getting orca and espeakup/speakup to work together

That's a long standing issue with Debian and derivatives' handling of audio cf. the Debian accessibility mailing list:
https://lists.debian.org/debian-accessibility/
And especially the thread that begins with this message:
https://lists.debian.org/debian-accessibility/2018/10/msg00000.html

I tried to help on that, to no avail.

Workarounds that I know as of today:
1) Remove pulseaudio and install the libspeakup-ng Debian package patched by Samuel:
https://lists.debian.org/debian-accessibility/2018/11/msg00065.html
Instructions:
https://people.debian.org/~sthibault/tmp/sid-tmp/libespeak-ng1_1.49.2+dfsg-7~0_amd64.deb
However, I don't know if theses packages can be used wuth stretch
@Samuel: please tell us if it's possible, or if you will backport these packages.
Caveat: I can't provide help to remove pulseaudio (which I generally do not recommend) as I am neither a Debian nor Ubuntu user.

2) Install a distribution that does not have this issue, like Slint.

Didier
--
Didier Spaier
http://slint.fr

What is the trick to getting speakup with software speech and orca to work at the same time. I have both debian stretch and ubuntu bionic systems and on both machines, I have to disable espeakup to get orca to work. If I disable espeakup via the command, "systemctl disable espeakup" and reboot, orca works as normal. If I re-enable espeakup via, "systemctl enable espeakup", then I get no speech on the login screen or after I log in. After I log in, even playing sounds via aplay does not work.
I have a vague memory of having to recompile espeakup or something to get that to work.

Samuel Thibault

2018-11-19 21:21:20 UTC

Hello,

https://lists.debian.org/debian-accessibility/
https://lists.debian.org/debian-accessibility/2018/10/msg00000.html
I tried to help on that, to no avail.
https://lists.debian.org/debian-accessibility/2018/11/msg00065.html
https://people.debian.org/~sthibault/tmp/sid-tmp/libespeak-ng1_1.49.2+dfsg-7~0_amd64.deb
However, I don't know if theses packages can be used wuth stretch

There is no need to install that package in Stretch. The issue that
that package is fixing is due to the pause implementation meant to work
around the very issue he is having. That pause implementation wasn't in
Stretch, it's only in Buster.

The pause implementation requires a Linux kernel change, so that won't
be backported to Stretch unfortunately.

Post by Didier Spaier
Caveat: I can't provide help to remove pulseaudio (which I generally do not recommend)

In Stretch, there is no real choice: unless configuring dmix, having
pulseaudio installed necessarily means conflicts between the espeakup
daemon and Orca.

In Buster, it's not a solution, since the version of firefox there
requires pulseaudio. Thus the pause implementation in espeakup, to
make it release the sound board for Orca to take it. Conversely, when
switching off from the X session to the Linux console, we'd need to make
Orca release the sound board. It actually happens already when switching
to a Linux console where the user is not logged in, because pulseaudio
releases the sound board in that case.

Now, as you mentioned, Slint does not have the issue, because it uses
the dmix plugin. Now, to repeat myself:

- it would be useful to document how to configure dmix on the wiki
https://wiki.debian.org/accessibility

- enabling dmix by default in Debian could be an option, it "just" needs
to be discussed with the alsa maintainers. If nobody takes the time to
do this, we'll stay with the issue.

Samuel

John G Heim

2018-11-20 16:13:46 UTC

I found a vague howto on getting dmix to work but it says that if you
use dmix, you have to configure each sound app to use it. Is that right?
It seems more trouble than it is worth.

Is this problem part of the kernel-space vs user-space problem? I would
settle for getting speech during start up. I can probably put something
in the X11 configuration to disable espeakup once the GUI starts up. But
I am wondering if there is a path forward on this.

future.

Post by Samuel Thibault
Hello,

There is no need to install that package in Stretch. The issue that
that package is fixing is due to the pause implementation meant to work
around the very issue he is having. That pause implementation wasn't in
Stretch, it's only in Buster.
The pause implementation requires a Linux kernel change, so that won't
be backported to Stretch unfortunately.

