If I downloading these modules, Orca and Espeak speaks correct my two
machines, but another user writing this letter with Orca mailing list
yesterday:
"I tested alpha 3 too a few days ago.
i had no sound too. no system or other sounds.
tried it on a computer with a sb live 5.1 player sound card and on a
laptop with a conexant chipset."

Why need the snd-pcm-oss and snd-mixer-oss modules? Sunday, I read this
letter with Orca mailing list:
"Hi Everyone
For those wishing to install Ubuntu 9.04 alpha 3, I've encountered the
following problem and workaround. This, for me, was preventing the
boot-up accessibility option from working.
Problem: Portaudio, on which espeak depends, is compiled to use OSS by
default and, via padsp, is wrapped to using Pulseaudio. However, the
snd-pcm-oss and snd-mixer-oss modules required for oss output are not
loaded, rendering espeak unable to speak. I'm unsure if this happens on
all systems, but I've verified it on mine, and this issue is still
present as of the most recent lived build. While padsp does wrap espeak
to use pulse, it needs to be able to communicate with the oss device for
it to do so, hense these modules need to be loaded.
Solution: Once the livecd is booted up (note, these alpha builds do not
seem to play a login sound) open gnome-terminal (alt+f2 followed by
gnome-terminal). Issue the following command:
sudo modprobe snd-mixer-oss snd-pcm-oss
Now, to test, issue the command:
espeak hello
Espeak should say hello. Now, run orca and go through the setup. When it
asks you to log out, say no as this isn't working properly on the
livecd, at least not for me.
To log out manually, do the following (note: ctrl-alt-backspace is
disabled so this shortcut won't work anymore to kill the X session):
alt+f1, left arrow, down arrow, l, alt+l.
The disk should spin and you should then get the drum sound, indicating
that you are at the login prompt. Login with the username "ubuntu"
without the quotes. press enter twice as this user has no password. From
here on out, it is exactly like installing Ubuntu 8.10: envoke
gnome-terminal, sudo su, start orca with --no-setup and run ubiquity.
I hope this helps someone out. I'm unsure if all systems encounter this
issue, but if you do encounter it hopefully this helps you get around
the bug."

I seeing another problem, but I don't know what the metter:
When loading these modules, the espeak speech-sinthesizer and Orca
screen reader (using gnome-speech espeak driver) very slow my machines
(for example writing apple word, very later beginning spokening the
letters (two or free seconds later, but apple word already written).

Oldest Ubuntu versions (Hardy and Intrepid tested) not produces this
simptoms.
The success login sound is not played my machines.

Another problem this workaround:
When Orca is running, my desktop machine produces following cpu usages:
I have got an Athhlon64 X2 3800 Mhz cpu and 2 GB ram, the first cpu
usage is 28%, the second cpu usage is 17%, I think this is not good.

If you would like test this problem, do following steps:
1. Download the current dayly live cd and try boot.
After pressing enter key your preferred language, press f5 key, and
choosing screen reader option.

What happen if you not loading this two modules? If your machine
producing my problems, Orca not speaking.
And what happen if loaded and rerun Orca? My machine Orca is speaking
this workaround, but producing wroted simptoms.

Have any ydeas?

Attila