnuRadio?
Yes, I think you could.

(2) How would I do this?  Can I just write a c++ program that makes
use of your Quadrature Demodulation class to do it for me:  Would it
look something like this:

//use GnuRadio libraries
my_demod = new VrQuadratureDemod( xxx ) ;

and link in with the appropriate libraries?  Does anyone have any
simple examples/templates and a Makefile ?
It is easier if you just use gnuradio for all the processing.
If you have the captured raw data already on your harddisk (shorts or floats) You could start rightaway using gnuradio blocks and input your data as a file_source (see example at end of mail)
If you want to do realtime analysing you would need to write your own gnuradio 
driver.

I had a somewhat similar thing I wanted to do.
I wanted to use standard videocapture cards as ADC and do some processing.
What I did was.
1. use the standard linux kernel video4linux driver.
2. Take some code of an existing videocapture example program
3. Take the mc4020 gnuradio "driver" (really is and interface to the mc4020 
kernel driver)
4. Replace the code which interfaces with the mc4020 kernel driver with my 
video4linux videocapture code.

Now I can use the new driver within gnuradio programs.
I didn't link in any libraries but of course you could link in any needed 
libraries in you new driver.

This way you can write any gnuradio code (python) and use your new 
driver/library and any existing gnuradio blocks.
Just look at all the gnuradio examples in the WIKI and in gnuradio-example in 
CVS.

If you don't want to write your program in python in the gnuradio framework, you could also just link in the gnuradio libraries and use them directly. This would mean some more work because the python interface is an integral part of gnuradio. Some parts are written in python and I don't know if it is easy to interface them back to C++ or objective C.
the recent java interface discussion on this mailinglist could be of interest 
to you for this. (Search the mailinglist archives for OSGi)

All above is only valid for gnuradio 2.x. The old gnuradio 0.x code can be used directly from C++ but I don't think there is any development going on in gnuradio 0.x.

Greetings,
Martin


from gnuradio import gr
from gnuradio import audio
#from gnuradio import mydriver
import sys

def build_graph ():
        sampling_freq = 10e6
        cfir_decimation = 100
        fg = gr.flow_graph ()

        #src0 = mydriver.source_f(0)
        src0 = gr.file_source (gr.sizeof_short, "captured_10.7Mhz_data.raw", 0)
        # compute FIR filter taps
        #(lowpass cutoff and width of transition band determine the with of the 
bandpass channel filter)
        channel_coeffs = \
                gr.firdes.low_pass (
                  1.0,                  # gain
                  sampling_freq,
                  50e3,                 # low pass cutoff
                  400e3,                 # width of transition band
                  gr.firdes.WIN_HAMMING )

        # input: short; output: complex
        chan_filter1 = \  # do downconversion(decimation) and bandpass channel 
filtering in one step
                gr.freq_xlating_fir_filter_scf (
                  cfir_decimation,#decimation_factor
                  channel_coeffs,
                  10.7*1e6, # wanted center frequency 10.7 MHz
                  sampling_freq )

my_processing = gr.complex_to_mag () #This is just an example, complex_to_mag would do an am-demodulation of a complex signal, you could add any processing block you want here

        dst = gr.file_sink (gr.sizeof_float, "processed_signal_out.raw")
        #dst = audio.sink ( 32000 )

        fg.connect ( src0, chan_filter1 )
        fg.connect ( chan_filter1,my_processing )
        fg.connect ( my_processing, dst )

        return fg

def main (args):

        fg = build_graph()
        fg.start()
        raw_input ('Press Enter to quit')
        fg.stop()

if __name__ == '__main__':
        main (sys.argv[1:])






reply via email to

[Prev in Thread] Current Thread [Next in Thread]