One of the advantages of a source method over observer method in blur perception measurements is better control of a stimulus blur level, which is achieved with computerised image processing unlike the observer method that requires optical defocusing of the observer. The aim of this study was to determine if paralysation of accommodation has effect on blur perception thresholds, thereby evaluating its necessity in such experiments. Blur perception thresholds (just noticeable blur, clear image, recognition, and non-resolvable blur thresholds) were evaluated with (using cycloplegia) and without paralysed accommodation to determine effect on blur perception. A computerised low-pass spatial frequency filter was used to process images of Landolt rings and gradually increase the blur level from the stage of clear image to the stage of unrecognisable orientation of Landolt ring opening. Ascending and descending psychophysical methods were used to evaluate the previously mentioned blur perception thresholds. Results for 10 participants revealed no significant effect of cycloplegia on any of the determined blur perception thresholds. Therefore, it may be considered that blur perception measurements with source method can be done without cycloplegia.