Voice Support for the Linux Client

Hi,


I'm lost in this (far too) technical topic...

I'm trying to have voice working on my Debian box, the trouble is that everything seems to work but i can't hear other people talking.

People can hear me perfectly and i "see" when they're speaking (green waves over the head, and they're in active speakers, too)

In the Voice chat tab, Device Settings window, Audio out, i don't have Alsa but "Default" and "null driver" only.(Alsa for input device)

Others sounds (audio streams, UI, wind) works perfectly..

I removed the .asoundrc file and updated all alsa related stuff (Lenny)

My soundcard is a SB pci128 (very common, indeed), there's also an integrated Ali M5455 on the motherboard.

i'm using Second Life First Look Dazzle 1.19.0(80267)
Any hints or links welcome..

What happens if you put on your headset and try this?

In one terminal run alsamixer and in another terminal run

$ arecord -f S16_LE -r 3200 | aplay -vv

This should create an "echo" type test. You should hear everything you say but with a slight delay.
If not, use the alsa mixer controls to get your capture device (mic) working and your render (headset)
device working.

Then try using SLVoice and capture all the LLVoiceClient lines in the debug window. You are looking for
the (something like) setrenderdevice and setcapturedevice messages.

Let us know if this helps.

thank you Ab, it helped me to isolate the problem,

I have voice now...
i have excluded OSS and ESD in the startup script and voice output works just fine.
(audio output was initialized with OSS, that does explain the "null driver" line in Voice chat tab, i guess)

...but i don't have the UI sounds and audio streams anymore. My guess is that SLvoice grabs the audio device for itself and doesn't let the viewer access it..

If i could access the other soundcard on my box for the UI sounds, that would be perfect, but i don't know if it's possible

Thanks for your help anyway

HmmmmKay... So a Linux Client is released with voice support finally (for testing)?
How come i can't find it on secondlife.com? The only unix client i can see is the same one i've been using for a while now, without the proper voice binaries, thus without voice support.

Am i missing something?

Yes, you're missing something. The Linux voice program is included with the release candidate, the windlight first look and the dazzle first look.

Refer to this file for more details:

http://www.paganstead.com/extras/linux_voice_update_2008-02-22.txt

Adamas

I'm a programmer, and do know the company who made the linux module should have already made a 64bit version. not doing so shows a complete lack of competance.

I use 64-bit CentOS and find having to boot to a 32bit partition greatly disturbing. Hopefully the company gets their act together and also makes a 64-bit version. Most are going to be running 64-bit now, so it only seems well... intelligent to write code for 64-bit.

Voice works on my Mac just fine, and seems to just not work on my laptop. My desktop doesn't work at all unless I boot in to the 32-bit CentOS.

"Yes, you're missing something. The Linux voice program is included with the release candidate, the windlight first look and the dazzle first look."

Ok, i must have overlooked that. Thanks for the information

Using Ubuntu 7.10 I got the same result. With no modifications, and voice enabled I could see that people were talking but not hear them. Then, after excluding ESD and OSS I can hear voice chat, but no UI sounds or streaming audio. So, it seems I can have one or the other but not both right now. Also, after excluding ESD and OSS from the startup script my framerates went from their normal 10-12 fps down to 2-3 fps. So, I can listen to voice chat, but I better not plan on doing anything or moving around.

Hi,
I have this situation that i cannot find out a solution:
A Gentoo system that acts as a firewall and a router for all computers in the LAN that are behind it. The Voice function of SL is working OK, as the needed ports are forwarded to its internal LAN IP (let's say the internal IP is 192.168.1.3).
At the same time there is an Ubuntu/Mandriva/etc. systems in the LAN segment (let's say one of its IPs is 192.168.1.5) that has a firewall running on it and if connected directly to the modem that supplies the outside internet connection has SL Voice running OK. But when it is connected in the LAN where the gentoo system is router, there is no Voice.
I suppose the solution has something to do with DNAT, NAT etc, or with running one SLVoice process on the Gentoo machine and all other clients in the LAN to connect to it, but i cannot achieve any result with all this.
So I will be glad if someone could help me to resolve this issue, which in short is "How to use the SL Voice function on many clients that are running in one and the same LAN segment and are behind a router?"

EDIT: I am talking about the latest SL Linux client version - 1.19.0.5 , and i have no Windows or Mac machine at all in that particular LAN.

I'd had to start off with 'man iptables' in a terminal and would still be too clueless on giving you a 'howto' ......

hey guys!

I have the same pb than Michelle & zhorglub voice need OSS to work...
So we have the choice voice chat only or no voice thats sux!

Is somebody has a clue about this issue?

EDIT : i found the solution with the ubuntu group, you have to install esound, restore the client defaut and it works perfectly !!!

HAVE FUN!

Might as well be the first to whine that the release candidate 1.19.1.0 has broken my working voicechat!

