In this paper, important improvements in sound level meter (SLM) calibration are presented. One of the most important problems that the calibration laboratories are facing today is the long time required for calibrating certain SLMs. Instruments without a dc output require that an operator reads the SLM response in every test. This makes the process slow and tedious, and induces errors due to the operators fatigue. By using an image acquisition card with a low cost analog camera to acquire an image of the sound level meters display, calibration time can be reduced significantly. Once an image of the SLMs display has been acquired, machine vision techniques (OCR) are used to recognize the sound level meter indication. This process reduces calibration time and errors produced by operators fatigue.