Most quantitative phase microscopy methods require the use of custom-built or modified microscopic configurations which are not typically available to most bio/pathologists. There are, however, phase retrieval algorithms which utilize defocused bright-field images as input data and are therefore implementable in existing laboratory environments. Among these, deterministic methods such as those based on inverting the transport-of-intensity equation (TIE) or a phase contrast transfer function (PCTF) are particularly attractive due to their compatibility with Köhler illuminated systems and numerical simplicity. Recently, a new method has been proposed, called multi-filter phase imaging with partially coherent light (MFPI-PC), which alleviates the inherent noise/resolution trade-off in solving the TIE by utilizing a large number of defocused bright-field images spaced equally about the focal plane. Despite greatly improving the state-ofthe-art, the method has many shortcomings including the impracticality of high-speed acquisition, inefficient sampling, and attenuated response at high frequencies due to aperture effects. In this report, we present a new method, called bright-field quantitative phase microscopy (BFQPM), which efficiently utilizes a small number of defocused bright-field images and recovers frequencies out to the partially coherent diffraction limit. The method is based on a noiseminimized inversion of a PCTF derived for each finite defocus distance. We present simulation results which indicate nanoscale optical path length sensitivity and improved performance over MFPI-PC. We also provide experimental results imaging live bovine mesenchymal stem cells at sub-second temporal resolution. In all, BFQPM enables fast and accurate phase imaging with unprecedented spatial resolution using widely available bright-field microscopy hardware.