I'm running ubuntu gutsy, with pulseaudio providing the ESD compatabiity layer. Up till now voice has been working fine based on an /etc/openalrc file of:

(define devices '(alsa))
(define alsa-out-device "dmix"
(define alsa-in-device "plughw:0,0"

and with /etc/pulse/default.pa being edited to enable the line

load-module module-alsa-sink device=dmix

WIth the new client this setup just repeatedly connects to the vivox server every 20-30 seconds. No actual error messages are evident in the logs. And no more green wavy signs when I talk (obviously).

If anyone has any bright ideas as to what has changed with the linux voice client (apart from it being "improved" or any suggestions as to what I should try changing as a result I would be pleased to hear them!

Frustrating business this voice on linux stuff.....

i just tried the new release candidate on my OpenSUSE 10.3 and, except for one crash about an hour into the session, it worked perfectly. Not only that, they fixed that damned popping noise I was getting whenever speakers with a strong signal talked.

I checked the binaries for the voice and found they changed two of the libraries (by comparing the filesizes of those files with those of previous versions).

Adamas

I am running Second Life on Ubuntu 7.10. It is not the latest version of SL, probably 1 (or 2) before, but it does have voice chat. At least it has the settings for it, I cannot get it to work. I also tried installing pulseaudio to see if that helps.
I cannot hear anyone speaking, nor can I even see in the list of Active Speakers who is speaking, the list is blank.
In the settings for voice chat, the only option for speakers is "default".

Suntan,

I can't get the audio output to dmix either...

I think the 1.19.1 client is much better at picking up all of your soundcards, so probably the defaults work best for the alsa and pulse settings.

I'm still adding:

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

to /etc/openalrc

I guess it depends on what you mean by working perfectly, Adamas. For me, mixing the output into one soundcard , e.g. usb headphones, is essential, although I do quite like listening to other people's interface sounds.

If anyone has any clues on the mixing front that would be appreciated.

Amber

Sorry, I only use one sound card. Everything comes pre-mix.

Adamas

I downloaded the latest version of SL and installed that, now the voice chat is working, sort of, except that so far my microphone is not working with SL. I can hear what I speak into the mic in the PC speakers, but SL is not picking it up.

I just want to ask, has anyone at all gotten voice to work fully along with UI sounds, ie. you can hear in game sounds, speak on voice, and hear on voice all at once.

So far I can't find anyone in SL or on these forums that meat that criteria.

Yes.

I have it all.

You just found someone who meets that criteria.

Adamas

Btw the voice blob of the WL RC 1.19.1.0 is far better in audioquality and much more configurable then the voice blobs before.

4 my lagtop isnt equipped enough for WL i tryed just to copy the voice files of 1.19.1.0 to the 1.19.0.5 viewer - and it works very well. If someone tries this, too, remember to backup the old files before, just in case.

Doubltfully so, will there be voice support for the x86_64 builds of the client?

Or will there be another 32/64 bit problem?

Several clients have removed FMOD due to the problem replacing it with openAL/gstreamer to solve previous 32/64 bit issues leaving voice as the last one.

After lots of tinkering, i got my Voice t work.

I have 2 Soundcards in my Computer, 1 is the Onboard Soundcard with ACL888 Chipset and the 2nd is a SB Audigy SE. I a Using Ubuntu 7.04 (XFCE Windowed) and have Compiled Alsa from Soure, now i at 1.0.16. I got the final Idea as i read in this Thread that SL and SL-Voice are seperated. Now i let SL-Voice work over the Onboard Chipset ACL888 (input and output) and let the SL-Client work over the Audigy as well as the rest of my Programms that use Sound.

heres my .asoundrc

#---- Soundblaster Audigy SE ----#

pcm.ca0106 {
type hw
card 1
}

ctl.ca0106 {
type hw
card 1
}

#---- Onboard Soundkarte ----#

pcm.acl888 {
type hw
card 0
}

ctl.acl888 {
type hw
card 0
}

#--- Filter ---#

pcm.Filter_DMix {
type dmix
ipc_key 1024
slave.pcm "ca0106"
}

pcm.Filter_Duplex_c1 {
type asym
playback.pcm "acl888"
capture.pcm "plughw:0"
}

pcm.!default {
type plug
slave.pcm "Filter_DMix"
}

heres my /etc/openalrc

(define devices '(alsa))
(define alsa-out-device "Filter_Duplex_c1"
(define alsa-in-device "Filter_Duplex_c1"

SL_Voice shows Alsa as input and outut available and SL-Game ist started with esd enabled and alsa/oss disabled

Just posting my experience.

So far I have yet to get voice working. I have two sound cards. One Integrated card and one Audigy SE. I black listed the integrated card because it was broken. Regular sound was working perfectly, but I couldn't hear or speak. So I decided to install PulseAudio in Ubuntu 7.10. Now the regular sound seems to have a bit of a delay, and Voice doesn't even show myself in the active speakers box.

Does anyone have any idea why?

You probably don't have esound running.

Adamas

How do I get esound running?

I thought PulseAudio was supposed to replace it?