I'm planning to write my program from scratch in C#. I understand the basics of the FFT now; I had always wondered how transient signals would show up in a window if they were significantly shorter than the time covered by an FFT window, and I see now that there is a tradeoff: more samples = higher frequency precision but fewer FFT windows can be drawn. I've observed this in HDSDR where RBW can be either frequency-precise or do fast windowing. But as everyone says, the FFT is tremendously complex, so I will probably use a 3rd-party library. I'm currently working out how to use the AForge .NET library.> Date: Tue, 22 Mar 2016 09:20:41 +0100 > From: vitt...@gmail.com > To: discuss-gnuradio@gnu.org > Subject: Re: [Discuss-gnuradio] Handling of IQ files > > Happy to read you replay Henry! > Feel free to rip/change/upgrade my work, but let me know about ur progress, > pse! > Remember that it's fairly resource hungry ( file larger that 250 Mb > crashes the app also on my I7/16Gb), but it's a starting point. > I think that with GNURADIO it's possible to refine this task... I've a > lot to study! > > Victor I3VFJ > > 2016-03-21 20:45 GMT+01:00 Henry Barton <kw...@outlook.com>: > > I like the concept of your program. It looks just like what I’m trying to > > write. > > > > Sent from Windows Mail > > > > From: Vitt Benv > > Sent: Sunday, March 20, 2016 4:16 PM > > To: discuss-gnuradio@gnu.org > > > > Hi, > > this [ https://sourceforge.net/projects/automodrecog/ ] is my little > > effort about handling IQ files. > > The input IQ file is recorded with HDSDR, very nice piece of sw, that > > as a good recording scheduler. By the way the file provided can be > > played with it. I do also some tests with IQ file produced by R&S > > EM100 receiver. > > I wrote this [horrible] application for personal use and it's very raw > > IMO.... > > I'm interested on HF monitoring and the main goal is to find SSB > > emission along time, not quite simple task. > > > > In brief: > > - read params from input file > > - split it in smaller chunk and save FFT for each chunk > > - sum (maxhold) or avg (average) all the FFTs > > - find relevant ( over threshold) carrier and try to "pack" them to > > find "bandwidth" associated > > - build a report as .html page with a .png file that represent the result. > > > > The most difficult part is to estimate the best "threshold", and at > > the moment I'm almost stuck there.... and moreover RL reclaims my > > presence :-) > > > > ... my euro "cent" on the subject. > > > > Victor I3VFJ > > > > _______________________________________________ > > Discuss-gnuradio mailing list > > Discuss-gnuradio@gnu.org > > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio > > > > _______________________________________________ > > Discuss-gnuradio mailing list > > Discuss-gnuradio@gnu.org > > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio > > > > _______________________________________________ > Discuss-gnuradio mailing list > Discuss-gnuradio@gnu.org > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
