Academic Journal
Προβολή σε EDS

Psychoacoustically inspired optimization framework for neural binaural synthesis.

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: Psychoacoustically inspired optimization framework for neural binaural synthesis.
Συγγραφείς: Lu X; Shanghai Institute of Artificial Intelligence for Education, East China Normal University, Shanghai, China., Liu F; School of Computer Science and Technology, East China Normal University, Shanghai, China., Xie X; School of Computer Science and Technology, East China Normal University, Shanghai, China., Na R; School of Computer Science and Technology, East China Normal University, Shanghai, China., Sang J; School of Computer Science and Technology, East China Normal University, Shanghai, China.
Πηγή: The Journal of the Acoustical Society of America [J Acoust Soc Am] 2026 Aug 01; Vol. 160 (2), pp. 1521-1534.
Τύπος έκδοσης: Journal Article
Γλώσσα: English
Στοιχεία περιοδικού: Publisher: American Institute of Physics Country of Publication: United States NLM ID: 7503051 Publication Model: Print Cited Medium: Internet ISSN: 1520-8524 (Electronic) Linking ISSN: 00014966 NLM ISO Abbreviation: J Acoust Soc Am Subsets: MEDLINE
Imprint Name(s): Publication: Melville, NY : American Institute of Physics
Original Publication: Lancaster, Pa. [etc.] : American Institute of Physics for the Acoustical Society of America
Ιατρικοί όροι (MeSH): Sound Localization*/physiology , Psychoacoustics* , Auditory Perception* , Models, Neurological*, Auditory Pathways/physiology ; Humans ; Cues ; Acoustic Stimulation ; Wavelet Analysis ; Fourier Analysis ; Perceptual Masking ; Signal Processing, Computer-Assisted ; Time Factors
Περίληψη: Neural binaural synthesis is a promising technique for immersive audio reproduction. However, current methods rely on signal-oriented optimization objectives that treat all time-frequency regions uniformly. This approach conflicts with the selective nature of human auditory perception, as it may overemphasize perceptually masked or sub-threshold components, resulting in audible artifacts in reverberant sound fields. To address this, a perceptual optimization framework inspired by human spatial hearing is introduced. It includes a coherence-gated spatial loss that uses interaural coherence to weight spatial cues according to their reliability. By reducing the contribution of low-coherence regions associated with diffuse reverberation, this mechanism forces the model to prioritize the reconstruction of direct sound and early reflections critical for localization. A multi-scale wavelet loss is also incorporated to measure reconstruction errors across multiple wavelet decomposition levels, addressing the fixed resolution trade-off of the short-time Fourier transform and helping preserve fine-grained temporal transients. The framework was validated on three distinct backbone architectures. Extensive objective evaluations and subjective listening tests under non-head-tracked binaural playback show consistent improvements over state-of-the-art baselines, demonstrating its effectiveness as a model-agnostic training strategy for improving both signal fidelity and spatial accuracy.
(© 2026 Acoustical Society of America.)
Entry Date(s): Date Created: 20260814 Date Completed: 20260814 Latest Revision: 20260814
Update Code: 20260814
DOI: 10.1121/10.0044589
PMID: 42599064
Βάση Δεδομένων: MEDLINE
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://resolver.ebsco.com/c/fiv2js/result?sid=EBSCO:cmedm&genre=article&issn=15208524&ISBN=&volume=160&issue=2&date=20260801&spage=1521&pages=1521-1534&title=The Journal of the Acoustical Society of America&atitle=Psychoacoustically%20inspired%20optimization%20framework%20for%20neural%20binaural%20synthesis.&aulast=Lu%20X&id=DOI:10.1121/10.0044589
    Name: Full Text Finder (for New FTF UI) (ns324271)
    Category: fullText
    Text: Full Text Finder
    MouseOverText: Full Text Finder
Header DbId: cmedm
DbLabel: MEDLINE
An: 42599064
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Psychoacoustically inspired optimization framework for neural binaural synthesis.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AU" term="%22Lu+X%22">Lu X</searchLink>; Shanghai Institute of Artificial Intelligence for Education, East China Normal University, Shanghai, China.<br /><searchLink fieldCode="AU" term="%22Liu+F%22">Liu F</searchLink>; School of Computer Science and Technology, East China Normal University, Shanghai, China.<br /><searchLink fieldCode="AU" term="%22Xie+X%22">Xie X</searchLink>; School of Computer Science and Technology, East China Normal University, Shanghai, China.<br /><searchLink fieldCode="AU" term="%22Na+R%22">Na R</searchLink>; School of Computer Science and Technology, East China Normal University, Shanghai, China.<br /><searchLink fieldCode="AU" term="%22Sang+J%22">Sang J</searchLink>; School of Computer Science and Technology, East China Normal University, Shanghai, China.
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%227503051%22">The Journal of the Acoustical Society of America</searchLink> [J Acoust Soc Am] 2026 Aug 01; Vol. 160 (2), pp. 1521-1534.
