Abstract
Four different methods of measuring speech and noise level in speech audiometry are compared. The methods differ with regard to the temporal characteristic of the integrator used for determining the level. The four methods are characterized by their so-called detector/indicator characteristics, being ‘fast’, ‘slow’, ‘impulse’, and long-term integration. It is concluded that the method using long-term integration is preferable. If the speech material consists of short speech segments such as isolated words separated by pauses, the long-term rms level should be measured without integrating over the pauses between the speech segments.
Key Words: