Voice Support for the Linux Client

I have been reading this thread for a while. I got voice to work, but it is instable as h*l. If you ar lucky it works, most of the time it wont work.
My guess is that we never get voice in linux.

Please update status of voice client.

Thank you Tofu.

There's nothing particularly new to report - we're re-iterating once again with the voice SDK vendor now that we have feedback from a round of internal testing.

I'm still convinced that releasing what we have right now will cause more frustration than happiness to the average Linux user, but having waited so long now I'm strongly inclined to release the next SDK iteration anyway, assuming no regressions from the current version.

Thanks Tofu,
I think releasing it into the wild might pay off. Linux users are a fairly resilient bunch (usually), and one of our more keen Linux code contributors might be able to knock it in to shape. It would certainly concentrate efforts on that rather than reverse engineering an open source solution.
Just stick a big WARNING sign on it when it gets released.

Fluf Fredriksson - This is somewhat what I've been saying over at the jira. Lets hope that it does come out on the next release. I'm tired of the current WINE solution.

~Ilobmirt Tenk

Since moving to Linux this year I have found that everyone is very friendly and wants to help, so if you were to release this I am sure we could all help with issues etc and get voice working better.

Thanks Tofo,

Is there any indication at what time the release may come?

Hilby

Would you be prepared to give us a rough idea of what the issues are? Don't have to go into too much detail, just so we can be ready

If you think it's valuable - here's an excerpt from the voice README I constructed. All details are subject to change (e.g. I hope there are fewer Linux-specific problems that need mentioning upon release).

KNOWN PROBLEMS
-=-=-=-=-=-=-=

* The 'Input Level' meter in the Voice Chat Device Settings dialog
does not respond to audio input.

* The Input/Output options in the Voice Chat Device Settings dialog
do not list all of the available audio devices, so only the 'default'
device may be selected.

TROUBLESHOOTING
-=-=-=-=-=-=-=-

PROBLEM 1: I don't see a white dot over the head of my avatar or other
Voice-using avatars.
SOLUTION: Ensure that 'Enable voice chat' is enabled in the Voice Chat
preferences window and that you are in a voice-enabled area (you
will see a blue headphone icon in the SL menu-bar).

PROBLEM 2: I have a white dot over my head but I never see (or hear!) anyone
except myself listed in the Active Speakers dialog when I'm sure that other
residents nearby are Voice users.
SOLUTION: This is an incompatibility between the Voice support and your
system's audio (ALSA) driver version/configuration.
a. Back-up and remove your ~/.asoundrc file, re-test.
b. Check for updates to your kernel, kernel modules and ALSA-related
packages using your Linux distribution's package-manager - install these,
reboot and re-test.
c. Update to the latest version of ALSA manually. For a guide, see the
'Update to the Latest Version of ALSA' section of this page:
<https://help.ubuntu.com/community/HdaIntelSoundHowto> or the official
documentation on the ALSA site: <http://www.alsa-project.org/> - reboot
and re-test.

PROBLEM 3: I can hear other people, but they cannot hear me.
SOLUTION:
a. Ensure that you have the 'Talk' button activated while you are trying to
speak.
b. Ensure that your microphone jack is inserted into the correct socket of your
sound card, where appropriate.
c. Use your system mixer-setting program or the 'alsamixer' program to ensure
that microphone input is set as the active input source and is not muted.
d. Verify that audio input works in other applications, i.e. Audacity

PROBLEM 4: Other people just hear bursts of loud noise when I speak.
SOLUTION:
a. Use your system mixer-setting program or the 'alsamixer' program to ensure
that microphone Gain/Boost is not set too high.

Hmmm, those problems are no other than I got with other sound programs... where can I download SLVoice for linux Tofu?
I am willing to give it a try, sounds way better than the wine solution.

This is the most I have heard from you in a while Tofu Linden.
Thanks for hearing us Linux users out.

Those errors are exactly what I have found in the Wine solution. Interesting...

~Ilobmirt Tenk

Definitely sounds like it might be worth throwing it out there, letting a few people play with it. Maybe popup a warning the first time you try to use it, saying it's not recommended for the average user, and that support may be limited or non-existant.

Hmmm... Tofu, is sound for both the regular "enviroment" and the voice component handled by the client (both in/out)? I'm wondering because anything other than the native sound card is put as a seperate card in ALSA, which on Mac and Win32 means I can select the USB headset for voice only and the enviroment sound is played over the regular speakers.

Tofu,

Thanks for the update. All the troubleshooting issues except #2 sound like general voice issues, not specific to Linux.

Thanks again.

Vince

Just tried it with ESD... it worked but no audio input. And this is on my main rig!

