The ATK Real-Time API for HTK
ATK is an API designed to facilitate building experimental
applications for HTK. It consists of a C++ layer sitting on top of
the standard HTK libraries. This allows novel recognisers built using
customised versions of HTK to be compiled with ATK and then tested in
working systems. Like HTK itself, it is portable across the main Unix
platforms and Windows.
ATK features include:
- Multi-threaded to allow efficient and responsive real-time operation.
- Synchronised audio input/output with barge-in support.
- Support for finite-state grammars and trigram language models.
- Ability to return recognition results word-by-word as they are
recognised to reduce latency
- N-best recognition output
- Support for HLDA
- Integrated Flite speech synthesis.
- Make files for single-build under Linux and Windows
Requirements: ATK supports both Linux and Windows in a single set of
sources so like HTK itself, there are no platform specific
distributions. Compilation under Linux requires that the ALSA audio
development package be installed.
Compilation under Windows using the supplied project files requires Visual
Studio 2005 to be installed.
Further information can be found on
the ATK home
page
The complete documentation is available as a
pdf
Registered HTK users can download the current version (1.6) from
here.