WEBVTT

00:00:00.250 --> 00:00:07.694
The probe study is independent of the active analysis. In this
simulation, we search near four point seven parts

00:00:07.694 --> 00:00:12.787
per million using a fixed local window and monitor each
spectrum's original maximum.

00:00:13.787 --> 00:00:21.474
Six groups remain visible: three usable ramps and three
diagnostic rows. Calculate all slopes lets us compare

00:00:21.474 --> 00:00:28.257
the fits while each acquisition retains its conditions. Review
the chosen calibration row without averaging

00:00:28.257 --> 00:00:30.518
different diffusion times or ramps.

00:00:31.518 --> 00:00:39.816
Refine R² targets 0.9999 while retaining enough points and
gradient range. This example proposes twenty-two

00:00:39.816 --> 00:00:47.560
points instead of twenty-three, excluding point eleven. The
proposed selection still needs your review.

00:00:48.560 --> 00:00:55.696
Apply selection changes the included points for this row and
invalidates its previous calibration. Measured

00:00:55.696 --> 00:01:02.831
values remain unchanged. Review the slope and conditions, then
evaluate the reference again before saving.

00:01:03.831 --> 00:01:11.243
Undo refinement restores the previous selection. This screenshot
follows the real undo action, with all

00:01:11.243 --> 00:01:18.654
twenty-three original points included again. An optimized R² does
not demonstrate a physically valid probe

00:01:18.654 --> 00:01:19.148
calibration.

00:01:20.148 --> 00:01:26.920
The simulation uses pure water at twenty-five degrees Celsius and
a compatible literature diffusion coefficient.

00:01:26.920 --> 00:01:34.595
Evaluation derives the gradient from the reviewed fit and
conditions. This demonstration does not save or share

00:01:34.595 --> 00:01:35.950
an experimental probe.