Jones et al (2006, MNRAS, 369, 25) present optical and near-IR galaxy luminosity functions from a flux-limited sample of 138226 galaxies from the 6-degree Field Galaxy Survey. Here are the interesting results:
"a Schechter function is unable to decline rapidly enough at the bright end and remain as flat at the faint end [of the galaxy LF]... We do not find as steep a faint-end slope as Kochanek et al. (2001), and suspect this is due to their shallower depth and subsequent brighter faint-end limit. The 6dFGS rF-band LF most closely matches those of the Las Campanas Redshift Survey (Lin et al. 1996) and SDSS (Blanton et al. 2005), although we find only marginal evidence for the faint-end upturn claimed by Blanton et al. (2005). In bJ, the 6dFGS LF has a nearly identical faint-end slope to those obtained by 2dFGRS (Norberg et al. 2002) and the ESO Slice Project (Zucca et al. 1997), although the 6dFGS normalization is closer to that found by Blanton et al. (2005). Neither this survey nor 6dFGS used evolutionary corrections. Furthermore, we see no evidence for a faint-end upturn in any of the 6dFGS LFs. "The faintest and brightest ends of the LF are where galaxy formation models have the most difficulty, and where "feedback" (either from supernovae or AGN) is often invoked as the solution. The figure below is Fig 10 from the Jones et al paper.
Figure 10. LFs for the 6dFGS, derived from the 1/Vmax (green open circles), SWML (red solid circles) and STY methods (blue dashed curve). The inset shows the 1, 2 and 3σ confidence contours of the STY fit. The upper panel shows the 1/Vmax and SWML residuals relative to STY (i.e. the deviations from the best-fitting Schechter function).
Jones, D. Heath, Peterson, Bruce A., Colless, Matthew & Saunders, Will
Near-infrared and optical luminosity functions from the 6dF Galaxy Survey.
Monthly Notices of the Royal Astronomical Society 369 (1), 25-42.
Its nice to see we don't have to worry about large numbers of faint galaxies. But it doesn't appear to answer the question I was most interested in: what is the actual shape of the LF if its not a Schechter function?