Post by Didier Spaier
Caveat: I can't provide help to remove pulseaudio (which I generally do not recommend)

In Stretch, there is no real choice: unless configuring dmix, having
pulseaudio installed necessarily means conflicts between the espeakup
daemon and Orca.
In Buster, it's not a solution, since the version of firefox there
requires pulseaudio. Thus the pause implementation in espeakup, to
make it release the sound board for Orca to take it. Conversely, when
switching off from the X session to the Linux console, we'd need to make
Orca release the sound board. It actually happens already when switching
to a Linux console where the user is not logged in, because pulseaudio
releases the sound board in that case.
Now, as you mentioned, Slint does not have the issue, because it uses
- it would be useful to document how to configure dmix on the wiki
https://wiki.debian.org/accessibility
- enabling dmix by default in Debian could be an option, it "just" needs
to be discussed with the alsa maintainers. If nobody takes the time to
do this, we'll stay with the issue.
Samuel

Zachary Kline

2018-11-20 16:33:35 UTC

I thought that Dmix was configured automatically in most recent versions of Alsa. The problem seems to specifically lie between espeakup and PUlseaudio, as far as I can tell.

I found a vague howto on getting dmix to work but it says that if you use dmix, you have to configure each sound app to use it. Is that right? It seems more trouble than it is worth.
Is this problem part of the kernel-space vs user-space problem? I would settle for getting speech during start up. I can probably put something in the X11 configuration to disable espeakup once the GUI starts up. But I am wondering if there is a path forward on this.
future.

Post by Samuel Thibault
Hello,

There is no need to install that package in Stretch. The issue that
that package is fixing is due to the pause implementation meant to work
around the very issue he is having. That pause implementation wasn't in
Stretch, it's only in Buster.
The pause implementation requires a Linux kernel change, so that won't
be backported to Stretch unfortunately.

Post by Didier Spaier
Caveat: I can't provide help to remove pulseaudio (which I generally do not recommend)

In Stretch, there is no real choice: unless configuring dmix, having
pulseaudio installed necessarily means conflicts between the espeakup
daemon and Orca.
In Buster, it's not a solution, since the version of firefox there
requires pulseaudio. Thus the pause implementation in espeakup, to
make it release the sound board for Orca to take it. Conversely, when
switching off from the X session to the Linux console, we'd need to make
Orca release the sound board. It actually happens already when switching
to a Linux console where the user is not logged in, because pulseaudio
releases the sound board in that case.
Now, as you mentioned, Slint does not have the issue, because it uses
- it would be useful to document how to configure dmix on the wiki
https://wiki.debian.org/accessibility
- enabling dmix by default in Debian could be an option, it "just" needs
to be discussed with the alsa maintainers. If nobody takes the time to
do this, we'll stay with the issue.
Samuel

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Samuel Thibault

2018-11-20 19:12:52 UTC

I found a vague howto on getting dmix to work but it says that if you use
dmix, you have to configure each sound app to use it. Is that right?

No, you can configure it globally.

Is this problem part of the kernel-space vs user-space problem?

No, it's espeakup, i.e. a userland daemon which accesses the sound
board.

I would settle for getting speech during start up. I can probably put
something in the X11 configuration to disable espeakup once the GUI
starts up. But I am wondering if there is a path forward on this.

The path forward was implemented in time for Buster: espeakup shuts
itself down when switching to graphical mode.

Samuel

Janina Sajka

2018-11-21 13:57:17 UTC

I believe it's for each device, not each app. Note that you will have
multiple devices on a given card, e.g. hdmi and analog.

There are a lot of poorly written docs out there.

I found a vague howto on getting dmix to work but it says that if you use
dmix, you have to configure each sound app to use it. Is that right? It
seems more trouble than it is worth.
Is this problem part of the kernel-space vs user-space problem? I would
settle for getting speech during start up. I can probably put something in
the X11 configuration to disable espeakup once the GUI starts up. But I am
wondering if there is a path forward on this.
future.

Post by Samuel Thibault
Hello,

There is no need to install that package in Stretch. The issue that
that package is fixing is due to the pause implementation meant to work
around the very issue he is having. That pause implementation wasn't in
Stretch, it's only in Buster.
The pause implementation requires a Linux kernel change, so that won't
be backported to Stretch unfortunately.

