Pulsed thermography (PT) is always considered as a quantitative testing approach in infrared thermal imaging. One of its main models called temperature-contrast function (TCF) method is usually found invalidated in defect depth measurement when meets a thin sample. To exhibit the existing fault during TCF modeling, one-dimensional finite analytical solutions were used to carry out calculations instead of the approximate solution adopted in original TCF model under an ideal impulse excitation. Two kernel parameters, named as peak temperature-contrast time (PCT) and peak temperature-contrast slope time (PST), were detailedly analyzed both in temperature-contrast curve and its derivative line. Comparison was made in their detecting abilities. The results show that PCT is governed by background-s depth, not by test point-s depth, so this parameter isn-t suitable for a quantitative test; and PST is also manifested to have service restriction in a depth ratio between the two points (test point/reference point). The upper limit of this depth ratio (approximately 0.5) was obtained by a criterion equation as a theoretical reference when PST was chosen in applications.