OBJECTIVE : To evaluate light transmittance as a function of wavelength for eight composite materials and compare the transmittance for blue light produced from two curing units with different emission spectra. METHODS : Light transmittance through 2- and 4-mm-thick composite specimens was recorded in real time during 30 s of curing using a broad-spectrum (peaks at 405 and 450 nm) and a narrow-spectrum (peak at 441 nm) LED-curing unit. The spectral resolution of 0.25 nm and temporal resolution of 0.05 s resulted in a large amount of light transmittance data, which was averaged over particular spectral ranges, for the whole measurement period. Statistical analysis was performed using Welch ANOVA with Games-Howell post hoc test, t test, and Pearson correlation analysis. The level of significance was 0.05 and n = 5 specimens per experimental group were prepared. RESULTS : Light transmittance varied as a function of wavelength and time, revealing significantly different patterns among the tested materials. Light transmittance for different parts of curing unit spectra increased in the following order of emission peaks (nm): 405 < 441 < 450. Of particular interest was the difference in transmittance between 441 and 450 nm, as these peaks are relevant for the photoactivation of camphorquinone-containing composites. A high variability in light transmittance among materials was identified, ranging from statistically similar values for both peaks up to a fourfold higher transmittance for the peak at 450 nm. CONCLUSIONS : Each material showed a unique pattern of wavelength-dependent light transmittance, leading to highly material-dependent differences in blue light transmittance between two curing units. CONCLUSIONS : Minor differences in blue light emission of contemporary narrow-peak curing units may have a significant effect on the amount of light which reaches the composite layer bottom.