In SDR# (SDR Sharp), dynamic averaging is a running average of successive FFT spectra. It smooths noise and stabilises the display, but it is not a fixed integration time.
Instead, it behaves like an exponential (IIR) integrator:
- Each new FFT contributes only part of the result
- Older data fades away gradually
So you get an effective integration time rather than a strict one.
Approximate relationship:
Let:
- f_FFT = FFT update rate (spectra per second)
- alpha = averaging factor (between 0 and 1)
Then:
T_int ~ 1 / ( (1 – alpha) * f_FFT )
What this means:
- If alpha is close to 1
-> long effective integration
-> slow response
-> smoother (less noise) - If alpha is small
-> short integration
-> fast response
-> noisier
Example:
If:
- f_FFT = 10 spectra/sec
- alpha = 0.9
Then:
T_int ~ 1 / (0.1 * 10) = 1 second
If alpha = 0.99:
T_int ~ 1 / (0.01 * 10) = 10 seconds
Important limitation:
Dynamic averaging in SDR#:
- does NOT give a fixed integration window
- does NOT preserve total accumulated power
- only affects the displayed spectrum
Bottom line:
Dynamic averaging is an exponential smoothing filter. As alpha approaches 1, the effective integration time increases, but it never becomes a true fixed-duration integration.