Three indirect light measurement methods were compared in mixed deciduous and coniferous forests with heterogeneous stand structure: tRAYci — a spatially explicit light model calculating percentage of above canopy light (PACL); LAI-2000 Plant Canopy Analyzer measuring diffuse non-interceptance (D1FN); and spherical densiometer estimating canopy openness (CO). Correlations between the different light variables were analyzed at several spatial scales (at 5 × 5, 10 × 10, 15 × 15, 20 × 20 and 30 × 30 m 2 ). Relationships between light variables and the cover of alight flexible plant, blackberry ( Rubus fruticosus agg.), as a potentially sensitive response variable for light conditions were also investigated. LAI-2000 (D1FN) and tRAYci (PACL) seemed the most appropriate for the description of the light environment in the investigated stands. D1FN and PACL had stronger correlations with each other and with blackberry cover than CO. Spatial heterogeneity of light (expressed with coefficient of variation) showed much stronger correlations than mean values both between the methods and between light intensity and Rubus cover. The correlation values between the methods increased towards coarser scales (from 5 × 5 to 30 × 30 m 2 ), while the correlation between light intensity and blackberry cover had a maximal response at the scale of 20 × 20 m 2 if a lower resolution of light estimation was used, and had also a maximum at smaller scales if the light was calculated for more points per plot by tRAYci. LAI-2000 can be recommended for the comparison of different stands, however, for fine scale description of light conditions of a stand tRAYci seems to be more appropriate.