Deriving stellar properties from photometry: maximizing information content and minimizing biases

I study the importance of the accurate calibration of photometric systems in order to produce meaningful comparisons between the observed colors + magnitudes and model SEDs. Possible sources of errors are discussed and two examples are analyzed. I show that well-calibrated Tycho-2 photometry is stable and precise enough for such comparisons. On the contrary, the available calibrations for Johnson UBV photometry yield relative large systematic errors, which has prompted me to develop a new, more precise calibration. The advantages of multicolor photometry over the standard single-color + magnitude diagrams for the derivation of physical properties of stars (elimination of degeneracies, inclusion of multiple parameters, avoidance of linearizing approximations, possibility of a more precise treatment of errors) are discussed through the use of CHORIZOS, a code developed specifically for this purpose.