November 09, 2010
GritTec has updated high level API of GritTec's Speaker-ID SDK up to version 2,990.
In this version GritTec has optimised handler of system
events with returning errors codes for the general functions and
procedures of high level GritTec's Speaker-ID SDK. Efficiency of the
system events handler was checked in the multi-threading mode of
voice identification for audio files with big sizes (~ 10mB).
Testing has been made on the PC computer with Windows XP (x86) and
system parameters: AMD Athlon 64 X2 6000t, 2 GB RAM. For testing the
new SDK GritTec's developers have created the special program with
GUI interface and supporting of multi-threading voice identification
and/or voice verification.
During running of GUI program with 20 streams of voice
identification the handler of system events has returned messages
about low memory resources for some streams. It has allowed to stop
or suspend completely streams at the level of creation of milti-threading
voice identification program for which there were such system
messages. Thus the updated high level SDK gives an opportunity to
developers to control work of GritTec's Speaker-ID engine more
Reminder, high level API represents a universal remedy of text independent voice
identification and/or voice verification, in which all functions and procedures
are presented in the unified form. It allows developers to
reduce integration time of identification engine under the another compilers
(distinct from the C/C ++), i.e. without creation of special wrappers modules
for using of GritTec's libraries.
GritTec Laboratory (GritTec Ltd.) specializes on research and
development of algorithms and technologies in the field of speech
and audio processing. GritTec's research is focused on speech
enhancement, speech concealment, voice biometric, speech recognition,
speech synthesis and other speech and audio technologies.