In this paper, the following four indicators describing noise annoyance in a community are considered: the percentage of people being highly annoyed (pHA), annoyed (pA), (at least) little annoyed (p LA), and the mean of the expected individual annoyance scores (m EA) in the population concerned. These indicators are calculated for a study population on the basis of maps of the noise measure DENL, using DENL-annoyance relationships. An approach is described for comparing the effect of noise reduction, using the above mentioned noise annoyance indicators. For the municipality of Leiden, the Netherlands, the results from applying this approach indicate that the calculated efficacy of a given noise abatement policy depends to some extent on the noise annoyance indicator that is used. A lower limit of 55 dB for the noise calculations has been suggested by the EU Directive on environmental noise. This leads to a large underestimation of the noise annoyance indicators and may affect priority setting regarding noise abatement. Beside the four noise annoyance measures, two other noise quality indicators are considered: the percentage of people being highly sleep disturbed (p HS) and (AREA50) - the percentage area where DENL exceeds 50 dB. The results for Leiden show that each of these two indicators put more focus on noise from motorways than the other noise indicators considered. © S. Hirzel Verlag.