I'm using the Carnegie Mellon Pronouncing Dictionary for a project and it stores words as a pair consisting of the word and then the phonemes in Arpabet style...
hello -HH AH0 L OW1
I'd like to be able to feed the phones back into a Text-To-Speech engine, but here I get a bit lost - my 'say' command on OSX doesn't accept the same set of phonemes (and nethier does espeak) - is there a (free) system for OSX that let's me input phonemes in Arpabet style and get audio out? (Or, for that matter, something that converts between different sets of phonemes?)
1 Answer
For eSpeak, you need a Arpabet to IPA table. With that you can convert the CMU-DICT to IPA and then use it with eSpeak. A different mapping might be needed for say as phone set system it uses might be different.
