I am trying to get Ubuntu/Lucid working well with applications for the blind and visually impaired. Orca is working quite well now with pulseaudio. The other critical application is speakup, which reads text on the Ctrl+Alt+F1-6 consoles. Speakup has issues with pulseaudio, and I need some advice about working them out. It is important, but not required, that speakup run before the user logs in, so it can read the login prompt and some system messages. Ideally, it does this with speechd-up which talks to speech-dispatcher which generates speech through pulseadio. Thus, once speakup starts, it sounds like I need to somehow launch pulseaudio, as some user like gdm. When the user logs in, speechd-up needs to switch to talking to that user's instance of pulseadio, so that speach from both the console windows and Orca can stream through the same instance of speech-dispatcher. However, when I try this, the gdm version of pulseaudio seems to grab the sound card, and the second instance is mute. The obvious solution is for me to just run one copy of pulseaudio (as root?) at boot time and make all users access that copy, just like the older sound systems. However, I understand that is frowned upon by the pulseaudio developers. What is the right solution? Remember, I need sound both before logging in, and after, and they need to play nice, as the console windows may not be logged in, while the Gnome desktop is. I also prefer sound from speakup to work as soon as the module is loaded. I understand the old system for doing this, but not how to make it work with pulseaudio. Any help is appreciated. Thanks, Bill
