Change synthesizer module



Currently we are using a command line script to call the Daisy Pipeline. We use a custom preprocess method and use a custom voice, build with CereVoice.

For better performance, we would like to swap SAPI for a CereProc API. Is it possible to do this with the Daisy Pipeline and how would we approach this in the best way?

Besides this, we use specific tags within our data, for example we use to use a different instance of a unit. When I directly feed this to our Cereproc Voice, it works perfectly. But when I process it through the daisy pipeline, it seems that these tags are stripped before they reach the TTS engine. How can we make sure that additional tags (also added to the DTD) will reach the TTS voice engine?