Population growth and urbanization are projected by the United Nations to add 2,5 billionpeople to the world’s urban population by 2050. We need to construct buildings in anunprecedented scale to meet global housing demand. Sustainable development is critical.Compared to traditional heavy constructions, lightweight wooden constructions are moreenvironmentally friendly and will play a key role in meeting future demands. However, thereare two major problems with lightweight constructions that need to be addressed: 1) Limitedlow frequency sound insulation and 2) Variations in sound insulation.Annoyance from walking sound tend to be higher in lightweight than in heavy constructionseven with the same measured sound insulation. The Swedish research program AkuLiteindicated that the correlation between measured sound insulation and annoyance wassignificantly improved by extending current evaluation methods from 50 Hz down to 20 Hz.Secondly, large variations in sound insulation between nominally identical lightweightconstructions are common, which leads to larger safety margins. By identifying and quantifyingunderlying causes, production costs can be minimized and the performance can be improved.The aim of the thesis is to develop a new evaluation method for impact sound insulation thatbetter correspond to rated annoyance, and to identify and control underlying causes forvariations in sound insulation. The thesis contains six papers.In Paper I and II, sound insulation measurements were carried out in a large number ofnominally identical rooms of two different industrially prefabricated lightweight woodenconstructions. The purpose was to assess and quantify the variations in impact and airbornesound insulation. In Paper I, 30 nominally identical apartments of a volume based system wasevaluated. The apartments on the highest floor achieved significantly better sound insulationdue to the extra weight on lower floors affecting the elastic connections between stories. InPaper II, 18 rooms of a cross-laminated timber system of plate elements were evaluated.Additionally, several potential parameters related to measurement uncertainty wereinvestigated.Paper III deals with measurement uncertainty. An empirical study of reverberation timemeasurements showed that current methods need to be improved, if sound insulationrequirements are to be extended to 20 Hz.Paper IV and V verified that the frequency range 20-50 Hz is important for walking soundannoyance, and that alternative frequency adaptation terms can improve the correlation betweenmeasured impact sound insulation and annoyance ratings. In Paper IV, the methodology was toperform extensive field measurements in apartment buildings of various construction types andto perform questionnaire surveys among the residents. In Paper V, the methodology was toevaluate annoyance based on binaural recordings of walking sound in a two-part listening test.In Paper VI, 70 measurements in a lightweight wooden system were evaluated to quantify thetotal variations in impact and airborne sound insulation from 20 and 50 Hz, respectively. It wasconcluded that the proposed metrics of impact sound insulation were primarily determined bythe impact sound level 20-40 Hz and that the measurement methods must be evaluatedthoroughly to avoid excessive safety margins.

A new evaluation method for impact sound insulation from 25 Hz, that correspond to the ratedannoyance for both heavy and lightweight constructions is proposed. By using the proposedmethod and attending the specific causes for variations, the lightweight industry will be able todevelop improved multi-story dwellings with higher perceived acoustic quality.