Toggle search
Toggle menu
4.8K
5.5K
229
46.5K
VR & AR Wiki
Toggle preferences menu
Toggle personal menu
Not logged in
Your IP address will be publicly visible if you make any edits.

Head-related transfer function

From VR & AR Wiki
More actions
See also: 3D audio and VR audio

A head-related transfer function (HRTF) describes how sound travelling from a point in space is filtered by a listener's head, torso and outer ears (pinnae) before it reaches each ear. In the definition used by Li and Peissig in their 2020 review, an HRTF is "an acoustic transfer function between a point sound source in the free-field and a defined position in the listener's ear canal"; its time-domain form is the head-related impulse response (HRIR).[1] A complete HRTF set holds one filter per ear for each measured direction, and it carries the cues the auditory system uses to judge where a sound comes from.

HRTFs are the basis of binaural rendering in virtual reality and augmented reality. A spatial audio engine filters a mono sound with the left and right HRTFs for the source's direction relative to the listener's head and plays the result over headphones, typically while the listener wears a head-mounted display, so the sound appears to come from a fixed place in the scene.[2] Headset makers including Meta and Microsoft build HRTF processing into their platform audio software.[3][4] Because HRTFs differ from person to person, much current research is about measuring individual HRTFs quickly or estimating them without a full measurement.[1]

The broader practice of spatial sound in headsets, including room acoustics, ambisonics and output hardware, is covered in 3D audio and VR audio. This article covers the transfer function itself.

Reviewed 29 September 2026. Facts checked against the cited papers (Crossref/PubMed metadata and abstracts, full text of Li and Peissig, CIPIC, SONICOM and NASA 2010) and the cited Meta, Microsoft, Valve, Apple, Resonance Audio, SOFA and press pages. About review dates.

Definition

Li and Peissig describe the HRTF as a linear time-invariant system (an approximation under certain conditions) between a point source in the free field, meaning without room reflections, and a defined point in the ear canal. The HRIR is obtained from the HRTF by an inverse Fourier transform, and filtering a signal with an HRIR is a convolution.[1] Because the HRTF is defined without room information, reverberation has to be added separately; Meta's audio documentation notes that HRTFs "provide strong directional cues, but without room effects, they often sound dry and lifeless."[3]

Localization cues

According to Li and Peissig, all of the acoustic cues used to localize real sound sources are contained in HRTFs: interaural level differences (ILDs), interaural time differences (ITDs) and monaural spectral cues.[1] The Resonance Audio documentation explains the division of labor. The time difference between the two ears helps a listener judge the horizontal position of low-frequency sounds; for higher frequencies the head casts an acoustic shadow, and the horizontal position is judged from the level difference instead. The changes in frequency content caused by the outer ear, called spectral effects, are used to judge the vertical position of a source.[5]

Interaural cues alone cannot separate every direction. Sources lying on the same "cone of confusion" produce similar time and level differences, so front and back, or up and down, are told apart mainly by spectral cues.[6] Head movement helps: Meta's design guidance notes that "simply turning our head changes difficult front/back ambiguity problems into lateral localization problems that humans are better equipped to solve."[2]

Near field and far field

For sources farther than about 1 m, HRTFs are almost independent of distance, and distance is usually simulated by changing the level according to the inverse-square law. Closer than about 1 m, in the near field, the HRTF changes noticeably with distance, so rendering nearby sources needs distance-dependent HRTFs.[1] The Meta XR Audio SDK includes a near-field rendering mode that "approximates the effects of acoustic diffraction to create a more realistic representation of audio sources closer than 1 meter."[3]

Measurement

HRTFs are usually measured in an anechoic chamber, which approximates a free field. The subject wears a pair of in-ear microphones, a loudspeaker plays an excitation signal from a known position, and the transfer function is computed from the recorded and original signals; the process is repeated for every direction in the set.[1] Excitation signals include pseudo-random sequences (MLS and Golay codes) and linear or exponential sine sweeps.[1] Meta's developer guidance gives the same outline: put microphones in a subject's ears, just outside the ear canal, place the subject in an anechoic chamber, and play sounds from each required direction.[2]

