In this method, it is the same with additional training in a sense that it uses a certain acoustic model. But, it is more of an adaptation method from a certain acoustic model to a target acoustic model by estimating the linear transformation matrix.
From HTK 3.3, the adaptation method changed from using HAdapt method to HERest method. Refer to the HTK book for more details.
See Also
HTK Speech Recognition Toolkit. For acoustic models for Julius, see source information of this document. For acoustic model construction for HTK, see Acoustic model construction for HTK For acoustic model construction for Sphinx, see acoustic model construction tutorial for Sphinx developed by CMU, acoustic model construction for Sphinx.