Pulse Audio

Robert Spangler spangler.robert at gmail.com
Sun Mar 9 21:10:03 EDT 2014 

Are you using Debian Sid?  Mine speaks on the login screen with gdm3.


On 3/9/2014 1:11 PM, Trevor Astrope wrote:
> Make sure your user is in the pulse-access group. I also have
> Debian-gdm, speech-dispatcher and root in pulse-access. I added these
> users to the audio group too, for good measure.
>
> For orca, I don't get speech until I login. It selects my user by
> default and places focus on the password field, so I just need to enter
> my password and then it will start speaking.
>
> On Sun, 9 Mar 2014, Robert Spangler wrote:
>
>> I am running Jesse/Sid.  I tried all of these things with no luck.  I
>> even tried Janina's suggestion of disabling Pulseaudio (using her
>> steps) and that just resulted in Orca not speaking.
>>
>> I thought I would try a little experimenting of my own, so I
>> recompiled espeak to use Pulseaudio.  I then configured Pulseaudio to
>> run as a system service.  When I attempted to restart espeakup, it
>> said it couldn't connect to PulseAudio: access denied.
>>
>> On another note: I noticed, when booting, that speech starts before
>> the Pulseaudio daemon.  Could this be causing a problem?
>>
>> I'll keep hammering away at this.  I agree with another poster that
>> someone should fix this bug.
>>
>> Thanks,
>> Rob
>>
>> On 3/7/2014 10:48 AM, Trevor Astrope wrote:
>>> On Fri, 7 Mar 2014, John G. Heim wrote:
>>>
>>>> I think the  classic fix for this probme is to recompile espeak to use
>>>> pulseaudio.  Here is a link to a bug report explaining (sort of) what
>>>> to do.
>>>> https://bugzilla.redhat.com/show_bug.cgi?id=481651
>>>>
>>>> I think we had a discussion of this point back in November of 2013 but
>>>> i can't find the thread on google. I am still running debian squeeze
>>>> on my every day work machine and all my other machines now use sonar
>>>> linux. So I haven't tried to get speakup and orca to play nice for
>>>> several months. But I know I got speakup and orca to work in wheezy by
>>>> following the advice on that bug report. The real problem was that the
>>>> version of orca in debian wheezy
>>>> didn't work.
>>>
>>> Below is the post from you and Mike Ray on how to configure and build
>>> espeak to use pulseaudio. I run pulseaudio as a daemon to get espeakup
>>> to start speaking as soon as it is loaded so I get speech at the login
>>> prompt.
>>>
>>> In Debian Jessie/Sid, you will need to edit /etc/default/pulseaudio to
>>> have this line:
>>>
>>> PULSEAUDIO_SYSTEM_START=1
>>>
>>> And in /etc/pulse/client.conf:
>>>
>>> autospawn = no
>>>
>>> It's been a while since I configured speech for orca, but I think I had
>>> to modify /etc/speech-dispatcher/speechd.conf to use a unix socket:
>>>
>>> SocketPath "/var/run/speech-dispatcher/speech-dispatcher.sock"
>>>
>>> And in my .bash_profile, I added:
>>>
>>> export
>>> SPEECHD_ADDRESS="unix_socket:/var/run/speech-dispatcher/speech-dispatcher.sock"
>>>
>>>
>>>
>>> I use hardware speech with speakup on the machine running orca and I
>>> don't run orca on the machine using espeakup, so I can't say that it
>>> will work for both orca and speakup with espeakup. Please follow up if I
>>> missed something.
>>>
>>> ----- Forwarded Message -----
>>> Date: Fri, 1 Nov 2013 22:17:03
>>> From: John G. Heim <jheim at math.wisc.edu>
>>> Reply-To: Speakup is a screen review system for Linux.
>>> <speakup at linux-speakup.org>
>>> To: mike at raspberryvi.org, Speakup is a screen review system for Linux.
>>> <speakup at linux-speakup.org>, Brandon McGinty-Carroll
>>> <bmmcginty at bmcginty.hopto.org>
>>> Subject: Re: speakup & orca in debian 7
>>>
>>> I got it working on my debian 7 (wheezy) system. Here is what I did:
>>>
>>> apt-get remove espeakup espeak
>>> apt-get install libpulse-dev libsonic-dev
>>> apt-get source espeak
>>> cd espeak-1.6.42/src/
>>> [edit Makefile as describe below]
>>> make
>>> make install
>>>
>>> cd ../../
>>> apt-get install espeakup
>>> cd espeakup-0.71/
>>> make
>>> make install
>>>
>>> modprobe speakup-soft
>>> espeakup
>>> Viola!
