Smoothing
|
Previous Top Next |
Setting
|
Effect
|
If the smoothing width
value is too large
|
... the peak data will be washed away.
|
If the smoothing width
value is too small
|
... the noise is not efficiently removed. This may result in a
detection of noise fluctuations as peaks.
|
Description
|
Graphics
|
Raw data
The image to the right shows a
close up of raw trace data. Two
local maxima (peaks) are visible.
Note, that there are two problems:
The trace curve runs in a zigzag
manner and the peaks are sitting on
a background noise.
|
|
Smoothing width set to 2
Here a smoothing width value of 2
has been applied.
The zigzag Pattern is still clearly
visible. So the noise fluctuations are
not efficiently removed. The
smoothing width value is too low.
Too low!
|
|
Smoothing width set to 6
With zigzag is weaker but still
present. This data could be used for
peak detection.
It would be better to adjust a little bit
greater smoothing width value.
Good!
|
|
Smoothing width set to 24
The zigzag almost completely
vanished but the peaks remain
clearly visible. This seems to be the
best adjustment of the smoothing
width value.
Best!
|
|
Smoothing width set to 50
Here the smoothing width value is
too great. The peaks are washed
away! So you will loose valuable
information.
Too high!
|
|
Smoothing width set to 100
The same as above: Smoothing
width value is too great. The peaks
are eliminated!
Much too high!
|
|