Working on and off using weekends and other spare time, I finally managed to import my voice into MaryTTSusing my MacbookPro without a VM! It’s still not working as I expected but the overall process completes from beginning to end. I feel like I’m back where I left off several months prior. Before I was working in a Linux environment and other audio issues were plaguing me as I could not hear the imported audio files in one of the import steps. This time around it looks like all of the tools work without nasty side effects. If I can only get my audio generation process further refined I might be onto something.
Advertisement
This is a totally useless and totally cool project ! Keep going and give us a a demo when you’re done
By: vincent daubry on November 4, 2010
at 4:56 am
Useless? Actually it has its uses. For example, the installed English voice doesn’t pronounce certain things accurately. You can build a voice that is specific to something like driving directions by building it with driving directions narrative. That way all turns would be pronounced clearly. Or you could build a voice specific to assembling a certain product. In this case all of the components, tools involved and assembly related narrative would be spoken with high quality. Call it domain specific voice building.
By: Cliff on November 11, 2010
at 4:42 pm