– Name: TypePub
  Label: Publication Type
  Group: TypPub
  Data: Journal Article
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: TitleSource
  Label: Journal Info
  Group: Src
  Data: <i>Publisher: </i><searchLink fieldCode="PB" term="%22American+Institute+of+Physics%22">American Institute of Physics </searchLink><i>Country of Publication: </i>United States <i>NLM ID: </i>7503051 <i>Publication Model: </i>Print <i>Cited Medium: </i>Internet <i>ISSN: </i>1520-8524 (Electronic) <i>Linking ISSN: </i><searchLink fieldCode="IS" term="%2200014966%22">00014966 </searchLink><i>NLM ISO Abbreviation: </i>J Acoust Soc Am <i>Subsets: </i>MEDLINE
– Name: PublisherInfo
  Label: Imprint Name(s)
  Group: PubInfo
  Data: <i>Publication</i>: Melville, NY : American Institute of Physics<br /><i>Original Publication</i>: Lancaster, Pa. [etc.] : American Institute of Physics for the Acoustical Society of America
– Name: SubjectMESH
  Label: MeSH Terms
  Group: Su
  Data: <searchLink fieldCode="MM" term="%22Sound+Localization%22">Sound Localization*</searchLink>/<searchLink fieldCode="MM" term="%22Sound+Localization+physiology%22">physiology</searchLink> <br /><searchLink fieldCode="MM" term="%22Psychoacoustics%22">Psychoacoustics*</searchLink> <br /><searchLink fieldCode="MM" term="%22Auditory+Perception%22">Auditory Perception*</searchLink> <br /><searchLink fieldCode="MM" term="%22Models%2C+Neurological%22">Models, Neurological*</searchLink><br /><searchLink fieldCode="MH" term="%22Auditory+Pathways%22">Auditory Pathways</searchLink>/<searchLink fieldCode="MH" term="%22Auditory+Pathways+physiology%22">physiology</searchLink> ; <searchLink fieldCode="MH" term="%22Humans%22">Humans</searchLink> ; <searchLink fieldCode="MH" term="%22Cues%22">Cues</searchLink> ; <searchLink fieldCode="MH" term="%22Acoustic+Stimulation%22">Acoustic Stimulation</searchLink> ; <searchLink fieldCode="MH" term="%22Wavelet+Analysis%22">Wavelet Analysis</searchLink> ; <searchLink fieldCode="MH" term="%22Fourier+Analysis%22">Fourier Analysis</searchLink> ; <searchLink fieldCode="MH" term="%22Perceptual+Masking%22">Perceptual Masking</searchLink> ; <searchLink fieldCode="MH" term="%22Signal+Processing%2C+Computer-Assisted%22">Signal Processing, Computer-Assisted</searchLink> ; <searchLink fieldCode="MH" term="%22Time+Factors%22">Time Factors</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Neural binaural synthesis is a promising technique for immersive audio reproduction. However, current methods rely on signal-oriented optimization objectives that treat all time-frequency regions uniformly. This approach conflicts with the selective nature of human auditory perception, as it may overemphasize perceptually masked or sub-threshold components, resulting in audible artifacts in reverberant sound fields. To address this, a perceptual optimization framework inspired by human spatial hearing is introduced. It includes a coherence-gated spatial loss that uses interaural coherence to weight spatial cues according to their reliability. By reducing the contribution of low-coherence regions associated with diffuse reverberation, this mechanism forces the model to prioritize the reconstruction of direct sound and early reflections critical for localization. A multi-scale wavelet loss is also incorporated to measure reconstruction errors across multiple wavelet decomposition levels, addressing the fixed resolution trade-off of the short-time Fourier transform and helping preserve fine-grained temporal transients. The framework was validated on three distinct backbone architectures. Extensive objective evaluations and subjective listening tests under non-head-tracked binaural playback show consistent improvements over state-of-the-art baselines, demonstrating its effectiveness as a model-agnostic training strategy for improving both signal fidelity and spatial accuracy.<br /> (© 2026 Acoustical Society of America.)
– Name: DateEntry
  Label: Entry Date(s)
  Group: Date
  Data: <i>Date Created: </i>20260814 <i>Date Completed: </i>20260814 <i>Latest Revision: </i>20260814
– Name: DateUpdate
  Label: Update Code
  Group: Date
  Data: 20260814
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1121/10.0044589
– Name: AN
  Label: PMID
  Group: ID
  Data: 42599064
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=42599064
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1121/10.0044589
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        StartPage: 1521
    Subjects:
      – SubjectFull: Auditory Pathways physiology
        Type: general
      – SubjectFull: Humans
        Type: general
      – SubjectFull: Cues
        Type: general
      – SubjectFull: Acoustic Stimulation
        Type: general
      – SubjectFull: Wavelet Analysis
        Type: general
      – SubjectFull: Fourier Analysis
        Type: general
      – SubjectFull: Perceptual Masking
        Type: general
      – SubjectFull: Signal Processing, Computer-Assisted
        Type: general
      – SubjectFull: Time Factors
        Type: general
      – SubjectFull: Sound Localization physiology
        Type: general
      – SubjectFull: Psychoacoustics
        Type: general
      – SubjectFull: Auditory Perception
        Type: general
      – SubjectFull: Models, Neurological
        Type: general
    Titles:
      – TitleFull: Psychoacoustically inspired optimization framework for neural binaural synthesis.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Lu X
      – PersonEntity:
          Name:
            NameFull: Liu F
      – PersonEntity:
          Name:
            NameFull: Xie X
      – PersonEntity:
          Name:
            NameFull: Na R
      – PersonEntity:
          Name:
            NameFull: Sang J
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 08
              Text: 2026 Aug 01
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-electronic
              Value: 1520-8524
          Numbering:
            – Type: volume
              Value: 160
            – Type: issue
              Value: 2
          Titles:
            – TitleFull: The Journal of the Acoustical Society of America
              Type: main
ResultId 1