function [diam,area,trolands] = PupilDiameterFromLum(lum,source) % [diam,area,trolands] = PupilDiameterFromLum(lum [, source]) % % Compute pupil diameter and area from photomic luminance. % Diameter is in mm, area in mm^2. % Luminance is in cd/m2. % Also returns photopic trolands. % % Source (string): % PokornySmith: (default) % Formula is Eq. 1 from: Pokorny and Smith, "How much light % reaches the retina", Colour Vision Deficiences XIII (C. % Cavonius, ed.), pp. 491-511. % % DeGrootGebhard: % Formula is De Groot and Gebhard's from % Eq. 2(2.4.5) of Wyszecki and Stiles, % 2cd edition (page 106). % % MoonSpencer: % Formula is Moon and Spencer's from % Eq. 1(2.4.5) of Wyszecki and Stiles, % 2cd edition (page 106). % % Notes: % a) The calculations of the DeGroot/Gebhard formula do not seem to agree with the % same calculations as expressed in Figure 2(2.4.5) on the same page of W+S. One would % need to go back to the original literature to sort out what is going on. % % b) In terms of the different methods, Joel Pokorny (1999, personal communication) says: % The average pupil diameter/luminance functions in the literature vary enormously. % This can be seen in the figures in % % Moon, P. and D. E. Spencer (1944). "On the Stiles-Crawford Effect." % Journal of the Optical Society of America 34: 319-329. % % and % % de Groot, S. G. and J. W. Gebhard (1952). "Pupil size as determined % by adapting luminance." Journal of the Optical Society of America % 42: 492-495. % % For example, the Reeves (1918, "The visibility of radiation." % Transactions of the Illuminating Engineering Society 13: 101-109) % pupil diameter function is displaced roughly 1.5 log units higher on % the luminance axis than Crawford's (1936, "The dependence of pupil % size upon external light stimulus under static and variable conditions." % Proceedings of the Royal Society B (London) 121(B): 376-395) average data. % % Both Moon and Spenser & DeGroot and Gebhard sought functions which were % compromises between existing data sets. LeGrand's function shows good % correspondence with the Reeves' data. These three functions nominally % describe pupil behavior for binocular view of large fields. In vision % science we most frequently use fields of limited extent and often use % monocular view. These stimulus manipulations lead to larger pupils than % the binocular large field condition. Thus it made sense to me to use the % LeGrand function. As is mentioned in "How much light..." pupil size % varies for all sorts of reasons and any estimate should be viewed as % having a large tolerance. % % 4/2/99 dhb Wrote it. % 5/8/99 dhb Consolidated different methods. % 7/8/03 dhb Accept strings without dashes. % 12/4/07 dhb Added dog case, with a place holder number of 8 mm. % Set default methods if (nargin < 2 || isempty(source)) source = 'PokornySmith'; end % Get diameter according to chosen source switch (source) case {'PokornySmith', 'Pokorny_Smith'} diam = 5 - 3*tanh(0.4*log10(lum)); case {'DegrootGebhard', 'DeGroot_Gebhard'} diam = 10.^(0.8558-4.01*1e-4*((log10(lum)+8.6).^3)); case {'MoonSpencer', 'Moon_Spencer'} diam = 4.9 - 3*tanh(0.4*(log10(lum)+1)); case 'PennDog' diam = 8; end % Compute ancillary information area = pi*(diam/2).^2; trolands = lum.*area;