Research areas
Building Acoustics
Based on the software BASTIAN® from DataKustik a software was developed, to auralise airborne sound insulation in buildings.
Screenshot of BASTIAN®
At first the user of the software creates a room situation by selecting the direction of sound transfer (horizontal, vertical or outdoors to indoors) and the building products for the partition and flanking walls.
The calculations in BASTIAN® are based on EN 12354 parts 1-3: "Building acoustics - Estimation of acoustic performance of buildings".
All kinds of sound transmitting components and systems being of importance for the particular room situation are considered. These are, for instance, partitioning and flanking ceilings, walls and roofs, wall linings windows and doors and also cable and ventilation ducts.
The picture below shows the resulting sound insulation being calculated from the above room situation:
BASTIAN® Resulting diagram
The Situation in the receiving room is considered by the dimensions and the reverberation time. Volume and reverberation time can be selected by the user from a list of 9 given rooms of which the binaural room impulse responses were measured.
BASTIAN® auralisation dialog
For every sound radiating component a filter is calculated from the sound insulation. A head related transfer function, a delay and a weighting is added according to the direction and the distance from the listener to the radiating component is added. This filter builds the direct sound component the listener perceives. To calculate the reverberation the direct sound of the real binaural room impulse response is cutted out and the remaining reverberation tail is filtered with the energetic sum of the 5 filters. Afterwaards the direct sound is cutted out and replaced with the newly calculated direct sound.
A convolution of a selected source signal with this filter results in the new timesignal in the listening room.
Such a calculation including the convolution lasts shorter than 10 seconds on a modern PC (circa half of the real time of the resulting signal length). FFT convolution is used which is mainly programmed in Assembler.
Listening examples :
The following examples were calculated with the software. Because the levels of resulting signals in the listening room are ca. 40 dB below the level in the source room, you should adjust the volume of your system so, that the levels given for the source room are reached during listening. Otherwise you wouldn't hear anything. Replay by earphones is mostly suited. The files are in WAVE-format.
The files above each are of ca. 1.3 MB size. Alternatively you can download them in MPEG Layer 3 format (ca. 250 kB). You need a suitable software (e.g. WinAmp) resp. a browser plugin.