Post by Didier Spaier
Caveat: I can't provide help to remove pulseaudio (which I generally do not recommend)

In Stretch, there is no real choice: unless configuring dmix, having
pulseaudio installed necessarily means conflicts between the espeakup
daemon and Orca.
In Buster, it's not a solution, since the version of firefox there
requires pulseaudio. Thus the pause implementation in espeakup, to
make it release the sound board for Orca to take it. Conversely, when
switching off from the X session to the Linux console, we'd need to make
Orca release the sound board. It actually happens already when switching
to a Linux console where the user is not logged in, because pulseaudio
releases the sound board in that case.
Now, as you mentioned, Slint does not have the issue, because it uses
- it would be useful to document how to configure dmix on the wiki
https://wiki.debian.org/accessibility
- enabling dmix by default in Debian could be an option, it "just" needs
to be discussed with the alsa maintainers. If nobody takes the time to
do this, we'll stay with the issue.
Samuel

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

--
Janina Sajka

Linux Foundation Fellow
Executive Chair, Accessibility Workgroup: http://a11y.org

The World Wide Web Consortium (W3C), Web Accessibility Initiative (WAI)
Chair, Accessible Platform Architectures http://www.w3.org/wai/apa

Janina Sajka

2018-11-20 15:31:03 UTC

Pulseaudio is supposed to eliminate this problem, but I've never been
able to rely on pulseaudio working for me.

Under alsa both espeakup and speech-dispatcher are likely unable to
share the same hardware device. This is usually the case for me,
requiring I send the output of each to a different device.

Check carefully with a command like:

aplay -l |grep card

You're looking for cards that are labeled "analog."

On one of my laptops I have both hw:0,0 and hw0,2 labeled analog. I
simply set speech-dispatcher to use hw:0,2 in
/etc/speech-dispatcher/speechd.conf and all works as expected.

Janina

What is the trick to getting speakup with software speech and orca to work
at the same time. I have both debian stretch and ubuntu bionic systems and
on both machines, I have to disable espeakup to get orca to work. If I
disable espeakup via the command, "systemctl disable espeakup" and reboot,
orca works as normal. If I re-enable espeakup via, "systemctl enable
espeakup", then I get no speech on the login screen or after I log in. After
I log in, even playing sounds via aplay does not work.
I have a vague memory of having to recompile espeakup or something to get
that to work.
--
--
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Samuel Thibault

2018-11-20 19:11:05 UTC

Post by Janina Sajka
Pulseaudio is supposed to eliminate this problem,

If espeakup was running as the same user as the one who is logged in,
yes. But it'd mean you wouldn't have screen reading before logging in.

Post by Janina Sajka
Under alsa both espeakup and speech-dispatcher are likely unable to
share the same hardware device.

Indeed, unless enabling the dmix plugin.

Using different sound boards is another solution indeed.

Samuel

John G Heim

2018-11-20 21:49:10 UTC

Is there any way to get speakup to work through the PC speaker?

Post by Janina Sajka
Pulseaudio is supposed to eliminate this problem,

If espeakup was running as the same user as the one who is logged in,
yes. But it'd mean you wouldn't have screen reading before logging in.

Post by Janina Sajka
Under alsa both espeakup and speech-dispatcher are likely unable to
share the same hardware device.

Indeed, unless enabling the dmix plugin.
Using different sound boards is another solution indeed.
Samuel

Gregory Nowak

2018-11-20 23:17:36 UTC

I remember there used to be a way to use the pc speaker as a sound
card. I seem to recall it was a kernel driver, but it could have been
through alsa itself. A quick web search gives me a linux journal
article from 1997, but nothing besides that. So, short of checking the
config choices for a modern 4.x kernel, I don't know if this is still
an option. Even if it is still an option, I suspect the speech
wouldn't be very intelligible.

Greg

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

--
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-***@EU.org

Jude DaShiell

2018-11-20 23:37:36 UTC

thinkpenguin.com sells sound cards that plug into usb slots and you plug
speakers into the other end of the card. On my system I have only one
analog card. Since I bought one of these usb cards I think I'll plug it
in and see if my system suddenly finds a second analog card.

Date: Tue, 20 Nov 2018 18:17:36
Reply-To: Speakup is a screen review system for Linux.
Subject: Re: getting orca and espeakup/speakup to work together
I remember there used to be a way to use the pc speaker as a sound
card. I seem to recall it was a kernel driver, but it could have been
through alsa itself. A quick web search gives me a linux journal
article from 1997, but nothing besides that. So, short of checking the
config choices for a modern 4.x kernel, I don't know if this is still
an option. Even if it is still an option, I suspect the speech
wouldn't be very intelligible.
Greg

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

Janina Sajka

2018-11-21 14:19:36 UTC

I routinely have two USB sound cards on my desktop, plus an old AV-710
based PCI card, plus an RME Hammerfall. And, yes, I use them all in
particular ways.