>>>
>>> Notes:
>>> 1. I couldn't figure out how to get the debian espeakup package to work
>>> with my custom compiled version of speakup. You can't install espeakup
>>> w/o the espeak package.
>>> 2. The apt-get source command downloads the source for the debian
>>> package into the current working directory.
>>>
>>> On 11/01/2013 07:28 PM, Mike Ray wrote:
>>>>
>>>> I should also say the command to compile is:
>>>>
>>>> make all
>>>>
>>>> not just make
>>>>
>>>> Mike
>>>>
>>>> On 02/11/2013 00:22, Mike Ray wrote:
>>>>>
>>>>> Here is what I did:
>>>>>
>>>>> Edit the Makefile and find the section which reads like this:
>>>>>
>>>>> # 'runtime' uses pulseaudio if it is running, else uses portaudio
>>>>> #AUDIO = runtime
>>>>> AUDIO = portaudio
>>>>> #AUDIO = portaudio0
>>>>> #AUDIO = portaudio2
>>>>> #AUDIO = pulseaudio
>>>>> #AUDIO = sada
>>>>>
>>>>> And simply change it to read thusly:
>>>>>
>>>>> # 'runtime' uses pulseaudio if it is running, else uses portaudio
>>>>> #AUDIO = runtime
>>>>> #AUDIO = portaudio
>>>>> #AUDIO = portaudio0
>>>>> #AUDIO = portaudio2
>>>>> AUDIO = pulseaudio
>>>>> #AUDIO = sada
>>>>>
>>>>> As you can see I just commented out portaudio and uncommented
>>>>> pulseaudio.
>>>>>
>>>>> I'm pretty sure that's all I did.  If you get errors about libraries
>>>>> just install what you have missing.
>>>>>
>>>>> I am currently having trouble configuring pulseaudio to autospawn
>>>>> properly on the Raspberry Pi.
>>>>>
>>>>> But I hope if I can get it to work switching from portaudio to
>>>>> pulseaudio will solve all the latency issues and crashes we currently
>>>>> have with Arch on the Pi.
>>>>>
>>>>> We have SpeakUp and Emacspeak running with eSpeak on Arch on the Pi
>>>>> but have had to blacklist firmware and newer versions of sound
>>>>> drivers because of a change to the driver which introduced some
>>>>> latency problems.
>>>>>
>>>>> The other way to make SpeakUp use pulse would be to ditch espeakup
>>>>> and use speechd-up to connect it to speech-dispatcher but I haven't
>>>>> had much success with speechd-up on Arch either on x86 or on the Pi.
>>>>>
>>>>> Mike
>>>>>
>>>>>
>>>>> On 02/11/2013 00:01, Brandon McGinty-Carroll wrote:
>>>>>> Mike,
>>>>>> If you have the configure commands/steps handy, I'd love them, as
>>>>>> would others, I'm sure.
>>>>>> If not, I'll go looking and post back what I find.
>>>>>>
>>>>>> Brandon McGinty-Carroll
>>>>>>
>>>>>> On Fri, Nov 01, 2013 at 10:22:05PM +0000, Mike Ray wrote:
>>>>>>> Hello,
>>>>>>>
>>>>>>> Espeakup doesn't use either portaudio or pulseaudio.  It just calls
>>>>>>> espeak, which in it's default configuration uses portaudio.
>>>>>>>
>>>>>>> It is possible to re-compile espeak to use pulseaudio.  I have just
>>>>>>> done exactly that in efforts to get tts to work properly on Arch
>>>>>>> Linux on a Raspberry Pi.
>>>>>>>
>>>>>>> espeak using portaudio suffers appalling latency on the Pi and
>>>>>>> sometimes crashes the kernel, but using pulseaudio the latency
>>>>>>> issues are gone.
>>>>>>>
>>>>>>> Mike
>>>>>>>
>>>>>>> On 01/11/2013 20:37, Gregory Nowak wrote:
>>>>>>>> If there is a way to use dmix to augment/replace pulse, I'd love to
>>>>>>>> know about it. The problem here is that espeakup uses alsa directly
>>>>>>>> instead of going through pulse. The only two ways I can think of to
>>>>>>>> fix this would be either to be able to run espeakup as a normal
>>>>>>>> user,
>>>>>>>> which should force it to use pulse by virtue of opening alsa as a
>>>>>>>> regular user from what I understand, or to add pulse support to
>>>>>>>> espeakup. Actually, espeak uses portaudio if I remember right, so
>>>>>>>> maybe it's
>>>>>>>> as simple as portaudio supporting pulse.
