Mixing different codecs – Nortel Networks NN44470-100 User Manual

Page 20

- 20 -

MAS commissioning

Nortel Media Application Server 6.0 for AS 5300

Fundamentals

NN44470-100 01.01 Standard

Release 6.0 03 June 2008

Mixing different codecs

The system can mix participants with potentially different codecs in the same
conference. To achieve this, the system transcodes the incoming streams into
a common format, mixes, and then provides each participant with the correctly
mixed audio. The system optimizes for multiple participants by transcoding the
resulting mixed audio based upon a grouping of conference participants. If
every participant uses identical codecs, at least three groups exist in a
conference.

The first group contains the primary speaker, the second group contains the
secondary speaker, and the third group contains the remainder of the
participants. The groups describe the different versions of the mixed audio
required in a conference. The first and second groups exist because the
primary and secondary speakers do not hear their own voice, thus they must
receive a different version of the audio than other participants. The third group
is for all other participants who require the mixed audio of both active
speakers.

The system must transcode all incoming audio streams to test for energy.
However, if the basic conference algorithm is used, the system must
transcode only the audio of each group and then replicate for each participant
as needed. This conserves system resources. If multiple codecs exist in a
single conference or if the premium conference algorithm is used, additional
groups are required. The system can, therefore, optimize conferences with
more than three participants because the additional system resources
required for each participant includes transcoding the incoming audio stream
and testing for energy. The additional participants are assigned to existing
groups as needed so they receive the same audio as other group members.

The MAS can provide additional media processing functions for conferences
on a global conference basis or for each participant. Optionally, the system
can allocate a global announcement port that provides broadcast
announcement capability to all participants. Additionally, the system can
allocate a pseudo resource for each participant that enables the system to
perform digit collection, speech recognition, and whisper-like functions for a
specific individual.