Due to standing waves, the sound pressure in a room may vary as much as 20-30 dB at low frequencies, which complicates measurement of noise. For assessment of annoyance, mainly the high-level areas of the room are of interest. These areas can however be difficult to locate, as different frequencies produce pressure maxima in different positions in the room. To obtain insight in the behaviour of sound in rooms, sound pressure distributions in rooms were investigated by simulations and scanning measurements. In addition measurements were performed in three-dimensional corners, as well as according to current Swedish and Danish guidelines. The level that is only exceeded in 10% of the space of a room is proposed as a reasonable target for a measurement method. The Swedish method showed results close to the target, but, due to a doubtful use of C-weighting in the scanning, it may give results below the target in case of complex sounds. The Danish method was found to have a high risk of giving results substantially below the target, unless complainants can precisely appoint measurement positions, where the sound is loudest/most annoying. An alternative method using measurements in four three-dimensional corners of the room is proposed. This easy and straightforward method seems to give reliable results close to the proposed target. Supported by STVF/FTP.