Has anyone had experience using TTS voices generated outside of the NeoSpeech voices that come with Cp? I am specifically looking at Amazon’s “Polly”. I suspect I can’t edit/generate within Cp, but will have to import .wav files.
I recently used it in a project I revised to demonstrate the interactive video in CP 2019. I had no issues with it. To me, Polly’s voices sound better than ones in NeoSpeech.
Does anyone know of a good “getting started” blog post for Polly? I have an AWS account, but am concerned that I’ll do something expensive while trying to generate text for a proof of concept. As I understand it, you pay by the character?
I just completed a brief proof-of-concept for a client using Polly TTS. It took a while to figure out a sensible use (for me) of the SSML input feature. I decided to skip mucking about with the lexicon and just added IPA pronunciation when needed by using tags.
As with all TTS, it helps to have a good handle on phonetics. (Thanks Mom and Dad for letting my fritter away my undergrad degree in linguistics!) I really like being able to add breath sounds, breaks, and prosodic features to make the voices sound more natural.
I’m a bit disappointed that Polly doesn’t output WAV files, but the MP3s sound okay when added to Captivate and published.
Hope these comments help. Give a shout, Alice, if you’d like more info: leslie@lesliebivens.com
You must be logged in to post a comment.
