Auditory Stimulation For Virtual And Augmented Reality · And Augmented Reality Why it’s...
Transcript of Auditory Stimulation For Virtual And Augmented Reality · And Augmented Reality Why it’s...
Auditory Stimulation For Virtual And Augmented Reality
Why it’s Important to Get it Right & How to Get it Right
Scott Isabelle, Knowles ElectronicsRobert Gilkey, Wright State University
Acknowledgement: Brian Simpson, US Air Force Research Labs
Key Functions of Hearing• Social/Communication
• Speech• Music
• Alerting/Orienting• Warning• Directing attention
• Connectedness• Monitoring• Presence
Deafness as a Model of VR Without sound
Ramsdell (1978) – War veterans with sudden, profound hearing loss• Problems associated with loss of communication abilities were often reported
• But felt that loss at a more primitive level was more important
• Patients often reported that they did not feel part of the world around them
• That the world seemed dead or unreal• That the progression of time was altered
after Gilkey & Weisenberger (1995)
Presence• Importance varies with application• In general, we prefer not to get comments about our VRs such as:
• It seems dead• It seems unreal• I don’t feel like I’m a part of it
• Note deaf individuals make these types of comments to describe an environment (i.e., the real world) in which vision, touch, olfaction, and taste are perfectlyrendered
• a standard we’re unlikely to achieve in virtual displays in the foreseeable future
after Gilkey & Weisenberger (1995)
Auditory Spatialization• Creating an immersive auditory environment to support presence
• Separating sound sources (e.g., voices) to support detection, recognition, and intelligibility (the cocktail party effect)
• Alerting and directing attention to significant objects and events
• Creating a realistic and externalized background
Auditory Localization Cues• Interaural Time Differences (ITDs)• Interaural Level Differences (ILDs) (Head Shadow)• Monaural Level (Head Shadow)• Monaural Spectral Shape (Pinna)• Monaural Spectral Shape (Shoulder/Torso)• Binaural Spectral Shape (Pinna)• Binaural Spectral Shape (Shoulder/Torso)• Dynamic cues from head and/or source movement• Distance
• Level• Floor Echoes• Direct‐to‐Reverberant Ratio• Vocal Effort• ILD for near‐field sources• Atmospheric absorption
Auditory Localization Cues• Interaural Time Differences (ITDs)• Interaural Level Differences (ILDs) (Head Shadow)• Monaural Level (Head Shadow)• Monaural Spectral Shape (Pinna)• Monaural Spectral Shape (Shoulder/Torso)• Binaural Spectral Shape (Pinna)• Binaural Spectral Shape (Shoulder/Torso)• Dynamic cues from head and/or source movement• Distance
• Level• Floor Echoes• Direct‐to‐Reverberant Ratio• Vocal Effort• ILD for near‐field sources• Atmospheric absorption
How to create spatial audio
Wenzel (1992)
‐40º 40º
90º‐90º
0º
HRIRs: Azimuth
Zahorik (personal communication)
‐40º 40º
90º‐90º
0º
HRTFs: Azimuth
Zahorik (personal communication)
HRTFs: Elevation from spectral cues
Simpson (2002)
Bandwidth
Brungart & Simpson, 2009Brungart & Simpson (2009)
• Localization degrades as bandwidth is reduced• and Localization is disrupted by noise
• Errors are nearly twice as large with non‐individualized HRTFs Brungart, Romigh, and Simpson (2009)
PERSONALIZATON OF DISPLAY
• Sense of Presence reduced with non‐individualized HRTFs Valjaamae, Larsson, Vastfjall, & Kleiner (2004)
Head‐tracking improves• Localization of virtual sources• Externalization of virtual sources
Head movements
Brimijoin et al. (2013)
Required Fidelity of Auditory Scene
• For randomly distributed sources, you can get by with much lower‐complexity systems.
• Having more to render requires proportionately fewer processing elements
Acceptance rate is:• Proportional to order
of the system• Sensitive to number
of sources and clustering
• Highly variable across subjects at lower orders
Wakefield & Roginska (2011)
Non‐auditory FactorsAudio‐Visual Match
• Recordings were judged to have greater realism when they were heard in the room where they were recorded
• Results depended on prior experience with the listening room.
Gilkey, Simpson, Weisenberger (2001)
SummaryVirtual audio
• Supports localization of auditory objects• Increases presence and realism of virtual environments• Improves intelligibility of separated voices
Requires adequate combination of • Spectral and binaural cue information (HRTFs)
• Improved experience/performance if individualized HRTFs• Bandwidth• Head‐tracking without excessive delay• Consistency of auditory scene with other modalities
• “Adequate” depends on application
Backup
Intelligibility
• Spatialized presentation improves intelligibility• Increasing performance advantage with increasingly complex environment
Simpson (2006)