Decent USB sound cards, with sufficient quality for good sounding
speech, can be had cheap.

Janina

Post by Jude DaShiell
thinkpenguin.com sells sound cards that plug into usb slots and you plug
speakers into the other end of the card. On my system I have only one
analog card. Since I bought one of these usb cards I think I'll plug it
in and see if my system suddenly finds a second analog card.

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

--
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Tom Fowle

2018-11-21 04:32:51 UTC

Since the PC speaker is driven from one bit with no smoothing or filtering,
the results would probably be nothing short of ghastly.

I seem to recall an experimental dos TTS that tried it with expected
results.
Tom Fowle

Post by Gregory Nowak
I remember there used to be a way to use the pc speaker as a sound
card. I seem to recall it was a kernel driver, but it could have been
through alsa itself. A quick web search gives me a linux journal
article from 1997, but nothing besides that. So, short of checking the
config choices for a modern 4.x kernel, I don't know if this is still
an option. Even if it is still an option, I suspect the speech
wouldn't be very intelligible.
Greg

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

--
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.
--
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

John G Heim

2018-11-21 13:34:41 UTC

Nah, it would be fine for emergencies. For example, suppose your GUI
interface freezes, orca dies or at least stops responding. You would
still be able to get to a character console and start killing processes.

Another thing would be that you could have speakup read boot messages
via the PC speaker so at least you'd know it is booting.

Post by Tom Fowle
Since the PC speaker is driven from one bit with no smoothing or filtering,
the results would probably be nothing short of ghastly.
I seem to recall an experimental dos TTS that tried it with expected
results.
Tom Fowle

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

--
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.
--
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Janina Sajka

2018-11-21 14:24:30 UTC

Well, another way to get similar results is to launch the gui by hand,
not via systemd. Long live startx.

The more things you start automagically, the more things there are that
can, and eventually will go awry. Why ask for more trouble? The
brilliance of Speakup is that it can be available to work even when the
boot fails. But, if you're starting the gui automatically, and it fails,
you have much less recourse.

Years ago I discovered the wisdom of this approach. I'd plug my laptop
into some hotel's ethernet, only to have my boot fail over some gui
based message about accessing their network. I discovered I could put
myself back in control by booting to a console login and then dealing
with bringing up the network.

Janina

Post by John G Heim
Nah, it would be fine for emergencies. For example, suppose your GUI
interface freezes, orca dies or at least stops responding. You would still
be able to get to a character console and start killing processes.
Another thing would be that you could have speakup read boot messages via
the PC speaker so at least you'd know it is booting.

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

--
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.
--
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Samuel Thibault

2018-11-20 23:42:06 UTC

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

John Covici

2018-11-21 15:05:37 UTC

How about using speech dispatcher for speakup? Would that make things
work better together since orca also uses speech dispatcher?

On Tue, 20 Nov 2018 18:42:06 -0500,

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

As a matter of fact: yes. You can use the snd-pcsp Linux kernel module,
which makes a sound board out of the PC speaker. You will need
to unload the pcspkr module to be able to load the snd-pcsp module.
Samuel
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici wb2una
***@ccs.covici.com

Samuel Thibault

2018-11-21 18:24:04 UTC

Post by John Covici
How about using speech dispatcher for speakup? Would that make things
work better together since orca also uses speech dispatcher?

John Covici

2018-11-22 07:13:03 UTC

I run orca with speech dispatcher using Alsa and if I want to play
audio, I enable spawn=yes in my /etc/pulse/client.conf. Maybe not the
best solution, but its what I have come up with so far.

On Wed, 21 Nov 2018 13:24:04 -0500,

Post by John Covici
How about using speech dispatcher for speakup? Would that make things
work better together since orca also uses speech dispatcher?

You'd have to run speech-dispatcher as a system service for both the
root-run espeakup and the user-run orca to be able to use it.
But then you can't play audio as user in your Xorg session since
speech-dispatcher will keep the audio open as root.
Samuel
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici wb2una
***@ccs.covici.com

Janina Sajka

2018-11-21 14:16:30 UTC

Man, those have become as rare as hen's teeth. And they're not about to
make a resurgence.

Furthermore, they tend to be buried in amongst all kinds of other
hardware. Is that really how you would want to orient your speaking monitor?
Deep in the bowels of some computer case?

I know I want to go exactly the opposite way. I've my audio out running
through a desktop Mackie mixer and usually talking through some quality
desktop speakers. If I need, I can route to some serious JBL IONs. I'm
contemplating adding Google Cast to my options.

