With the implementation of the Double Hot Thermocouple Technique (DHTT) it is possible to simulate service conditions, which occur during the continuous casting of steel inside the mold. For that purpose a slag film is stretched between two U-shaped thermocouples, which are heated to different temperatures. Therefor an investigation of the crystallization behavior of transparent slags depending on different temperature gradients is possible. In this thesis, an H-shaped one replaced one of the U-shaped thermocouples, so that a thin 3mm long slag film could be stretched in between, in order to analyze transparent slags, as well as translucent ones. The crystallization of two slag types was investigated, evaluated and the results were presented graphically in such a way, that a comparison of the crystallization behavior of these slags was possible. The first series of tests were carried out using a congruent melting 3K2O∙11GeO2-slag having a transparent liquid phase and opaque crystals. The second ones were carried out using an incongruent melting, synthetic, fluorine-free slag which only appeared transparent in the area around the stretching device, the rest however appeared only translucent. In order to carry out isothermal experiments, the H-shaped thermocouple was heated to 1150°C when using the 3K2O∙11GeO2-slag or to 1300°C for the use of the F-free slag and held at that temperature. The furnace chamber together with the stretching device was set to 600°C for the first experiment, for further ones to 700°C, 800°C and 900°C, respectively. 3 experiments were carried out for each temperature gradient. After stretching of the slag film a ten minute picture recording started. For data analysis 30 pictures were used out of it and inserted into an image editing program, where the area in which crystallization can occur was framed in red and the crystals were colored in blue. These color-coded images were divided into sections of a defined width of 0.1mm, cut out separately and with the help of an interface detection program, which detects the red and blue areas, the percentage of the crystallized area in each part was calculated. In order to compare the crystallization behavior of different slags, three modes, in which the results could be displayed, have been developed. For mode 1, the maximum crystallinity per section and temperature gradient was determined and a graph was drawn as the absolute crystallinity in dependence on the position within the slag film. For mode 2 the absolute mean values of the experimentation time were standardized to 100% and a diagram was drawn as the crystallinity in dependence on the relative time per section and temperature gradient. Out of this, the percentage of the absolute crystallinity when 95%, 50% and 0.5% of the absolute experimental time passed by was read off and displayed in dependence on the position within the slag film. For mode 3 instead of the time values, the absolute mean values of the crystallinity were standardized to 100% and the time in seconds after which 95%, 50% and 0.5% of each section was crystallized was read off and displayed again in dependence of the position within the slag film. Comparing all of the three modes to each other, mode 2 would be the least suitable. This method is dependent on the actual end time of the experiment and on the time when the last crystallization occurred. On the contrary, mode 3 together with mode 1 is very useful for the comparison of the crystallization behavior of different slags.