Hello all, I mentioned in an earlier post that software speech with speakup running under a kernel iwith one of the pre-emption models turned on is dodgy, as it produces unwanted spelling of words in inappropriate places, and can be sluggish. How would you overcome this? Cheers, Josh