2005MNRAS.357..174D

I report on the capabilities of the near-infrared (near-IR) surface brightness technique to predict reliable stellar angular diameters as accurate as ≲2 per cent using standard broad-band Johnson photometry in the colour range -0.1 ≤ (V-K)_O≤ 3.7 including stars of A, F, G, K spectral type. This empirical approach is fast to apply and leads to estimated photometric diameters in very good agreement with recent high-precision interferometric diameter measurements available for non-variable dwarfs and giants, as well as for Cepheid variables. Then I compare semi-empirical diameters predicted by model-dependent photometric and spectrophotometric (SP) methods with near-IR surface brightness diameters adopted as empirical reference calibrators. The overall agreement between all these methods is within approximately ±5 per cent, confirming previous works. However, on the same scale of accuracy, there is also evidence for systematic shifts presumably as a result of an incorrect representation of the stellar effective temperature in the model-dependent results. I also compare measurements of spectroscopic radii with near-IR surface brightness radii of Cepheids with known distances. Spectroscopic radii are found to be affected by a scatter as significant as ≳9 per cent, which is at least three times greater than the formal error currently claimed by the spectroscopic technique. In contrast, pulsation radii predicted by the period-radius (PR) relation according to the Cepheid period result are significantly less dispersed, indicating a quite small scatter as a result of the finite width of the Cepheid instability strip, as expected from pulsation theory. The resulting low level of noise strongly confirms our previous claims that the pulsation parallaxes are the most accurate empirical distances presently available for Galactic and extragalactic Cepheids.