Work by Møller and by Hammershøi and Møller showed that the direction-dependent part of the path from a source to the eardrum ends at the entrance of the ear canal, while transmission along the canal does not depend on direction. HRTFs of human subjects are therefore usually measured with microphones at the entrance of blocked ear canals (the "blocked ear technique"), which is easier than measuring deep in the canal.[1] Measuring a dense HRTF set one direction at a time is slow. Faster methods include interleaved and overlapped multiple sweeps played through many loudspeakers, and a "reciprocity method" that swaps the roles of source and receiver, placing miniature loudspeakers in the ears and microphones at the measurement positions, which can capture many directions within a few seconds; because miniature loudspeakers perform poorly at low frequencies and have a limited playback level, most laboratories still prefer direct measurement.[1]

HRTFs can also be computed. In 2001 Brian Katz used the boundary element method (BEM) to calculate part of an individual's HRTF from precise geometric data of the head and pinna, and compared the results with measurements of the same person.[7][8]

Storage format

HRTF sets are commonly exchanged in the Spatially Oriented Format for Acoustics (SOFA), a file format for spatial acoustic data such as HRTFs and binaural or spatial room impulse responses. The Audio Engineering Society standardized it as AES69-2015 and reaffirmed it as AES69-2020 and AES69-2022.[9] Valve's Steam Audio loads custom HRTFs from SOFA files as an alternative to its built-in HRTF.[10]

History

In 1974 E. A. G. Shaw brought together data from 12 studies, covering 100 subjects measured in five countries over a 40-year period, into average curves of the sound pressure transformation to the eardrum in the horizontal plane between 0.2 and 12 kHz.[11]

An early real-time HRTF renderer for virtual environments was developed at NASA Ames Research Center, where work on virtual environment systems had begun in 1983. Between 1986 and 1988 Elizabeth Wenzel of NASA Ames worked with Scott Foster of Crystal River Engineering on a prototype real-time virtual audio system called the Convolvotron. According to a 2010 retrospective by NASA Ames researchers, it could spatialize up to four sources at a 50 kHz sample rate using HRTF FIR filtering; its HRTFs covered 24 azimuths (every 15 degrees) on six elevation rings spaced 18 degrees apart, with interpolated values updated from a Polhemus Isotrak electromagnetic head tracker. The authors describe it as one of the first large-scale digital convolution engines fast enough for interactive audio.[12]

In 1989 Frederic Wightman and Doris Kistler published a two-part study that measured free-field-to-eardrum transfer functions for ten subjects at 144 source positions and used them to build headphone stimuli. Their simulations matched free-field ear-canal waveforms "within a few dB of magnitude and a few degrees of phase at frequencies up to 14 kHz."[13] In the second part, eight listeners localized the headphone stimuli in nearly the same positions as real loudspeakers, though with more front-back confusions and slightly less well defined elevation.[14]

Public HRTF datasets have been released since the 1990s. The SONICOM authors noted in 2023 that most public sets contained fewer than 100 human subjects, which is often too few for modern machine learning approaches, and built larger ones:[15]

Dataset Released Subjects Directions Notes
MIT Media Lab KEMAR[16] Measurements completed May 1994 One dummy head (KEMAR) 710 positions, elevations -40 to +90 degrees Loudspeaker 1.4 m from the head; maximum length sequences at 44.1 kHz; also published in JASA in 1995[17]
CIPIC HRTF Database (UC Davis)[18] 2001 45 (43 people plus KEMAR with large and small pinnae) 1,250 at about 5 degree spacing 200-sample HRIRs (about 4.5 ms) at 44.1 kHz; 27 anthropometric measurements per subject
SONICOM HRTF Dataset (Imperial College London)[15] 2023 120 Also headphone transfer functions, 3D scans of ears, heads and torsos, and depth pictures
Extended SONICOM HRTF Dataset[19] 2025 300 Adds synthesized HRTFs for 200 subjects generated with Mesh2HRTF and processed 3D head and ear scans

