In reading through prior posts I see that VA cannot support some other speech engines (Loquendo was noted in that post). Is this to say that VA is exclusive to the Microsoft Speech API bundled with Windows? Or would it be possible to consider other SAPI 5.x/5.1 compliant engines? When I listen to what can be achieved with cloud-based engines like Tacotron 2, the MS API sounds very dated by comparison. I have played with several different voices from IVONA and find a couple, notably the British ones, to strike my ear as more fluid, but nothing like the Tacotron stuff. Having now done a fair bit with SSML, my results are improved but the tools are limited and require a fair bit of tweaking iterations to get best results.
Given that Star Citizen just confirmed their desire to provide an API, I am in hopes we will see something like EDDI coming out for SC, which really cements the preference for an all TTS approach rather than a mix of TTS and voicepack recordings. (Thats just me, I would prefer a single consistent voice than an alternating one.)
So, apologizing for the ramble:
1. Can we use an alternate speech engine?
2. Has anybody tried them to the point of finding better or worse performance?
3. If yes to 1 and 2, any recommendations?
many thanks...!