Moving Average: Difference between revisions

From GNU Radio
Jump to navigation Jump to search
Line 13: Line 13:


; Length  (''Interactive'')
; Length  (''Interactive'')
: The size of the moving average window to use. Can be changed at runtime.
: The size of the moving average window to use.


; Scale  (''Interactive'')
; Scale  (''Interactive'')
: Factor to scale the sum of the last (Length) samples. To get an actual moving average, the scale should then be set to 1/Length. Can be changed at runtime.
: Factor to scale the sum of the last (Length) samples. To get an actual moving average, the scale should then be set to 1/Length.


; Max Iter
; Max Iter

Revision as of 14:23, 11 July 2019


Computes a moving average of the input:

Output[i] = scale * sum(input[i-length: i])

In its default parameters, the block actually computes a moving sum.

Parameters

Input Type
Complex, Float, Int or Short
Length (Interactive)
The size of the moving average window to use.
Scale (Interactive)
Factor to scale the sum of the last (Length) samples. To get an actual moving average, the scale should then be set to 1/Length.
Max Iter
The maximum number of samples the block will treat in one call of its work function. Larger numbers can improve throughput at the cost of latency and potential numerical instability with float or complex input.
Max Iter can be smaller than Length without issues.
vlen
Used if the input samples are vectors and corresponds to the length of those vectors.
When operating on vectors, the average is done using only the numbers from the same vector index:
Output[i][vector_index] = scale * sum(Input[i-length: i][vector_index])

Example Flowgraph

Source Files

C++ files
All inputs
Header files
All inputs
Public header files
All inputs
Block definition
GRC yaml