The baselining procedure tries to eliminate backgound noise (Fig. 1).
The baselining process reduces each data point in the data by the local minimum value. The 'baselining width' defines the number of data positions that are to be considered for searching
the local minimum (+ the central data point). The value at the actual data point position is set to
its original value minus the minimum value.
For example, a baselining width of 20 lets the algorithm search for the smallest Y-value out of the
actual data point position, 10 data points to the left and 10 data points to the right of the actual
data point position.
The influence of different baselining widhts to the original trace data (Fig. 1) is shown in Figs. 2-6.
Tab. 1.
|
Setting
|
Effect
|
|
If the baselining width
value is too large ...
|
... the background will not be eliminated efficiently because
the minima found during the baselining process are not due to
the base points of the local peaks.
As a result the peaks will remain on a background noise.
|
|
If the baselining width
value is too small ...
|
... severe artifacts are generated!
If the baseline width is smaller than the width of a peak, the
local minimum values in the center of a peak will be detected
in the peak flank regions.
This leads to split peaks (see example with baselining width
10 and 20 below).
|
Tab. 2. Influence of the baselining width on the appearance of the final data.
|
Format
|
Example
|
|
Fig. 1. Smoothed raw data
The raw data has already been
smoothed.
The peaks are sitting on a
background as indicated in green
color.
|
|
|
Fig. 2. Baselining width set to 10
Here the baselining width value is too
low. This results in split peaks and
drastic reduction of the peak heights.
Too low!
|
|
|
Fig. 3. Baselining width set to 20
Here the baselining width value is still
too low. Both peaks are still split their
height is too small.
Still too low!
|
|
|
Fig. 4. Baselining width set to 50
The entire background is removed.
The peaks are not split and they sit
on the baseline.
Best!
|
|
|
Fig. 5. Baselining width set to 100
No split peaks. Peaks sit on base
line. Background removed.
Still good!
|
|
|
Fig. 6. Baselining width set to 500
The baselining width value is too
great and, therefore, the peaks do not
sit on the baseline.
Too great!
|
|
