Reconstruction in multispectral optoacoustic tomography has become an critical area of importance, given the development of real-time imaging and visualization techniques. Speed of sound calibration is an intrinsic problem associated with the reconstruction process. Traditionally, the calibration has been user mediated, making it a tedious and offline affair. In this manuscript, we aim to introduce autofocusing and wavelet based measures to automatically calibrate the speed of sound. Further, it is observed that the temperature of the coupling medium (water) often drift during the signal acquisition, severely straining the image quality. The measures address these problems by iteratively determining the speeds with the changing boundary conditions with time.