[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Discuss-gnuradio] The power of N log N
From: |
Matt Ettus |
Subject: |
Re: [Discuss-gnuradio] The power of N log N |
Date: |
Tue, 14 Jun 2005 09:40:10 -0700 |
User-agent: |
Mozilla Thunderbird 1.0 (X11/20041206) |
Lamar Owen wrote:
> Yeah, been tracking CVS on a machine here, and noticed a load of differences.
>
> One quick question: now that tweak_adc_offset is no longer in the repository,
> is the ADC DC offset automatic? I still see a DC spike in the FFT output.
> In the large FFT case, I notice that the resulting trace is significantly
> thicker, and it doesn't seem to average out. I'll have to look, but an
> 'fft_smoothing' argument (to thin the trace) would be what I'd need to get
> the resolution I need.
What you are seeing is that ALL points end up being plotted. If you
assume you have a horizontal screen resolution of 1024 point, but are
plotting a 128 K-point FFT, you end up with 128 points being plotted in
every X position. Each of those points is likely to be slightly
different in amplitude, so you see a thick line. Some sort of "video
averaging", to borrow a term from spectrum analyzers, would be necessary
to clean that up.
Matt