The concept of reproducing sound with more than two loudspeaker setups is notnovel. It has been a while since the first time surround reproduction systems havebeen installed for domestic and automotive use, with the 5.1 setup being the mostfrequently used 1.
The evolving technology has also yielded audio signal transmis-sion with broad spectral bandwidth, highly effective spatial sound recording tech-niques and configurations incorporating multiple loudspeakers. For instance, in 7.1the angle is decreased among loudspeakers, whereas as mentioned in 2, elevatedloudspeakers are introduced in systems like 9.1 11.1.
In addition, the fast paced de-velopment of virtual reality video has emerged the need for 3-D audio reproduction.Spatial sound has also gained ground in critical applications associated with routeguidance of people having seriously defective vision 3 and with flight navigationsystems 4. It is apparent that, spatial sound systems have to be able to interact with dif-ferent audio channel formats, because it’s not always technically-and financially-advantageous to use recording techniques, typical for every reproduction system11.
Thus, compatibility turns out to be a very important issue. Besides, the vastmajority of commercial music is recorded in stereo format. Therefore, certain signalprocessing techniques have to be employed, in order to convert the stereo recordingsto multichannel formats, maintaining the quality of audio information and enhancingthe listening experience. Extensive research has been conducted for spatial audio processing 6, 7 and spatialaudio coding, which comprises spatial audio scene coding (SASC) 5 and directionalaudio coding (DirAC) 8. The aim is to decompose the input audio signal into non-diffuse and diffuse sound.
The calculated diffuse sound in the channels differs forpositive and negative correlations, which deviates from the ambience definition givenin the primary-ambience extraction framework investigates in the present thesis. InSASC, the the Gerzon localization vector 9 is used to carried out localizationanalysis, separately for both the separated components. In DirAC, the primarysound is rendered applying vector base amplitude panning (VBAP) 10, while thediffuse sound is decorrelated and played back by rear loudspeakers. The aim forthese signal processing techniques is to achieve spatial sound reproduction withevery sound system configuration