Hello, my 3 questions:
-Why if im listening music i have no sound in quake? --> (ubuntu)
-Why i only can play smooth in my screen native resolution? --> (ubuntu)
-Can i connect NetQuake servers with ezQuake client? how? (dont tell me nqconnect, this command is not in ezquake)

/nqconnect is available in the ezQuake source. You just have to recompile it with the proper #DEFINE, I think... anyone have a binary with this feature enabled?
FTE or DarkPlaces can also ZQuake already though.

Because ezQ uses OSS as Sound-Standard, which doesn't support more than 1 audio-stream per soundcard.
ezQ also supports ALSA, but that is to laggy to use it in a nice way... I think a better ALSA support + PulseAudio would just be nice... (I know PulseAudio is sometimes buggy as hell :S)

Nonono, OSS does more than one sound PERFECTLY. I switched from ALSA to it and it resolved all my sound issues. No lag, no "blocking", only good. Of course it always depends on your soundcard. I bought a soundcard for QW (cheap Ensoniq ES1371, so I did not have to use my onboard crap). However not even that made the ALSA woes go away. I switched to OSS after I bought the card so maybe it would work well with OSS every with the onboard one.

It's probably ALSA's OSS emulation that sucks.

- I can listen to music and still have game-sound in ezquake, but I don't have a normal soundcard, it's a recording artist/producer card with headphone input/output, the Echo AudioFire 2. It works with ASIO drivers. Perhaps ASIO For All driver could make a difference with normal soundcards as well ?? (have no idea really)

- for netquake and SP, I got Qrack. I love it. I've tried all these other clients, but Qrack is just a breeze to setup and have decent fps on an older system (while still having eyecandy with the extra pak file). Look for it on www.quakeone.com. It also has built-in mapbrowser, which is something I'd love to see in ezquake too. Also: the common netquake servers are in available in the built-in server list, like Euroquake.

OSS and ALSA should both support multiple simultaneous channels - this is however ONLY if the sound hardware supports this. Most modern low cost cards (especially on board chips) seem to support only one channel and only one sample rate (48khz).

I can't say much to OSS as I haven't used it for a long time - but ALSA supports this by mixing and resampling multiple applications via virtual configured sound devices.. and all applications that want to access the hardware at the same time should use this interface. Try setting the ALSA interface for all applications to "dmix". That should be a device configured by default that usually takes care of this mixing.

See the ALSA documentation for hours of fun! Other sound managers might work similar (ESD, PulseAudio? or what is this exactly).

Alsa sucks big time, at least for games that require lower latencies.
You have to choose between low latencies, and software mixing, by using different devices. FTE defaults to lower latencies (as the latency is significant with alsa's mixing).
The default depends on your alsa config.
If you don't want to have to choose between latency and mixing, use OSS4. Its just a better API. ezquake is meant to support either, and FTE certainly does. Your milage may vary, and you may need to install the oss versions of various apps instead.

Wrong.
OSS4 supports mixing on single-channel cards, while ALSA can support mixing on single channel cards on the condition that you don't mind the added latency. OSS3-Free does not.
48khz is a limitation of the hardware, not of the API. All versions of ALSA+OSS4 will resample accordingly. OSS3-Free does not resample when in mmap mode.

Using mixing in ALSA induces much higher latencies. OSS4 seems to avoid high latencies when mixing.
OSS4 Apps should generally just use "/dev/dsp" for the default sound card without having to compensate for channels.

OSS3-Free is deprecated, use OSS4.

Yes, I meant that the frequency limitation is in the hardware. The bottom line is that both APIs - ALSA and OSS(4?) do support mixing and resampling. It is just a matter of configuration and what works for you better.

Don't know about the latency difference between those two APIs. I never negatively noticed a delay in the ALSA case. But then again I'm a div6 player and use OSX anyway ;-)

Is OSS4 part of the kernel btw? I thought that the kerne herd dropped OSS(3?) in favour for ALSA a trillion years ago - do they support both now?

ALSA mixes things in user space.
OSS4 mixes things in kernel space.
This means ALSA needs to wait for the given process to become active in order to mix sound.
OSS4 mixes as needed, and doesn't need to buffer as much sound. Thus achieves lower latencies.
I think.
Quake's sound engine works by DMA/mmap, rewriting its sound buffer constantly from the start time to a time specified by _snd_mixahead.
You find the OS mixer chunks of samples at a time, and buffers them. The smaller the chunks it grabs (and mixes+buffers) the lower the latencies that you get. The OS mixer does not re-mix the same samples should a new player appear and start shooting a rocket at you, unlike quake will. I can't fully remember how quake works, but you want the lowest mixahead you can get away with. And that means OSS. Its less likely to consume a huge chunk.

My experience of alsa is that of sound lag or disabled mixing.
Might be just due to ALSA's sound driver... But the AC97 codec hardware is reasonably common.


OSS4 is not part of the official kernel. Not the linux one anyway. However, it is GPLed (or BSDed!) and may be an optional kernel module for your distribution. And is freely downloadable.
Its a bit of a pain to install, due to having to kill ALSA first. Needs runlevel 1 or something.

Actually I found the installation incredibly easy. It did everything by itself iirc.
When I tried helping some qw rookie install it on Ubuntu some while ago, it did not work too well though. But there is a guide in their forums.