HMP Linux 1.2 Release Update, Rev 05 — September 2006September 2006 31
disable the active talker feature, which is always enabled. It only disables the notification to
the application program. The active talker feature sums the 3 most active talkers in a
conference, so that the conversation doesn’t get drowned out when too many people talk at
once. Active talker notification provides data on active talkers through thedcb_gettalkers()
and dcb_GetAtiBitsEx() function s, which can be used by an application program to identify
active talkers; for example, to provide a visual display highlighting the active talkers in a
conference. Active talkers are determined by their loudness; i.e., the strength of their “non-
silence” energy.
Note: In some cases, it is desirable to inactivate the active talker feature, such as for a
background music application program. Although you cannot directly disable the
active talker feature, you can set the noise level threshhold by which signals are
recognized as either speech or noise.
3.4.3 Continuous Speech Processing API for Linux and Windows Operating Systems Library Reference
For more information on DV_TPT termination conditions when using ec_reciottdata() and
ec_stream() functions, see the DV_TPT data structure description in the Vo ice A PI for
Host Media Processing Library Reference.
(IPY00010008 = PTR 33269)
The code examples for the functions ec_reciottdata() and ec_stream() specify a data
format that is not valid on HMP. The example code should be revised as follows:

Original

..
xpb.wFileFormat = FILE_FORMAT_VOX;
xpb.wDataFormat = DATA_FORMAT_PCM;
xpb.nSamplesPerSec = DRT_8KHZ;
xpb.wBitsPerSample = 8;
..

Revised

..
xpb.wFileFormat = FILE_FORMAT_VOX;
xpb.wDataFormat = DATA_FORMAT_MULAW;
xpb.nSamplesPerSec = DRT_8KHZ;
xpb.wBitsPerSample = 8;
..
(IPY00009264 = PTR 33379)