Send to

Choose Destination
Klin Monbl Augenheilkd. 1998 Apr;212(4):190-5.

[Determining visual acuity using European normal values: scientific principles and possibilities for automatic measurement].

[Article in German]

Author information

Universit├Ąts-Augenklinik Freiburg.



According to the European standard EN ISO 8596 the Landolt-C in 8 different orientations has to be used to measure visual acuity. With decreasing size of the Landolt-C the hit rate declines from 100% to the chance level of 12.5%. This gradual transition is described by the "psychometric function". The steepest point of the psychometric function is in the middle between 100 and 12.5, i.e., at 56.25%. This point of the psychometric function (approximated by 5 of 8 Landolt-Cs) has been selected as the threshold for visual acuity, because it is there that the visual acuity is influenced least by (incidental) fluctuations. The subject has to answer by forced choice; a response like "I cannot detect anything" is not acceptable. "NORMAL" VISUAL ACUITY: Cannot be assigned to a certain value, like 1.0 or 6/6. With the standard test procedure, visually healthy, young subjects achieve a visual acuity of about 2.0 or 12/6, while in senior subjects 0.5 (3/6) may be "normal".


Logarithmic, not arithmetic, scaling of visual acuity approximates the perceptual metric. Consequently, visual acuity values may not be averaged arithmetically. Instead, three steps are required: all values have to be converted to logarithms, then averaged, and finally the average can be reconverted. Geometric averaging is equivalent. "MINIMUM ANGLE OF RESOLUTION" NOT NECESSARY: MAR is the reciprocal of visual acuity. In many studies, clinical outcome has been assessed using log(MAR). Though statistically correct, this term is unnecessary, as log(acuity) has identical statistical properties. Furthermore, log(MAR) is contra-intuitive as its value becomes smaller when vision improves.


Facilitates complying with the EN ISO 8596. For instance, the Freiburg Visual Acuity Test relieves the examiner from observing whether 5 responses have been correct, and that not more than 8 tests are given per level.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Georg Thieme Verlag Stuttgart, New York
Loading ...
Support Center