Validity of digital subtraction of transcranial plain films in quantification of positional changes of the mandibular condyle.