The CIPIC authors noted that most commercial HRTF systems of the time used a single "standard" HRIR, often from the public KEMAR data, although HRTFs "vary significantly from person to person."[18] The KEMAR, CIPIC and ARI (Acoustics Research Institute) databases are among the public sets Meta lists for developers.[2]

Generic and individualized HRTFs

An individualized HRTF is measured or modeled for one listener; a generic (non-individualized) HRTF comes from a dummy head or another person. Li and Peissig note that HRTFs are unique to each person because of individual anatomy, especially the shape of the pinna, and that non-individual HRTFs can reduce localization accuracy and externalization, the sense that a sound comes from outside the head.[1]

The trade-off was already being studied in the early 1990s. In a 1993 study by Wenzel, Arruda, Kistler and Wightman, 16 inexperienced listeners judged the direction of noise bursts rendered with the HRTFs of one representative subject. Once confusions were resolved, virtual sources were localized about as accurately as real ones for 12 of the 16 listeners, but many listeners showed high rates of front-back and up-down confusions, which increased significantly with the virtual sources. The authors concluded that interaural cues survive the use of another person's HRTFs, while the spectral cues that resolve position on a cone of confusion are distorted.[6] A 2001 NASA Ames study by Begault, Wenzel and Anderson compared head tracking, individualized HRTFs and simulated reflections for virtual speech, and found significant effects on azimuth and elevation error, reversal rates and externalization.[20]

Later studies used VR headsets directly. In a 2020 study published in JMIR Serious Games, Jenny and Reuter had 39 participants wearing VR headsets rate their own HRTFs against the MIT KEMAR HRTF and other people's HRTFs from the ARI database. Sounds filtered with the individual HRTFs were rated easier to localize and externalize, more natural in timbre and more realistic.[21] A 2016 conference abstract by Mehra, Nicholls, Begault and Zannoli reported benefits of personalized over generic HRTFs for listeners who could move their heads in virtual environments, a case earlier static-listener studies had not covered.[22] A 2025 audio augmented reality study by Vincent Martin and Lorenzo Picinali, which paired virtual speech sources with real loudspeakers, found that individual HRTFs improved perceived realism but not localization when the head was still, and that the opposite held when head movements were allowed.[23]

The auditory system can also adapt to unfamiliar cues. In a 1998 Nature Neuroscience study, Hofman, Van Riswick and Van Opstal changed subjects' pinnae with molds; elevation localization collapsed at first and then recovered steadily, and the subjects could still localize with their original ears afterwards.[24] Researchers at Microsoft Research tested the same idea in VR. Using an HTC Vive, Berger and colleagues found in 2018 that pairing a moving sound with a synchronized visual object improved later sound localization for users of generic HRTFs, while exposure to the sound alone or to unsynchronized audio and video did not; they concluded that generic HRTFs "may be enough to enable good auditory source localization in VR."[25]

Personalization methods

Because a full acoustic measurement is slow and needs special facilities, several shortcuts are used or under study:

Approach How it works Examples
Selection from a library The listener picks, or a system chooses, the closest HRTF from a set of measured profiles Sony created five HRTF profiles for the launch of the PlayStation 5 from data on more than 100 people[26]
Scaling from body measurements Anthropometric data, such as head size, are used to adjust a generic HRTF CIPIC database anthropometry;[18] HoloLens adjusts HRTFs for head size using the wearer's interpupillary distance[4]
Numerical simulation HRTFs are computed from a 3D model of the head and ears Boundary element calculations by Katz (2001);[7] Mesh2HRTF-synthesized HRTFs in the extended SONICOM dataset[19]
Camera scan A phone camera captures the head and ears to build a personal profile Apple Personalized Spatial Audio, set up with the iPhone TrueDepth camera[27]
Neural networks A model predicts an HRTF from sparse measurements, ear shape or images Survey by Lu et al. (2025);[28] generative model trained from sparse home measurements by Zandi et al. (2022)[29]

