SIMBAD references

We present a new, fast and easy to use NLTE line formation code for ``unified atmospheres'' with spherical extension and stellar winds, developed for the (routine) spectroscopic analysis of luminous blue stars, covering the spectral range from ``A'' to ``O'' and including central stars of planetary nebulae. The major features of our code are: Data driven input of atomic models; consistent photospheric stratification including continuum radiative acceleration and photospheric extension; ``β-velocity law'' for the wind; comoving frame or Sobolev plus continuum line transfer; fast solution algorithm for calculating line profiles, allowing for a consistent treatment of incoherent electron scattering. We describe the code and perform thorough tests for models with H/He opacity, especially with respect to a comparison with plane-parallel, hydrostatic models in cases of thin winds. Our conclusions are: Due in particular to our numerical treatment of the radiative transfer in the ionization and recombination integrals, the convergence rate of the solution algorithm is fast. The flux conservation is good, (maximum flux errors of order 2 to 3%), unless the atmospheric conditions are extreme, either with respect to mass-loss or to a large extension of the photosphere. (In these cases, our treatment of the temperature structure has to be improved). A comparison with plane-parallel results shows perfect agreement with the thin wind case. However, this comparison also reveals two interesting effects: First, the strength of the Hei lines in hot O-stars is very sensitive to the treatment of electron scattering in the EUV. This might affect the effective temperature scale of early O spectral types. Second, the effects of photospheric extension become decisive for the gravity determination of stars close to the Eddington limit. Finally, we demonstrate the differences in using the Sobolev vs. the comoving line transfer in the rate equations. We conclude that, in cases of moderate wind densities, comoving frame line transfer is inevitable for accurate quantitative work.