Silence trimming should trim even more when using one core

[hozosch]

Is your feature request related to a problem? Please describe.

Silence trimming in conjunction with Wasapi, in my view, is the best thing that's been done in a while. I love how much more responsive things have become, and this is also very noticeable with one core. However, when compared to things such as sapi5 ESpeak, it still isn't as responsive.

Describe the solution you'd like

I suspect that there still isn't enough silence being trimmed with one core. But this is also voice specific, for example, the austrian German voice is nice and responsive. However, other voices are lagging a bit behind. I suspect that there's still silence which could be trimmed, but I don't know how easy it is to be implemented. Nevertheless, I wanted to bring this up for it to be investigated.

Describe alternatives you've considered

We could leave it as is, as the way it is now already has a great impact. That's the only other way I can think of, but responsiveness should be a top priority for a good screen reader.

Additional context

[gexgd0419]

Could you provide some more detail of what voice you think is lagging a bit behind? For example, the name and the language of the voice? If the voice has a SAPI5 version, does the SAPI5 version lag as well?

[hozosch]

I think I notice it most with the swiss German voice. It's the only one for that language, and there is no sapi 5 version of it, unless you would do a registry hack.