HMM-based Speech Synthesis System (HTS) - Extensions

Extensions †

Some of HTS users have implemented substantial extensions to HTS. On this page we provide short descriptions and links to further information. Many thanks to the authors of these extensions.

If you provide some free extension to HTS for download and would like to see it listed on this page, please contact us.

DFKI MARY †

DFKI MARY speech synthesis system provides voice import tools for creating your own HMM-based voices for MARY. It uses versions of the speaker dependent and adaptive demo scripts provided by HTS adapted to the MARY platform. For details, please see tutorial.

Marc Schroeder

Multi-process supporting patch for HTS-demo_CMU-ARCTIC-SLT †

This patch uses JM command of HHEd in HTS-2.1 to provide an option for clustering in parallel. In a training environment with multi-core CPUs, the clustering time of HHEd could be generally reduced. The patch is applied to HTS-demo_CMU-ARCTIC-SLT (ver.2.1) only. Please visit this page.

Xuchen Yao and Feng Ding

Open-Source LVCSR Engine Julius †

The open-source LVCSR engine Julius now supports multi-stream HMMs with stream-depdendent parameter tying structure and multi-space probability distribution HMMs (MSD-HMMs) built by HTS. You can use acoustic models trained by HTS directly. Please check announce and update pages on the Julius user forum for details.

Akinobu Lee

Prototype definition modification for HTK compatibility †

Due to the different format of the prototypes used by HTS and HTK, the demo for testing for a successful installation provided on the HTK website will not work with the HTS extension added. With the modification made in this, HTKDemo will run correctly with the HTS extension. This will enable users who are installing HTS to determine if they have installed everything correctly.

Gavin Potter