Existing methods for glare-free daylighting design rely on analyses of physically based lighting simulations employing tools such as Radiance. The rendered image is an accumulation of luminance values from a fixed point of view, which creates a basis for luminance-based metric analysis such as discomfort glare models. A major challenge using luminance images for glare prediction analysis lies in the image processing steps for deriving different parameters of the predictive discomfort glare metrics. One of these challenges is to define and identify glary pixels and zones in the image. The glare source detection algorithms adopted in evaluation tools like evalglare, search for pixels of luminance value that are x-times larger (we call it threshold multiplier in the following) than the average luminance of a pre-defined zone, e.g. the monitor screen. As a second step, the potential glare pixels detected within a search radius (also predefined) are considered and evaluated as one glare source. The implemented default value of a threshold-multiplier and the search radius are decided intuitively and so far no validation has been made on their accuracy. The objective of the present study is to establish a sensitivity analysis on the threshold and search radius parameters. In two series of experiments we took luminance images very 30 second under different lighting conditions and with different façade systems. We also gathered the participants’ subjective assessments of the glare conditions using a Likert scale. Thereafter, we processed the images using 15 different combinations for the threshold multiplier and search radius (threshold multiplier with 5 and search radius with 3 treatment levels). The preliminary results show that there is a significant effect of threshold on all lighting conditions and an effect of search radius on lighting conditions with shading systems. But the results also show, that the choice of the ”right” detection parameters underlies also some effects, which cannot be explained so far. The consequence of this is, that further investigations are needed and as long as no clear rule exist to choose ”right” detection parameters the user of glare evaluation tools has to check the detected glare sources for each investigated situation.