Lu and colleagues group recent deep learning work into two paradigms: explicit methods that predict a personalized HRTF and then use it in a conventional renderer, and end-to-end methods that map source signals directly to binaural output.[28] In 2020 Facebook Reality Labs Research, part of what is now Meta's Reality Labs division, said it hoped to "develop an algorithm that can approximate a workable personalized HRTF from something as simple as a photograph of [your] ears," describing an accurate individual measurement as needing "specialized tools and a lengthy calibration procedure."[30]

Applications in VR and AR

In a real-time renderer the HRTF must change as the listener's head turns. Tracking data from the headset gives the source direction relative to the head each frame, so the sound stays fixed in the virtual world as the user moves.[2] Because HRTF sets are measured at discrete directions, most implementations either snap to the nearest measured HRTF, which Meta notes "exhibits audible discontinuities," or interpolate between neighbors.[2] Microsoft states that for headphones in mixed reality "it's essential to use an HRTF-based technology for accuracy and comfort," and that the low-latency head tracking of mixed reality headsets, including HoloLens, supports high-quality HRTF-based spatialization.[4]

Platform or engine Company HRTF use
Meta XR Audio SDK Meta HRTF-based spatialization of objects and first-order ambisonics for Meta Quest headsets, with an HRTF intensity control and near-field rendering;[3] "Universal HRTF" announced in July 2023 for version 55[31]
Steam Audio Valve Built-in HRTF or a custom HRTF loaded from a SOFA file, with optional RMS volume normalization[10]
Resonance Audio HRTFs combined with ambisonics; plugins for Unity, Unreal Engine, FMOD, Wwise, Android, iOS and the web[5][32]
Windows spatial sound Microsoft HRTF-based spatialization through the ISpatialAudioClient API; hardware-accelerated on HoloLens 2[4]
Personalized Spatial Audio Apple Personal profile from a TrueDepth camera scan (iOS 16 or later), used by supported AirPods and Beats and by Apple Vision Pro[27][33]
Tempest 3D audio engine Sony Interactive Entertainment HRTF processing on the PlayStation 5, with five HRTF presets at launch[26][34]

Meta

In 2014 Oculus VR licensed VisiSonics' RealSpace 3D Audio engine. According to a VisiSonics press release dated 7 October 2014, Oculus announced the deal at Oculus Connect, its first developer conference, in Los Angeles. VisiSonics said the engine combined head-related transfer functions, room models and head tracking, and that it grew out of about a decade of research at the University of Maryland.[35] Oculus CEO Brendan Iribe said at the time: "Audio is an essential ingredient for immersive virtual reality."[35] The current Meta XR Audio SDK uses HRTFs for both object and ambisonic spatialization.[3] In July 2023 UploadVR reported that Meta would replace the SDK's HRTF, which had been based on publicly available data, with a "Universal HRTF" built from "a larger and more precise dataset of the HRTFs for over 150 people." It was to ship in version 55 of the SDK for Unity, and developers only had to update the SDK and rebuild their apps to use it.[31]

Apple

Apple's Personalized Spatial Audio uses the TrueDepth camera on an iPhone running iOS 16 or later to create a personal profile; according to Apple, the camera data is processed on the device and the images are not stored.[27] UploadVR described the feature as letting users "generate a custom HRTF using an iPhone TrueDepth face scan."[31] When Apple announced the Vision Pro in June 2023, it said that two individually amplified drivers in each audio pod "deliver Personalized Spatial Audio based on the user's own head and ear geometry."[33] The profile can be used on Vision Pro running the latest version of visionOS.[27]

Microsoft

Microsoft's documentation describes HRTFs as manipulating level and phase differences between the ears across the frequency spectrum, based on physical models and measurements of head, torso and ear shapes. To improve accuracy, HoloLens uses the interpupillary distance measured for the displays to adjust the HRTFs to the wearer's head size, and HoloLens 2 accelerates HRTF processing in hardware.[4]

