Organisation internationale de normalisation

3D Audio Phase 1

Monday

Joe Hilpert, FhG-IIS, presented

m35019

Software for MPEG-H 3D Audio CO RM4

[e-mail],Michael Fischer, Adrian Murtaza, [e-mail],Johannes Hilpert, [e-mail],Christian Ertel, Simone Fueg, Michael Meier, Alexander KrÌ_ger, Florian Keiler, Sven Kordon, Jeongil Seo, Henney Oh, Taegyu Lee, Nils Peters, Deep Sen, Charles Verron, Gregory Pallone

H-SW

The contribution summarized the work done in the AhG period to bring the reference software up to date with the functionality of the 3D Audio DIS text. The list of changes is shown here:

• 
  Incorporate MHAS extensions as in m33601
  
• 
  Incorporate update on delay handling as in m34271
  
• 
  Incorporate SUPPORT_MAE_GROUP_PRESET as in m33602
  
• 
  Incorporate multiLanguage support as in N14746
  
• 
  Incorporate decoder interface for user interaction as in m34291
  
• 
  Incorporate rendering interface for local loudspeaker setup information as in m34293
  
• 
  Incorporate local screen information as in m34296
  
• 
  Incorporate proposed updates to dynamic priority as in m34254
  
• 
  Incorporate imaginary loudspeakers as in m33205
  
• 
  Incorporate object spread as in m34286
  
• 
  Incorporate OAM frameLength signaling fix as in m34264
  
• 
  Incorporate profile level indication in mpegh3daConfig
  
• 
  Incorporate active downmix setting signaling as in m33228
  
• 
  Incorporate proposed update of loudnessInfoSet as in N14746
  
• 
  Incorporate proposed downmixID matching scheme as in m34249
  
• 
  Incorporate receiverDelayCompensation as agreed at the AhG Meeting in Paris
  
• 
  The VQ V-vector coding method, proposed in m34105, has been integrated into the HOA encoder and decoder software
  
• 
  A new software module for encoding and decoding of HOA rendering matrices, as in m33620, has been added. Thus the HOA rendering matrix can now be signaled in the bitstream and used by the HOA renderer for suitable rendering setup.
  
• 
  The time-domain DRC-1 for HOA from m33151 has been added into the HOA Decoder
  
• 
  The adjustment from m34239 and m34240 have been integrated into the gain control module
  

In addition, the software was re-organized so that the processing delay in the time or QMF domain can be managed with greater visibility.

The contribution presented numerical differences between RM3 and RM4. This was characterized for each change. Most changes resulted in a difference (e.g. -40 dB) that is not expected to be perceivable.

This contribution implements these complete set of changes

• 
  CO core functionality
  
• 
  HOA core functionality
  
• 
  TD binauralization
  

The presenter proses to additionally integrate all other software modules available at this meeting:

• 
  M35008
  
• 
  M35014
  
• 
  M35028
  

The Chair noted that all of the above contributions will be discussed and the Audio subgroup will decide on the scope of a software integration effort.
Sang Bae Chon, Samsung, presented

m35008

Reference Software of MPEG-H 3D Audio Immersive Audio Rendering

Sang Bae Chon, Sunmin Kim

H-SW

The contribution noted which reference software files are changed and which directory of software (i.e. an immersive audio) is new. The presenter noted that the integration has been checked with RM3, and will be checked with RM4.

It was the consensus of the Audio subgroup to integrate this software into RM4.

Taegyu Lee, Yonsei University, presented

m35014

Updates for Frequency-domain Binaural Rendering Software

Taegyu Lee, Young-cheol Park, Dae Hee Henney, Jeongil Seo

H-SW

The contribution reports on how the FD Binaural interface was incorporated into the 3D Audio reference software. It included a software re-organization in which the FD and TD engines are integrated in a single structure. When possible, source code filenames are changed to match a related syntax element in the text, and source code variables are changed to match a related syntax element in the text.

• 
  Interface
  
  • 
    Read WAV-format impulse response file
    
  • 
    Write normative impulse response format
    
• 
  Parameterization
  
  • 
    FIR representation to Parameter representation
    
• 
  Renderer
  
  • 
    Actual binauralization renderer
    

It was the consensus of the Audio subgroup to integrate this software into RM4. The authors of this contribution will check the integrated implementation, which should be bit-identical to previous software implementations.

The contribution identified an editorial error a Figure of the DIS text:

• 
  Correct subscript in Figure 53 (LN changed to Ln)
  

Finally, the software integration process identified two bugs in the DIS text:

• 
  Subclause 13.2.2, no limit on k
  
• 
  Subclause 13.2.3.4, change k=63 to k=47
  

The Chair requested that the authors re-submit these “bug fixes” as a contribution to the 111^th MPEG meeting.
Toru Chinen, Sony, presented

m35028

Proposed addition of a software of dynamic object priority generator in 3D Audio reference software

Yuki Yamamoto, Toru Chinen, Masayuki Nishiguchi

H-SW

The contribution proposes to extend the RM4 OAM file format to include

• 
  Dynamic object priority field
  

The presenter noted that revised software, suitable for integration with RM4, is. The Chair suggested that this new software be included in a revision of the contribution and uploaded by the presenter (now).

It was the consensus of the Audio subgroup to integrate this software into RM4. The authors of the contribution will integrate the new zip archive of software and also provide the code to transfer dynamic priority in the OAM file to the appropriate bitstream field via modifications of the RM4 core software.