>>>>>>>>
>>>>>>>> Greg
>>>>>>>>
>>>>>>>>
>>>>>>>> On Sat, Nov 02, 2013 at 04:52:48AM +0900, Devon Stewart wrote:
>>>>>>>>> Isn't there a way to use alsa's dmix plugin to either replace, or
>>>>>>>>> at least augment, pulse? Also, this is assuming that Alsa is
>>>>>>>>> being used from the CLI.
>>>>>>>>>
>>>>>>>>> -Devon
>>>>>>>>>
>>>>>>>>> On 2013/11/02, at 4:41, "John G. Heim" <jheim at math.wisc.edu>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> If I kill pulseaudio, do I still get speech with orca?
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 11/01/13 14:23, Trevor Astrope wrote:
>>>>>>>>>>> Are you using software speech with both speakup and orca? I've
>>>>>>>>>>> had this
>>>>>>>>>>> problem with sound not working on the command line after
>>>>>>>>>>> running orca.
>>>>>>>>>>> In my case, gnome/orca started a pulseaudio process running as
>>>>>>>>>>> my user,
>>>>>>>>>>> even though I have pulseaudio and speechd-up configured to run
>>>>>>>>>>> as a
>>>>>>>>>>> system daemon.
>>>>>>>>>>>
>>>>>>>>>>> Killing the pulseaudio process running as my user gets sound
>>>>>>>>>>> back. You
>>>>>>>>>>> may need to restart speechd-up service as well, as I sometimes
>>>>>>>>>>> have a
>>>>>>>>>>> problem with this too, but I usually use hardware speech, so I
>>>>>>>>>>> don't
>>>>>>>>>>> remember the scenario where that is necessary to do as well.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On Fri, 1 Nov 2013, John G. Heim wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Sorry if this is an FAQ but is there a solution to that
>>>>>>>>>>>> problem with
>>>>>>>>>>>> running both speakup & orca in debian 7? Every time I've
>>>>>>>>>>>> installed
>>>>>>>>>>>> debian 7 (aka wheezy), I can use speakup fine unless I log in
>>>>>>>>>>>> at the
>>>>>>>>>>>> GUI and run orca.  At that point, speakup stops talking and
>>>>>>>>>>>> nothing
>>>>>>>>>>>> I've found short of rebooting gets it working agin. I
>>>>>>>>>>>> understand this
>>>>>>>>>>>> has something to do with pulse audio run in the GUI.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> --
>>>>>>>>>>>> ---
>>>>>>>>>>>> John G. Heim, 608-263-4189, jheim at math.wisc.edu
>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>> Speakup mailing list
>>>>>>>>>>>> Speakup at linux-speakup.org
>>>>>>>>>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Speakup mailing list
>>>>>>>>>>> Speakup at linux-speakup.org
>>>>>>>>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>>>>>>>> --
>>>>>>>>>> ---
>>>>>>>>>> John G. Heim, 608-263-4189, jheim at math.wisc.edu
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Speakup mailing list
>>>>>>>>>> Speakup at linux-speakup.org
>>>>>>>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>>>>>>> _______________________________________________
>>>>>>>>> Speakup mailing list
>>>>>>>>> Speakup at linux-speakup.org
>>>>>>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Michael A. Ray
>>>>>>> Analyst/Programmer
>>>>>>> Witley, Surrey, South-east UK
>>>>>>>
>>>>>>> I KEEP six honest serving-men, They taught me all I know. Their
>>>>>>> names are What and Why and When and How and Where and Who.
>>>>>>> -- Rudyard Kipling (paraphrased)
>>>>>>>
>>>>>>> Interested in accessibility on the Raspberry Pi?
>>>>>>> Visit: http://www.raspberryvi.org/
>>>>>>>
>>>>>>>  From where you can join our mailing list for visually-impaired Pi
>>>>>>> hackers
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Speakup mailing list
>>>>>>> Speakup at linux-speakup.org
>>>>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>>>
>>>>>
>>>>
>>>>
>>> _______________________________________________
>>> Speakup mailing list
>>> Speakup at linux-speakup.org
>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>
>>> _______________________________________________
>>> Speakup mailing list
>>> Speakup at linux-speakup.org
>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>> _______________________________________________
>> Speakup mailing list
>> Speakup at linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>
>>
> _______________________________________________
> Speakup mailing list
> Speakup at linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

More information about the Speakup mailing list