Sony

The PlayStation 5's Tempest 3D audio engine creates 3D audio using HRTFs, which PS5 lead architect Mark Cerny discussed in a technical presentation in March 2020.[26][34] What Hi-Fi reported that Sony created five HRTF profiles for the launch of the console, based on data from more than 100 people, and that the console helps each player find the best-matching profile.[26] Cerny described possible future approaches, including: "Maybe you'll be sending us a photo of your ear, and we'll use a neural network to pick the closest HRTF in our library."[34]

See also

References

  1. ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 Song Li, Jürgen Peissig (2020-07-21). "Measurement of Head-Related Transfer Functions: A Review". Applied Sciences, vol. 10, no. 14, article 5014. MDPI. https://doi.org/10.3390/app10145014. Retrieved 2026-09-29.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 2.5 "Spatial audio". Meta Horizon OS Developers. Meta Platforms. https://developers.meta.com/horizon/design/spatial_audio/. Retrieved 2026-09-29.
  3. ↑ 3.0 3.1 3.2 3.3 3.4 "Meta XR Audio SDK Features". Meta Horizon OS Developers. Meta Platforms. https://developers.meta.com/horizon/documentation/unity/meta-xr-audio-sdk-features/. Retrieved 2026-09-29.
  4. ↑ 4.0 4.1 4.2 4.3 4.4 "Spatial sound overview - Mixed Reality". Microsoft Learn. Microsoft. https://learn.microsoft.com/en-us/windows/mixed-reality/design/spatial-sound. Retrieved 2026-09-29.
  5. ↑ 5.0 5.1 "Fundamental Concepts". Resonance Audio. https://resonance-audio.github.io/resonance-audio/discover/concepts.html. Retrieved 2026-09-29.
  6. ↑ 6.0 6.1 Elizabeth M. Wenzel, Marianne Arruda, Doris J. Kistler, Frederic L. Wightman (1993-07). "Localization using nonindividualized head-related transfer functions". The Journal of the Acoustical Society of America, vol. 94, no. 1, pp. 111-123. https://doi.org/10.1121/1.407089. Retrieved 2026-09-29.
  7. ↑ 7.0 7.1 Brian F. G. Katz (2001-11). "Boundary element method calculation of individual head-related transfer function. I. Rigid model calculation". The Journal of the Acoustical Society of America, vol. 110, no. 5, pp. 2440-2448. https://doi.org/10.1121/1.1412440. Retrieved 2026-09-29.
  8. ↑ Brian F. G. Katz (2001-11). "Boundary element method calculation of individual head-related transfer function. II. Impedance effects and comparisons to real measurements". The Journal of the Acoustical Society of America, vol. 110, no. 5, pp. 2449-2455. https://doi.org/10.1121/1.1412441. Retrieved 2026-09-29.
  9. ↑ "SOFA (Spatially Oriented Format for Acoustics)". SOFA Conventions. https://www.sofaconventions.org/mediawiki/index.php/SOFA_(Spatially_Oriented_Format_for_Acoustics). Retrieved 2026-09-29.
  10. ↑ 10.0 10.1 "HRTF - Steam Audio C API documentation". Steam Audio. Valve. https://valvesoftware.github.io/steam-audio/doc/capi/hrtf.html. Retrieved 2026-09-29.
  11. ↑ E. A. G. Shaw (1974-12). "Transformation of sound pressure level from the free field to the eardrum in the horizontal plane". The Journal of the Acoustical Society of America, vol. 56, no. 6, pp. 1848-1861. https://doi.org/10.1121/1.1903522. Retrieved 2026-09-29.
