Overview of Time–Frequency Domain Parametric Spatial Audio Techniques

Spatial sound reproduction is mainly concerned with delivering appropriate spatial cues using a reproduction system, loudspeakers or headphones, in order to provide a convincing, immersive, and engaging aural experience, close to the intentions of the producer for a synthetic sound scene, or to the original experience for a recorded one. State‐of‐the‐art parametric methods should be flexible, able to handle a variety of input formats, either multichannel content or recordings coming from a variety of recording setups. Many methods in enhancement, up‐mixing, and reproduction of recordings impose a model of one dominant directional source and some diffuse/ambient sound at each time‐frequency tile. The extracted spatial parameters allow the whole arsenal of audio engineering tools for perceptual sound spatialization to be used optimally for a variety of target setups, such as appropriate panning functions, head‐related transfer functions (HRTFs), and reverberation algorithms.