Issue link: https://digital.copcomm.com/i/1521235
M OT I O N P I CTU R E S O U N D E D I TO R S 75 a sound object with only their eyes whilst keeping their head position static, binaural cue anomalies can be introduced leading to a divergence of aural and visual localisation. Better externalisation during binaural sound reproduction was observed in virtual rooms that offered visual congruence with audio localisation exposing the participants to a VR wasp with the headphone audio rendered in different modes. The binaural audio was shown to be capable of significantly increasing self-reported anxiety. Another internal study demonstrated that externalisation can reduce auditory fatigue among call centre operatives. The relatively narrow field of view offered by current MR headsets mitigates this to a degree by 'forcing' the head to turn to bring emitters into the line of sight, and eye tracking within an HMD is becoming more commonplace too, although a physical HMD has been shown to degrade audio localisation. However, the deployment of spatial audio provides information about events beyond the field of vision and in virtual environments, spatialised audio significantly increases a listener's sense of presence. Audio Cognition There are challenges of auditory- scene analysis, amongst which are understanding the human assimilation of sequential and concurrent sonic elements to form a perceptual stream, and then segregating actual conflicting simultaneous sources or temporal sound patterns—to actually identify them. It was demonstrated that binaural listening can mitigate the 'cocktail-party effect'—source determination amongst numerous competing audio stimuli compared to conventional playback using a single headphone. Using over-ear headphones, speech intelligibility over noise has been shown to be as effective in binaurally decoded first-order Ambisonics (FOA) as in higher orders inferring that the higher channel count of the latter is unnecessary for intelligibility, although it might still be preferable if higher spatial resolution is required. Dynamic placement of the audio resulted in increased spatial release from masking for listeners, offering enhanced speech intelligibility when spatialised against noise. Some research compared between simultaneous versus sequential presentation of multiple audio streams, spatially localised audio sources versus a single point of the presentation, and dynamic movement versus fixed locations of sources. The investigation revealed a trend for participants preferring the spatially fixed yet also noted that the use of spatial audio improved recall against single-point presentation with simultaneous streams. Dynamic binaural synthesis with head movement can compromise the perceived authenticity of audio, although this was most noticeable for pink noise, in preference to music and mitigated by reverberation. VR users typically underestimate distance and increasing the reverberation time associated with an audio object can help to mitigate this, although, in the near field, this can cause audio-visual sensory segregation. However, due to its pass-through nature, in MR, users have a precise visual perspective and expect audio cues to match, and techniques are being developed to facilitate this further. If the reverberation introduces interaural differences beyond those of the source object, externalisation is improved. Some research suggests that details of the personalisation of HRTF's and how they relate to localisation resolution are a critical component of binaural audio, and they determine much of the quality of any binaural 3D sound that we experience. Human pinnae are as unique as fingerprints and therefore, HRTF's are highly idiosyncratic as well. Mixed-Reality Acoustic Modelling Spatial mapping—whereby a system can analyse the physical space around it—has been evolving for quite some time. For example, Magic Leap and Hololens headsets offer a contemporary implementation, first modelling static elements of the surroundings with point cloud world features, and then with world models, increasing detail with meshes and detecting planes. For example, tabletops or wall surfaces that are useful for augmenting with virtual objects. The current implementation cannot associate audio reflections with individual planes, although the application programming interface is evolving and an environmental reverberation model is being parameterised. For example, to characterise the reverb gain, the total energy of the reverb when the listener and the source are collocated. This has been extended by the addition of clustered reflections to provide distance simulation for individual audio objects, allowing 6DoF navigation in the Metaverse. Now, let's look at the technical