SL Voice works but...

I upgraded to the last version and I was so happy to see that voice is working now on linux! Thank you very much developers!
My only problem is that after few minutes the voice seems to disconnect from a kind of server. In fact my "Active speakers" list becomes empty and I just can see myself listed.
I have tried to restart the voice connection changing the voice options but nothing changes. The only way to make it work again is to completely restart the SL client.

Is there any way to make the voice connection more stable?
Can I restart manually this connection (maybe through the console) ?

I have read the Voice Readme post and looked for an answer but I couldnt' find anything useful.

thank you very much!!
Lanz
(on Ubuntu 7.10)

Here, this makes voice stable and reliable for me:

http://www.adamascarter.com/extras/sl_voice_linux.txt

It's written for OpenSUSE, but it shouldn't be too hard to translate to just about any distro.

Adamas

Gosh! Isn't sound still a mess in the Linux world? (dons asbestos lab coat!)

I have been getting this behavior when I haven't explicitly told openal (the 3d sound engine) which sound card it should be handling. This is probably because I've needed to use an additional sound card, e.g. usb headset, for voice output as I haven't been able to get the voice output to mix with the second life sounds on the default sound card.

Adding something like the following to /etc/openalrc (as per comments in the file regarding additional sound cards) seems to fix this:

(define alsa-out-device "hw:1,0"
(define alsa-in-device "plughw:1,0"

Having said that I tried a fresh install of openSuse (1.2) w/ Gnome this morning and everything worked perfectly including the mixing of voice and sounds on the default sound card. As far as Adamas's howto goes, it's getting a little dated now since we got a native binary for slvoice... and there shouldn't be any need to install both esd and pulse. That's just nutty!

Anybody else getting success with a fresh distro and which ones?

I haven't had to define anything with regard to openal to make voice work. Also, unless the esound module for pulseaudio has been fixed and actually does what the real esound does, you still need both.

And, for the record, my howto *is* updated to account for the native Linux binaries. I left out the instructions for using the Windows binaries with Wine.

In case you missed it, here is the link (to the updated howto) again:

http://www.adamascarter.com/extras/sl_voice_linux.txt

When you follow the instructions, it just works.

Adamas

thank you very much. i tried your suggested guide but I just made a big mess with my sound options in ubuntu. So I returned to alsa and .aoss which partially works (until voice crashes).

I think I'll wait for an ubuntu SL voice guide.

I am very disappointed to hear in another post that SL voice would not work in Ubuntu 8.04. It will be released in a ten of days and this is not a good news.

If somebody has got a SL voice working stable on ubuntu please say hallo!

EDIT: same results with the RC client. ...

Ubuntu 8.04 is still in beta, so there's a chance its problems will be fixed. One alternative is to switch to OpenSUSE. It just works. Another is to find a way to make it work with what you already have and stick with that without upgrading to 8.04. I gather you're using Gnome (that's default in Ubuntu) so I'd suggest some Gnome users could step up and offer some advice.

If you switched to KDE (Kubuntu) my method would probably work.

Adamas

Same problem right here with debian...
Well I think all of this is no problem with pulseaudio or esd oder something like that... to me it seems to be a problem with something completely different.
The reason is as follows:
When I got voice started the first time, everything was fine, i thought. But after a couple of seconds i couldn't speak anymore. About 1 minute later i got disconnected from the voice server.
What i found out was, that whenever there is a peak on the cpu usage, I'm unable to speak after that. So I can force this by changing the focus from SL to a programm which is a bit complex. Like Firefox for example. The cpu usage does peak usually when switching between tabs within the firefox or something like that. After that, i can't speak anymore within SL... and about 1 minute later, I get disconnected completly.

I think this might have something to do with a deeper system... maybe even the kernel. Something what does not happen when using an openSUSE patched kernel.
So it seems like we can only wait an hope until the vivox guys found out what's going on.

Anyway... the simplest solution is running the Windows-voice-client through wine.
Basically everything you need is working wine-software and the right files from the windows client. That's everything i needed to do.

Using an AMD Athlon 64 System with an SB Live! Soundcard and a VIA-Chipset Mainboard with an Debian System running a vanilla Kernel.
So what are you guys using? Those where it runs and those who have the same problems like me.

I have a SB Live and a amd64 x2 too, running under Gentoo. The Voice binaries from the 1.18.0.5 were the last stable version for me. [Edit: stable = not stopping after a few minutes.]

Copying them over any of the 1.19 installations works, only the low volume for each new speaker is annoying.


Cheers,
Torrid