Evaluation

There is need for a speech synthesis toolkit, like the modules that are already available to speed up speech recognition development.

The community has made an excellent start with Festival and MBROLA , but in order to push the technology forward faster we need more people involved to further build on these efforts. Work is needed on:

• Standards for component interfaces
• Tools (automatic segmentation, statistical training algorithms)
• Corpora (both speech and text, raw and annotated)
• TTS modules
• Add-on modules (email pre-processing)

This page has been set up to hold contributed modules, interfaces, components and tools

It is currently still at the proposal stage, but we are waiting for your contributions

Another source of information on Speech Synthesis evaluation is the TTS page in the Human Language Technologies Evaluation web site.