PTB-News 1/2000 (373 kB) PTB-News 1/2000, english edition, April 2000PTBnews2000 1d (307 kB) PTBnews 1/2000, Deutsche Ausgabe, April 2000To a steadily increasing extent, computer programs are used to predict the acoustic qualities of rooms already at the planning stage. However, the results obtained in this way are still afflicted with shortcomings. By performing objective tests of this kind of software, PTB supports the further development of the software and helps with the valuation of the program’s capabilities.
Within the scope of an international intercomparison coordinated by PTB, the quality of software for the simulation of room acoustics has recently been tested by 16 participants from nine different countries. The computer programs used had to prove their efficiency on a concrete object: the participants had to calculate the acoustic properties of a Swedish concert hall and compare them with independent measurement results determined by PTB and two other teams.
For various combinations of sound transmission and auditor positions, nine parameters of room acoustics, such as the reverberation time and the definition, had to be calculated in six octave bands. The results were compared with the measured values. The comparison revealed that the deviations are clearly greater than the difference which is subjectively perceptible, in particular in the lower frequency range (125 Hz octave). The large errors at low frequencies are due to a general deficiency of the simulation programs which cannot yet take diffraction effects into account. Possible applications of the software are, therefore, restricted to obstaclefree, large rooms. Another factor which influences all computer programs is the strong dependence of the results on the input data, for example, the acoustic absorption properties of the surfaces. In most cases these are known only as estimates.
Both software developers and software users can gain by a comparison of this type. Comparing their results with those of other simulation programs and with thoroughly measured data, software developers can identify sources of errors and further improve their simulation programs. On the other hand, the comparison helps acousticians, architects and civil engineers to assess the quality of the software they use.
The data of the Swedish concert hall will be made available at the PTB WWW server within the next months. Along with the room data and the measured values, recorded real sound patterns will be supplied which allow to test room-acoustics simulation software by means of a so-called auralisation: Sound signals recorded with low reverberation are convolved with the impulse response of the room and conditioned for earphone reproduction like an artificial head recording.