/me quickly grabs an elderly man in red and white riding a sled loaded with prezzies , pulls a gun , points it at the old man' head and threatens " .. Gimme voice for X-mas or Santa's gonna get it .. Don't push it ... "

Why not give SLVoice to us in the famous 210 jira call. We'll have a go and wont have do do some stuff with the elderly red-white man.

/me loads the gun - and looks suspiciously at the calendar on the wall and the portly, bearded, red faced gent in the snazzy red and white suit.

/me taps his foot impatiently.

Why all this nannying? Let us play with it!! We're complaining much worse for not having it than we ever would for having it and not getting it to work!
Linux users are tinkerers, tweakers and coders!
We are not Windoze users, we don't need "Shiny" or "Smooth". Linux is raw still, so give us what you have already - it can't be any more raw than the rest of Linux is!
The "no device settings" comes as no surprise... I had expected just that to be the problem a while ago..
Bu please consider: We don't need a (^%"&^£ drop down box. Give us a text entry field, where we can TYPE the name of the ALSA device - unlike Windows users, we _do_ know the system device names!!

Please Tofu, give us that present... /me puppy dog eyes looks at Tofu... please

I'd like to encourage people to try running SLVoice with wine. There are some problems (and work arounds) but my feeling is that these are not going to be any worse than problems with the native client and perhaps some of this information will be helpful in getting this to work a little more satisfactorily.

I am using SLVoice virtually all of the time and have been pleased with the quality of both record and playback.

I can only get recording to work with ESD or PulseAudio. This seems to be because the other methods, ALSA/OSS, do not seem to find the correct microphone hardware source. I get very poor results using ESD for playback and otherwise I find I can only set the output to the default ALSA sink. Now here lies a big problem because this hardware will only be available if the main SL client hasn't already scooped up the device. You will therefore need to switch all of the audio options off in the secondlife script (!) or set the ESD sink to another card or another output device e.g. hw:0,1 on the same card.

This is not as annoying as it sounds because, although it's vaguely forgivable using external speakers for the SL sounds, you should really be using headphones for the SLVoice return otherwise all sounds on the sim will start reverberating like a canyon.

The steps are as follows:

1. Install the windows version of the SL client using wine. The SLVoice binaries aren't changing so I haven't been doing this for every new release.
2. Make a file called SLVoice in the Linux distribution directory. You may as well copy all of the .dll's in the windows installation over to the Linux client directory as many of these are checked for existence by the client before it will talk to SLVoice
3. Start ESD or PulseAudio with the default devices you have decided to use for SLVoice input and the normal SL client output
4. Make sure ALSA and ESound drivers are chosen in the audio tab of winecfg
5. Start wine SLVoice.exe -p tcp -i 127.0.0.1:44124
6. Start the SL client
7. Use the Voice wizard to choose ESD for input/ the ALSA default for output
8. Sing from the rooftops!!

Thanks amber. this worked. I can now hear ppl at least and at this point im happy with that.

I know how to make this thing work, stably and reliably, in KDE.

The procedure used in Gnome won't work in KDE. I've got a few more tests to make before I post a complete, detailed procedure on how to do it, but here are a couple of things to do:

ESD (Enlightenment Sound Daemon):
You need this. Make sure it is installed. Configure it with ESD_BIND value of whatever you're using in the SLVoice command line. For most (if not all) it will be 127.0.0.1. Do NOT include that port number. It will break ESD. In OpenSUSE, the setting can be found in YaST / System / /etc/sysconfig Editor. The file, if you need to do it manually, is /etc/sysconfig/esound.

KDE Control Panel (usually called something like "Personal Settings" or some such):
Open this BEFORE starting either the SLVoice program or the SL client. Browse to the sound system and find the little box labeled "Enable networked sound." It doesn't matter if that box is checked or unchecked. Click it to change it. Click the "Apply" button and exit Control Panel.

Now (assuming you've set up the script properly) start SLVoice. If you want, you can check to see if it is running by running kpm (KDE program manager) and scrolling down to find "SLVoice.exe" and "wineserver." Then, start the client and make sure to enable voice. Just keep that enabled, there's no reason to unclick it unless you don't want to use voice.

The big problem we have is there's a bug in the KArtsServer (artsd) that starts up on boot with networked sound shut off no matter the setting in Control Panel. I'm still tracking down that bug, and if I find it I will post a workaround or fix for it (if possible).

I will post again later today (assuming some disaster doesn't happen) with complete, detailed instructions on how to set up an easy-to-use (hopefully one-click) way to run sl with voice. I will also include a list of all the .dll's necessary to run voice. It turns out the client doesn't look for these, only the Vivox program does.

Stay tuned for more details.

Adamas

Well, it works, but I keep getting disconnected from the voice server, and the sound is uber-choppy. Using Ubuntu Gusty.

Negative for me on the 64-bit side. Will try 32-bits once I'm done coverting 9 seasons of shows to my iPod.