I’m trying to tune the response time performance between whisper and piper. What i’ve noticed is, I get under 3 seconds response when using a “low” voice but using for instance alba (Medium) adds a good 12 seconds to the action. 2 things:
- if i use the “Try Voice” option with alba (medium) it speaks a pretty complex sentence within like 1-2 seconds. so i don’t think its a resource issue.
- let’s say it is a resource issue just for a second. Why does it wait 12+ seconds just to turn off the light? if i change the voice to alan (low) it turns off in like 1 second. Can’t it just do the action ASAP, THEN process the voice?
for whisper, i’ve settled on tiny / 5 beams for now. its very quick with a low voice, as mentioned. This is an x86 machine.
I really like Alba’s voice! help me use it 