Microplasmas in liquid with microsecond and nanosecond duration are generated by the application of fast rise time voltage pulses to a micro-tipped-electrode immersed in the liquid with single and double spark gap switching respectively. The time dependent light emission occurring during the microplasma's generation in the liquid with different energy inputs are investigated in this research. Salt water is used as the liquid for microplama generation with different salt (NaCl) concentrations. The size of the microelectrode tip, applied voltage waveform, and conductivity determines the energy input and energy density to initiate microplasma at the tip. The microplasma light emissions are detected at high time resolution in two ways. First via a microscope based optical system in a dark environment. The microsecond and nanosecond duration of microplasma light emission is captured by a high speed gated ICCD camera. From the photos taken in the dark environment, the light emission is clearly recorded with different delay and exposure times of the ICCD. Using backlighting, variation of densities commensurate with the microplasma formation is detectable. A second method for light detection employs a photomultiplier tube (PMT) with optical fiber and confirms the ICCD camera results with regards to light emission. Though not spatially resolved the PMT results have the advantages of: 1) being from a single discharge rather than multiple discharges with variable gate delay as in the case of the ICCD; and 2) not requiring complicated pre-triggering schema to overcome inherent ICCD triggering delays and measure the earliest phases of plasma initiation. However, significant efforts (including multiple layers of shielding and long fiber optic lengths) had to be employed to decrease electromagnetic interference (EMI) in the PMT signals. The light emission from microplasma experiences a gradually brighter and then gradually dimmer process. Based on the voltage pulse duration and energy inputs, the duration of light emission increases correspondingly. As energy to initiates discharge continued decreasing, the light intensity captured by ICCD and PMT also reduced. A threshold energy is deduced, under which no light emission is detected and no microplasma will be initiated. However, the energy released from such voltage appears to induce density variations, i.e., the microbubble generation. These results indicate that, at least within the range of rise times and conductivities tested, density variations (i.e. microbubble generation) precedes plasma generation. © 2013 IEEE.