We present a method to measure the Hubble parameter $H(z)$ and the angular diameter distance $D_A(z)$ simultaneously from the two-dimensional matter power spectrum from galaxy surveys with broad sky coverage. We validate this method by applying it to the LasDamas mock galaxy catalogs. Then we apply this method to Sloan Digital Sky Survey (SDSS) Data Release 7 and obtain measurements of $\Omega_mh^2=0.1268 \pm 0.0085$, $H(z=0.35)=81.3\pm 3.8$km/s/Mpc, $D_A(z=0.35) = 1037\pm44$Mpc, without assuming a dark energy model or a flat universe. We also find that the derived parameters $H(0.35)r_s(z_d)/c=0.0431 \pm 0.0018$ and $D_A(0.35)/r_s(z_d)=6.48 \pm 0.25$. These are in excellent agreement with similar measurements from the two-dimensional correlation function of the same data.

We measure the effective distance to z=0.35, D_V(0.35), from the overall shape of the spherically-averaged two-point correlation function of the Sloan Digital Sky Survey (SDSS) Data Release 7 (DR7) luminous red galaxy (LRG) sample. We find D_V(0.35)=1428_{-73}^{+74} without assuming a dark energy model or a flat Universe. We find that the derived measurement of r_s(z_d)/D_V(0.35)=0.1143 \pm 0.0030 (the ratio of the sound horizon at the drag epoch to the effective distance to z=0.35) is more tightly constrained and more robust with respect to possible systematic effects. It is also nearly uncorrelated with \Omega_m h^2. Combining our results with the cosmic microwave background and supernova data, we obtain \Omega_k=-0.0032^{+0.0074}_{-0.0072} and w=-1.010^{+0.046}_{-0.045} (assuming a constant dark energy equation of state). By scaling the spherically-averaged correlation function, we find the Hubble parameter H(0.35)=83^{+13}_{-15} km s^{-1}Mpc^{-1} and the angular diameter distance D_A(0.35)=1089^{+93}_{-87} Mpc. We use LasDamas SDSS mock catalogs to compute the covariance matrix of the correlation function, and investigate the use of lognormal catalogs as an alternative. We find that the input correlation function can be accurately recovered from lognormal catalogs, although they give larger errors on all scales (especially on small scales) compared to the mock catalogs derived from cosmological N-body simulations.