  12. ↑ Durand R. Begault, Elizabeth M. Wenzel, Martine Godfroy, Joel D. Miller, Mark R. Anderson (2010-10). "Applying Spatial Audio to Human Interfaces: 25 Years of NASA Experience". AES 40th International Conference, Tokyo. NASA Technical Reports Server. https://ntrs.nasa.gov/api/citations/20110008295/downloads/20110008295.pdf. Retrieved 2026-09-29.
  13. ↑ Frederic L. Wightman, Doris J. Kistler (1989-02). "Headphone simulation of free-field listening. I: Stimulus synthesis". The Journal of the Acoustical Society of America, vol. 85, no. 2, pp. 858-867. https://doi.org/10.1121/1.397557. Retrieved 2026-09-29.
  14. ↑ Frederic L. Wightman, Doris J. Kistler (1989-02). "Headphone simulation of free-field listening. II: Psychophysical validation". The Journal of the Acoustical Society of America, vol. 85, no. 2, pp. 868-878. https://doi.org/10.1121/1.397558. Retrieved 2026-09-29.
  15. ↑ 15.0 15.1 Isaac Engel, Rapolas Daugintis, Thibault Vicente, Aidan O. T. Hogg, Johan Pauwels, Arnaud J. Tournier, Lorenzo Picinali (2023-05). "The SONICOM HRTF Dataset". Journal of the Audio Engineering Society, vol. 71, no. 5, pp. 241-253. doi:10.17743/jaes.2022.0066. https://spiral.imperial.ac.uk/server/api/core/bitstreams/dfd02d49-9dd7-447e-a225-f1743906fb6d/content. Retrieved 2026-09-29.
  16. ↑ Bill Gardner, Keith Martin. "HRTF Measurements of a KEMAR Dummy-Head Microphone". MIT Media Lab. https://sound.media.mit.edu/resources/KEMAR.html. Retrieved 2026-09-29.
  17. ↑ William G. Gardner, Keith D. Martin (1995-06). "HRTF measurements of a KEMAR". The Journal of the Acoustical Society of America, vol. 97, no. 6, pp. 3907-3908. https://doi.org/10.1121/1.412407. Retrieved 2026-09-29.
  18. ↑ 18.0 18.1 18.2 V. R. Algazi, R. O. Duda, D. M. Thompson, C. Avendano (2001-10). "The CIPIC HRTF Database". 2001 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, New Paltz, New York. https://www.ece.ucdavis.edu/cipic/wp-content/uploads/sites/12/2015/04/cipic_WASSAP_2001_143.pdf. Retrieved 2026-09-29.
  19. ↑ 19.0 19.1 Katarina C. Poole, Julie Meyer, Vincent Martin, Rapolas Daugintis, Nils Marggraf-Turley, Jack Webb, Ludovic Pirard, Nicola La Magna, Oliver Turvey, Lorenzo Picinali (2025-07-07). "The Extended SONICOM HRTF Dataset and Spatial Audio Metrics Toolbox". arXiv. https://arxiv.org/abs/2507.05053. Retrieved 2026-09-29.
  20. ↑ Durand R. Begault, Elizabeth M. Wenzel, Mark R. Anderson (2001-10). "Direct comparison of the impact of head tracking, reverberation, and individualized head-related transfer functions on the spatial perception of a virtual speech source". Journal of the Audio Engineering Society, vol. 49, no. 10, pp. 904-916. https://pubmed.ncbi.nlm.nih.gov/11885605/. Retrieved 2026-09-29.
  21. ↑ C. Jenny, C. Reuter (2020). "Usability of Individualized Head-Related Transfer Functions in Virtual Reality: Empirical Study With Perceptual Attributes in Sagittal Plane Sound Localization". JMIR Serious Games, vol. 8, no. 3, e17576. https://doi.org/10.2196/17576. Retrieved 2026-09-29.
  22. ↑ Ravish Mehra, Aaron Nicholls, Durand Begault, Marina Zannoli (2016-10). "Comparison of localization performance with individualized and non-individualized head-related transfer functions for dynamic listeners". The Journal of the Acoustical Society of America, vol. 140, no. 4 (supplement), pp. 2956-2957. https://doi.org/10.1121/1.4969129. Retrieved 2026-09-29.
