3.3. H sensitivities and error budgets of the continuum subtraction
From the direct comparison of the distribution of eDIG in the galaxies in our survey, with those already observed by Lehnert & Heckman (1995) and Rand (1996), we can conclude that we have reached similar and in a few cases somewhat better sensitivities. This is mostly true for the galaxies, which were obtained with DFOSC at La Silla. The Calar Alto observations, however, have been acquired through relatively wide H filters in several cases ( ~ 168Å), so the sensitivity is not as good as for the La Silla observations, which were made through H filters of ~ 62Å. The estimated mean sensitivities of the galaxies observed with DFOSC are of the order of a few cm-6 pc.
Typically there are some uncertainties associated with the scaling procedure, to be applied for the continuum subtraction. The scaling factor has been determined from the ratio of the H and continuum countrates of individual stars in the object frames. Of course, there arise uncertainties as the actual continuum spectrum of a galaxy will be different from that of individual stars, as the galactic continuum is the superposition of all stars of the underlying stellar population. However, practice has shown, that this is a good method and the uncertainties typically are of the order of 5-10%. A careful analysis, however, is necessary in order to match the ideal scaling factor. The determined factor from the intensity ratio of the stars in the H and continuum frame has been slightly changed in a few cases where the derived scaling factor did not seem to match perfectly, as artifacts were observed in the bulge region, which is a good indicator for the accuracy of the scaling process. Then the galaxy profile was analyzed carefully, as the scaling factor was changed in very small steps to find the optimal (in a somewhat conservative manner) value, not to under- and oversubtract the emission. This is a well established procedure (for more details see e.g., Hoopes et al., 1999; Rossa & Dettmar, 2000).
Alternatively, it is possible to determine the scaling factor using continuum regions in the galaxy. This is something which would work much better for face-on galaxies, but is somewhat more difficult for edge-on galaxies, as here occasionally the prominent dust lane makes it difficult to assess whether there is underlying emission, which is strong enough to be seen after subtracting the continuum. Furthermore, the true H emission is only revealed after the continuum subtraction process. Hence, knowing regions free of emission is somewhat difficult to judge, prior to the subtraction process. However, testing this method, applied to a few individual galaxies, have shown that there is a good agreement between using stars or continuum light within the galaxies for the determination of the scaling factor.