Define how to load custom voices

[jankapunkt]

Currently I can't see how to load custom voices and I think it is currently not defined and therefore not implemented by any vendor.

We have the use-case where we want to create a custom voice (for example using the mozilla/tts stack) and load this voice so we can have a unified user experience.

[Eddcapone]

It is a pitty that this is still without a reply.

[jankapunkt]

Still no reply here. In the meantime I tried to setup some remote server, serving some voice files I could extract from eSpeak, however I never got a custom SpeechSynthesisVoice implemented.

It always tells me

Uncaught TypeError: SpeechSynthesisUtterance.voice setter: Value being assigned does not implement interface SpeechSynthesisVoice.

when I try to add my voice implementation to a SpeechSynthesisUtterance instance.

In my opinion this is such a crucial feature, especially for non-commercial projects that have not the funding to pay these online services to generate TTS for all those browsers that have no sufficient voices available.

[Neruell]

Agree, would love to see some sort of functionality like that.