SIMBAD references

We consider three corrections to the disk sizes estimated at a given frequency using accretion models. They are due to a color correction, a disk truncation at an inner radius larger than the innermost stable circular orbit, and disk winds, which we apply to the standard disk model. We apply our results to the estimates of the disk sizes based on microlensing. We find that these three effects combined can explain the long-standing problem of the disk sizes from microlensing being larger than those estimated using the standard disk model (i.e., that without accounting for the above effects). In particular, an increase of the color correction with the increasing temperature can lead to a strong increase of the half-light radius even if this correction is close to unity at the temperature corresponding to an observed frequency. Our proposed formalism for calculating the half-light radius also resolves the long-standing issue of discrepancies between the disk size estimates based on the accretion rate and on the observed flux.