Looking through the crowded mask: investigating the effect of distractor number and position in object substitution masking