Janina

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

Post by Janina Sajka
Pulseaudio is supposed to eliminate this problem,

If espeakup was running as the same user as the one who is logged in,
yes. But it'd mean you wouldn't have screen reading before logging in.

Post by Janina Sajka
Under alsa both espeakup and speech-dispatcher are likely unable to
share the same hardware device.

Indeed, unless enabling the dmix plugin.
Using different sound boards is another solution indeed.
Samuel

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

John G Heim

2018-11-21 14:52:06 UTC

I don't think PC speakers are rare at all. I am responsible for over 100
desktops and every one of them has a PC speaker. Of course, they are all
Dells. But even the machines I built myself for my own use with Asus
mobos have PC speakers. I find it hard to believe there exist mobos that
don't have a PC speaker connector. The machine might not have anything
connected to it but the connector is there. It's like with serial
ports. There are plenty of mobos available with serial port header
blocks. You just have to buy a cable to get the DB9 port. You might
have to buy a PC speaker if you're building your own machine.

It's worth a couple of extra bucks to get a machine with a PC speaker
because then you can configure a grub init tune.

Post by Janina Sajka
Man, those have become as rare as hen's teeth. And they're not about to
make a resurgence.
Furthermore, they tend to be buried in amongst all kinds of other
hardware. Is that really how you would want to orient your speaking monitor?
Deep in the bowels of some computer case?
I know I want to go exactly the opposite way. I've my audio out running
through a desktop Mackie mixer and usually talking through some quality
desktop speakers. If I need, I can route to some serious JBL IONs. I'm
contemplating adding Google Cast to my options.
Janina

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

Post by Janina Sajka
Pulseaudio is supposed to eliminate this problem,

If espeakup was running as the same user as the one who is logged in,
yes. But it'd mean you wouldn't have screen reading before logging in.

Post by Janina Sajka
Under alsa both espeakup and speech-dispatcher are likely unable to
share the same hardware device.

Indeed, unless enabling the dmix plugin.
Using different sound boards is another solution indeed.
Samuel

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Jude DaShiell

2018-11-21 17:38:16 UTC

If you get the right case, you can get something other than a 2 inch pc
speaker too.

Date: Wed, 21 Nov 2018 09:52:06
Subject: Re: getting orca and espeakup/speakup to work together
I don't think PC speakers are rare at all. I am responsible for over 100
desktops and every one of them has a PC speaker. Of course, they are all
Dells. But even the machines I built myself for my own use with Asus mobos
have PC speakers. I find it hard to believe there exist mobos that don't have
a PC speaker connector. The machine might not have anything connected to it
but the connector is there. It's like with serial ports. There are plenty of
mobos available with serial port header blocks. You just have to buy a cable
to get the DB9 port. You might have to buy a PC speaker if you're building
your own machine.
It's worth a couple of extra bucks to get a machine with a PC speaker because
then you can configure a grub init tune.

Post by John G Heim
Is there any way to get speakup to work through the PC speaker?

Post by Janina Sajka
Pulseaudio is supposed to eliminate this problem,

If espeakup was running as the same user as the one who is logged in,
yes. But it'd mean you wouldn't have screen reading before logging in.

Post by Janina Sajka
Under alsa both espeakup and speech-dispatcher are likely unable to
share the same hardware device.

Indeed, unless enabling the dmix plugin.
Using different sound boards is another solution indeed.
Samuel

_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Janina Sajka

2018-11-21 14:12:07 UTC

Post by Janina Sajka
Pulseaudio is supposed to eliminate this problem,

If espeakup was running as the same user as the one who is logged in,
yes. But it'd mean you wouldn't have screen reading before logging in.

I suppose I could try launching espeakup as my ordinary user, but I'm
loathe to go through the hassle of putting pulseaudio back on this box
just to test something. It's a pernicious app that can be hard to get
rid of, and I never found it did anything useful for me on those
occasions when I did have it working.

Nowadays I'm rather starting espeakup via systemd.

On those occasions when my system doesn't come up talking, I assume I'm
being asked for root password to perform system maintanance.
Fortunately, I still get a beep on backspace to confirm this once I give
the root passwd. At that point I can:

modprobe speakup_soft; espeakup

After dealing with the problem, usually an e2fsck as I have a failing
drive, a ctrl-d takes me to login.

In this situation, I guess espeakup is running as root?