pulseaudio and speech: performance issues

[Lex]

Hi all,

I am using ubuntu 9.10 with speech-dispatcher and espeak installed from 
vinux repository, as suggested by Bill Cox. Speech-dispatcher is running 
with pulse audio output. In general, speech responsiveness on my system 
is good but not ideal. Of course it is better than with 
speech-dispatcher that comes with ubuntu 9.10 by default, but it is 
noticeably slower than NVDA + espeak on windows. Also, i remember that 
speech-dispatcher + alsa when pulseaudio is removed runs smoothier.
So my question is following: experts, do you think speech output with 
pulseaudio can be further improved or current behavior is the best ever 
possible with pulseaudio? If it is the case, i will think about 
switching to distro without such deep pulse integration.
May be i can improve the situation by editing pulseaudio config to 
decrease latency or something like that? Or will rewritten pulse 
speech-dispatcher backend fix the situation? Can we achieve the same 
perfect responsiveness as with alsa?


Thanks!

Lex