More progress, I learned the audio support for espeakup is provided by espeak a separate package and it's supposed to be possible to compile espeak against pulseaudio and rebuild and once done flexibility then ought to increase on this end. More later. On Sun, 6 Mar 2016, Tanu Kaskinen wrote: > Date: Sun, 6 Mar 2016 05:49:56 > From: Tanu Kaskinen <tanuk at iki.fi> > To: Jude DaShiell <jdashiel at panix.com>, > pulseaudio-discuss at lists.freedesktop.org > Subject: Re: [pulseaudio-discuss] nearly solved pulseaudio configuration > > On Sat, 2016-03-05 at 20:19 -0500, Jude DaShiell wrote: >> Thanks to tanuk at iki.fi I got vlc playing over my usb speakers. The screen >> reader speech is still going out over the 3.5MM jack speakers. My guess >> is I may have to fix asound.state or asound.conf I'll find out in a little >> while. My /etc/pulse/default.pa to get this far follows the last line I >> modified with the help of tanuk at iki.fi to get this working to the point it >> works now. The rest is as installed on talkingarchlinux at >> http://talkingarch.tk/ and I put an r.f.e. in with that distro to have usb >> ports checked for speakers and doing sound card output search selection: > > Usually applications don't choose the device themselves when using > pulseaudio. I can think of three explanations why the screen reader > would not be affected by the default sink setting: > > 1) The screen reader actually does explicitly select the internal sound > card sink rather than the usb sink. In this case contacting the > talkingarchlinux folks or the screen reader developers is the right > approach. > > 2) You have at some point manually moved the screen reader stream to > the internal sound card sink, and pulseaudio remembers that as your > preference. In this case the default sink setting is ignored. You can > fix this by passing paramter "restore_device=no" to module-stream- > restore, so change this line in default.pa: > > load-module module-stream-restore > > to > > load-module module-stream-restore restore_device=no > > That change makes it so that manual stream moves are not remembered, so > the default sink setting always determines where new streams are > routed. > > 3) The screen reader doesn't go through pulseaudio, but accesses the > internal sound card directly. You can check this with this command: > > lsof /dev/snd/pcm* > > The command prints a table of information, with the first line > containing the table column headers. The subsequent lines contain > information about the processes that are using the sound cards. The > first column contains the process name, and if the screen reader > appears there, it's using the sound card directly. > > -- > Tanu > --
