Back

Tell your friends, The post on my site at https://eurpod.com/piper-voices-of-retro-speech-synthesizers/ has been updated to feature a link to the releases page for the amazing driver Sam Tupy (@samtupy) is producing for BESTSpeech (Keynote gold). I honestly don't know if I'll maintain the older driver. There's certain hacks and character conversions that I could in theory apply and improve it a little, but I don't think that will fix the squealing problem with rapid speech interruptions, because that's a direct problem of rapid teardown and re-initialization of the DLL, along with not having access to the raw data. In technicality, we also didn't have any async callback to the DLL (or really any callback at all), just a straight-up, "here, speak this chunk of text, and hopefully after you're done, we can close the engine. But when you're done, we have no idea of, so hopefully we get to close it in time before you're actually done!" So the work that Sam did to "modernize" this synthesizer is nothing short of a miracle, to be honest, now it behaves way closer to what you would see in Eloquence, DECTalk, any comprehensive synthesizer, and for that I'm grateful.