  23. ↑ Vincent Martin, Lorenzo Picinali (2025-10-10). "Impact of HRTF individualisation and head movements in a real/virtual localisation task". arXiv. https://arxiv.org/abs/2510.09161. Retrieved 2026-09-29.
  24. ↑ Paul M. Hofman, Jos G. A. Van Riswick, A. John Van Opstal (1998-09). "Relearning sound localization with new ears". Nature Neuroscience, vol. 1, no. 5, pp. 417-421. doi:10.1038/1633. https://pubmed.ncbi.nlm.nih.gov/10196533/. Retrieved 2026-09-29.
  25. ↑ Christopher C. Berger, Mar Gonzalez-Franco, Ana Tajadura-Jiménez, Dinei Florencio, Zhengyou Zhang (2018-02-02). "Generic HRTFs May be Good Enough in Virtual Reality. Improving Source Localization through Cross-Modal Plasticity". Frontiers in Neuroscience, vol. 12, article 21. https://doi.org/10.3389/fnins.2018.00021. Retrieved 2026-09-29.
  26. ↑ 26.0 26.1 26.2 26.3 "PS5 3D audio: what is it? How do you get it?". What Hi-Fi?. 2020-09-24. https://www.whathifi.com/features/ps5-3d-audio-what-is-it-how-do-you-get-it. Retrieved 2026-09-29.
  27. ↑ 27.0 27.1 27.2 27.3 "Listen with Personalized Spatial Audio for AirPods and Beats". Apple Support. Apple. https://support.apple.com/en-us/102596. Retrieved 2026-09-29.
  28. ↑ 28.0 28.1 Xikun Lu, Yunda Chen, Zehua Chen, Jie Wang, Mingxing Liu, Hongmei Hu, Chengshi Zheng, Stefan Bleeck, Jinqiu Sang (2025-08-30). "Deep Learning for Personalized Binaural Audio Reproduction". arXiv. https://arxiv.org/abs/2509.00400. Retrieved 2026-09-29.
  29. ↑ Navid H. Zandi, Awny M. El-Mohandes, Rong Zheng (2022-03-21). "Individualizing Head-Related Transfer Functions for Binaural Acoustic Applications". arXiv. https://arxiv.org/abs/2203.11138. Retrieved 2026-09-29.
  30. ↑ Ben Lang (2020-09-03). "Facebook Wants to Build an AR Headset to Supercharge Your Hearing, Create a Custom HRTF from a Photograph". Road to VR. https://www.roadtovr.com/facebook-ar-headset-supercharge-hearing-spatial-audio-context-noise-cancellation/. Retrieved 2026-09-29.
  31. ↑ 31.0 31.1 31.2 David Heaney (2023-07-18). "Meta Upgrades Quest Spatial Audio To Be More Realistic". UploadVR. https://www.uploadvr.com/meta-quest-universal-hrtf-spatial-audio/. Retrieved 2026-09-29.
  32. ↑ "Resonance Audio". Resonance Audio. https://resonance-audio.github.io/resonance-audio/. Retrieved 2026-09-29.
  33. ↑ 33.0 33.1 "Introducing Apple Vision Pro: Apple's first spatial computer". Apple Newsroom. Apple. 2023-06-05. https://www.apple.com/newsroom/2023/06/introducing-apple-vision-pro/. Retrieved 2026-09-29.
  34. ↑ 34.0 34.1 34.2 Nick Pino (2020-03-18). "PS5 won't use Dolby Atmos in games and may require you to send Sony a picture of your ears". TechRadar. https://www.techradar.com/news/ps5-wont-use-dolby-atmos-in-games-and-may-require-you-to-send-sony-a-picture-of-your-ears. Retrieved 2026-09-29.
  35. ↑ 35.0 35.1 "VisiSonics' RealSpace 3D Audio Software Licensed by Oculus for Virtual Reality". PR Newswire. VisiSonics. 2014-10-07. https://www.prnewswire.com/news-releases/visisonics-realspace-3d-audio-software-licensed-by-oculus-for-virtual-reality-278413231.html. Retrieved 2026-09-29.
Retrieved from "https://vrarwiki.com/index.php?title=Head-related_transfer_function&oldid=46112"