International organisation for standardisation

4.2.2MPEG-D Spatial Audio Object Coding

SAOC patent statement

The Chair noted that since SAOC progresses to FDIS, companies that feel that they hold patents that read on the specification are required to submit patent statements. The appropriate form for the patent statement and the party to which the form should be submitted can be found here:

Go to SC29 website: http://www.itscj.ipsj.or.jp/sc29/

Click on link “Procedures”

Click on link “Common Patent Policy for ITU-T/ITU-R/ISO/IEC”

Click on link “Patent statement and licensing declaration form for ITU-T/ITU-R Recommendation | ISO/IEC Deliverable.” This is the patent statement form and includes submission instructions.

Leonid Terentiev, FhG, presented

m17263

Changes for editorial consistency of SAOC FCD text

Jonas Engdegård Heiko Purnhagen Jürgen Herre Andreas Hölzer Leonid Terentiev Cornelia Falch Oliver Hellmuth Jeroen Koppens

This contribution proposes numerous editorial corrections such as ISO template formatting and spelling errors. It also proposes addition of missing syntax elements. Finally, it proposes a small addition to the informative annex that provide guidelines to normalize the relative energy of the audio objects in the downmix.

It was the consensus of the Audio subgroup to incorporate the suggested changes into the SAOC FDIS text.

Leonid Terentiev, FhG, presented

m17264

Report on corrections for SAOC

Jonas Engdegård Heiko Purnhagen Jürgen Herre Andreas Hölzer Leonid Terentiev Cornelia Falch Oliver Hellmuth Maria Luis Valero Jeroen Koppens

The contribution proposes numerous minor technical changes:

• 
  Change formulas to insure that decoder calculations will never be subject to numerical instability (e.g. division by zero)
  
• 
  Correct bitstream so that low delay syntax is in line with high-quality syntax
  

It was the consensus of the Audio subgroup to incorporate the suggested changes into the SAOC FDIS text.

Henney Oh, LG, presented

m17285

Proposed corrections on MPEG SAOC text

Henney Oh

The contribution proposes two categories of changes:

• 
  Minor technical corrections
  
• 
  Editorial changes for consistency and clarity
  

In each category numerous changes and corrections are proposed. A break-out group will review the contribution and the result of that work will be reviewed later in the week (e.g. when reviewing the DoC on SAOC FCD ballot.

Henney Oh, LG, presented

m17286

LG-ETRI Crosscheck listening test report on SAOC DCU

Henney Oh Jeongil Seo

The contribution reviewed the test methodology. It was noted that this is NOT a MUSHRA test, in that there is no hidden reference and no anchors. The results of the listening test were that the mean score for the DCU system was greater than that of the “NoLim” system at the 95% level of significance. However, the DCU system was not different from the downmix system. Finally, the contribution notes that it was decided during the AhG period that the cross-check test outcome was positive and therefore, as agreed at the 90^th MPEG meeting, the DCU technology would be incorporated into the Study on SAOC FCD text.

Henney Oh, LG, presented

m17287

Information and comments on SAOC DCU

Henney Oh

The contribution notes three problematic issues in the SAOC specification:

• 
  First, it gave an example of how the current DCU specification can produce results that are very far from what the user has requested as an output mix. Since this issue was identified only recently, there was not sufficient time to develop a satisfactory solution.
  
• 
  Second, the contribution notes that it is essential to exclude the Enhanced Audio Object (EAO) from the DCU process
  
• 
  Third, the contribution gives an example GUI that will display what the DCU module actually implemented. In this respect, the contribution requests to standardize an interface between DCU a possible GUI control.
  
• 
  Fourth, that it be possible to switch off the DCU at the decoder.
  
• 
  Finally, it suggests that the DCU quality evaluation should use a strict MUSHRA methodology.
  

Juergen Herre, FhG, pointed out that g_DCU is typically chosen to be small (e.g. 0.3). In this case, the example of the contribution (in which g_DCU is 1.0) will never occur. There was considerable discussion of the other points raised by the contribution.

Jeongil Seo, ETRI, presented

m17187

Comments on the Usage of DCU in SAOC

Jeongil Seo Kyeongok Kang

First, the presenter demonstrated an implementation of the Interactive Music Application Format (IMAF) that uses SAOC, and particularly showed a GUI that provides visual feedback of the DCU decision.

The contribution is joint work of ETRI and Audizen. It notes that there are applications in which a an over-ride of DCU control may be what the user wishes (e.g. if the user wants to hear only the guitar so to be better able to learn the guitar part of a song). Hence the contribution requests that there be a DCU “on-off” switch at the SAOC decoder.

Leonid Terentiev

m17267

Additional information on SAOC DCU

Jonas Engdegård Heiko Purnhagen Jürgen Herre Cornelia Falch Leonid Terentiev Oliver Hellmuth

The contribution proposes to

• 
  Rename DCU-related bitstream elements to better express their functionality
  
• 
  Add a syntax element that permits either static (i.e. once in header) or dynamic (i.e. periodically in the frame information) update of DCU parameters.
  
• 
  Clarify the semantics of DCU to better express the intended purpose of the DCU.
  

These proposals were discussed, and it appears that there are differing opinions as to whether the DCU information specified in the bitstream shall be normative and mandatory or whether a decoder, via local user control, might over-ride or modify the DCU controls and DCU processing.

The Chair recommended that interested experts discuss in a break-out points presented in the five previous contributions and report back to the Audio subgroup.