International organisation for standardisation

5.2.3MPEG-D SAOC

Rendering (Monday Afternoon)

Yang-Won Jung, LG, presented

m15403

Thoughts on SAOC applications using preset rendering information

Yang-Won Jung Henney Oh

This presentation started with a demo of an LG proprietary implementation of stereo remix technology. It is similar to SAOC but is not SAOC. However, the presenter indicated that the functionality should be possible using SAOC technology.

The presentation noted that there is no “default” rendering matrix for SAOC. This is totally up to the decoder. It proposes to adopt the concept of “presets” to enable a default set of rendering matrices. Encoder delivered presets would enable a simple “Karaoke” or “A Cappella” functionality without the need of a user to identify the (perhaps several) objects that would need to be attenuated. User created presets (which could be saved) would eliminate the need for a user to re-create a desired sound for e.g. a favorite song. In a teleconferencing scenario, the MCU could send a preset to the terminals.

The contribution proposes the following requirements for presets:

• 
  Preset in bitstream
  
• 
  Multiple presets
  
• 
  Time-varying (dynamic) presets
  
• 
  User-created (and saved) presets
  

The Chair noted that there appears to be a need to standardize the communication of the rendering matrix over two channels: over the communication channel (i.e. in the bitstream) and to/from a local storage medium.

Yang-Won Jung, LG, presented

m15404

A proposed CE on preset in SAOC

Yang-Won Jung Henney Oh

This contribution proposes specific changes to the SAOC WD to support presets delivered via the bitstream. It shows how the rendering matrix can be derived from a small number of gain parameters and enumerated these for the case of mono, stereo and 5.1 channel. Syntax and semantics for the proposed preset method was given. Further, it suggests that there be a mandatory preset in all SAOC bitstreams.

Leonid Terentiev, FhG, noted that the parameter to rendering matrix decoding entails some log and square root calculations.

Heiko Purnhagen, Dolby, presented

m15354

Proposal for support of the preset rendering information data by the MPEG SAOC system

Oliver Hellmuth Johannes Hilpert Leonid Terentiev Cornelia Falch Heiko Purnhagen Jonas Engdegård Jeroen Koppens

This contribution discusses another method of how to convey presets in the bitstream. It proposes to make the preset semantics informative. The rendering matrix is transmitted in the native format used by SAOC. An alternative format is a generic container to carry user information. The proposal transmits preset information as header information only, such that dynamic presets are not possible.

The Chair noted that the proposal is informative, and questioned the benefit of doing that.

Jeongil Seo, ETRI, presented

m15369

Proposal for Transmitting Rendering Matrix Information within SAOC Bitstream

Jeongil Seo Seungkwon Beack Kyeongok Kang Kwang-ki Kim

The presenter did not wish to repeat the points covered by the previous two presentations, and so went directly to the conclusions: that there should be support for multiple presets in the bitstream and that the syntax and semantics be normative.

The Chair summarized the open issues:

• 
  Transmission format
  
• 
  Normative or Informative
  
• 
  Static or Dynamic