The luminosity calibration of the HR diagram revisited by Hipparcos

The HR diagram luminosity calibration basically relies on trigonometric parallaxes and kinematical data (proper motions and radial velocities). Hipparcos data allow to re-calibrate the absolute magnitude for groups of stars selected in the HR diagram according to spectroscopic criteria. A new Maximum Likelihood luminosity calibration algorithm (the LM method) developed by Luri et al. (1996) has been applied to the Hipparcos survey stars. This algorithm uses all the available parallax and kinematical data and takes into account the sample biases and the observational errors. We have estimated the accuracy of the calibrations based on the MK classification and found that the relation between absolute magnitude and luminosity class has a large intrinsic dispersion. Given a sample, the LM method not only provides a luminosity calibration relationship but also an unbiased estimation of the individual absolute magnitude and distance for each star.