Alex Deligiannis wrote:
Dear Alberto,
V1.130 running here OK but I have a question about
Spectran V1 which I use more than Argo. I found a difference of 110 to 120 mH
between Argo readings and Spectran as follows:
MSF > Argo 799.73 Hz Spectran 799.84 Hz
HBG > Argo 799.71 Hz Spectran 799.84 Hz and
DCF77 > Argo 799.71 Hz Spectran 799.84 Hz (all CW mode 800 HZ pitch)
I'm wandering why I have this difference between the two programs and second
how I can calibrate spectran.
Dear Alex,
there are a couple of things to consider. First, the program
settings. The FFT bins are
spaced according to the FFT length and sampling rate chosen, and the
peak indicated
is that of the nearest bin. So you should compare Argo and Spectran at
the same FFT
length and sampling rate. Spectran reports in its control panel those
values, Argo doesn't.
Argo, when in horizontal mode, uses 5512 as sampling rate and FFT
lengths as follows :
QRSS3 : 16384
QRSS10 : 65536
QRSS20 : 131072
QRSS30 : 262144 etc. etc.
But there is a second point to consider : when I said that the nearest
bin is used for
indicating the peak frequency, I cheated a little. That is true for
Spectran.
In Argo I tried to be a bit smarter... the peak indicated is computed
with a
quadratic interpolation of three adjacent bins, the highest magnitude
one and its two
neighbours, using their relative magnitudes as weights.
So, all in all, may be Argo is a tad more precise than Spectran, at
least at comparable
resolutions.
As for calibration, the current version of Spectran doesn't have that
facility. May be
in a forthcoming one...
73 Alberto I2PHD
|
