Joint Baltic-Nordic Acoustics Meeting 2004, 8-10 June 2004, Mariehamn, Åland BNAM2004-1 PREDICTION OF SPEECH TRANSMISSION INDEX IN OPEN-PLAN OFFICES Valtteri Hongisto, Jukka Keränen, Petra Larm Finnish Institute of Occupational Health Laboratory of Ventilation and Acoustics Lemminkäisenkatu 14-18 B, 20520 Turku, Finland valtteri.hongisto@ttl.fi 1. INTRODUCTION According to questionnaire studies, speech is the most annoying sound source in open-plan offices [1]. The most distracting speech comes from the nearest workstations. Therefore, speech privacy between workstations should be as high as possible, that is, the speech transmission index, STI, should be low. This presupposes that the speech is properly attenuated, e.g. using absorbents, screens, and masked by using sufficient but not annoying background noise. Also the mutual distances and orientations of the workers are important. There are practically no regulations regarding the acoustical design of open-plan offices in Finland. Conventionally, HVAC designers aim at 35 dBA regarding HVAC noise and architects choose the absorption materials mainly by their visual properties. Office screens are seldom higher than 130 cm. As a result, workers complain about the speech disturbance, both from close and distant workstations. The acoustical reasons for the complaints are insufficient speech attenuation and speech masking. In single rooms, speech noise is not the most disturbing noise source because the speech levels are, inherently, lower and one can reduce the noise by closing the door, whenever needed. Continuous distraction due to speech leads to reduced work performance and productivity of work. Therefore, the need for better acoustical design is evident. At the moment, there are no simple models, which could help the architects in room acoustical design of open- plan offices. The aim of this study is to present a simple and validated model that can predict the STI between two nearby workstations and facilitate the all-inclusive design of open-plan offices. 2. METHODS 2.1. The model Speech intelligibility was described using the Modulation Transfer Function method, which gives the STI, or the rapid speech transmission index, RASTI. The value of STI can be within 0 and 1. High value of STI means high speech intelligibility and vice versa. E.g. in classrooms, high STI is desired, but between workstations in open-plan offices we choose an opposite scale. We should aim at low STI value, that is, high speech privacy. The prediction of STI presupposes that the signal-to-noise ratio of speech and the early part of the local reverberation time at the listener's location are known. The speech level propagating different paths and masking noise level from different sources are calculated using conventional room acoustical principles. The speech and masking noise paths are presented in Figure 1. Speech through the screen (1), reflected speech via ceiling (2), diffracted speech over the screen (3) and reverberant speech (4) were involved. Masking noise sources comprise, e.g., ventilation, office equipments, and artificial masking system (5-7). The full description of the theory is given in our recent paper [2] of which a summary is given below.