Sonars are used in a wide range of marine applications. The advantage of using sound waves in water has made the sonar technique ideal to detect underwater objects, imaging of the seafloor and detection of vessels. Sound waves can propagate over long distances in water unlike other waves. The ability to see underwater using acoustic waves has made the sonar technique important for marine and military industry. Most sonars uses the far-field when imaging. To characterize a sonars ability to depict one can measure the sonars far-field characteristics. This is usually done in big water tanks because the far-field boundary often is several meters away. This thesis investigates the possibility of characterizing the sonar using near-field measurements. This is done with simulations in matlab and measurements on a Simrad SH-90 transducer in a [ ] water tank at the University in Oslo. This was the first time measurements like these were done in the water tank so development of hardware and procedures were necessary. A method of simulating the wave field with parameters as close to the conditions in the water tank as possible was developed. The result of simulations showed that the system was sensitive to changes in the sound velocity. We found that it was possible to characterize a transducer if we had a calibrated system were the only unknown parameter is the speed of sound. By using the least square method of errors it was possible to characterize a transducer by finding the elements weights and phase. This gave us a basis for trying to match up the data in the measurement experiment. When trying to match up the simulated and measured data we found that there were too many unknowns in the measuring experiment. Due to time limitations of this thesis we were not able to do an optimal calibration of the measuring system. The work on this thesis showed us that it was possible to characterize a transducer in the simulations but more work has to be done on the measuring system.