A very common question that comes up when trying Debian GNU/Linux is how the heck do you get Advanced Linux Sound Architecture (a.k.a. ALSA) sound modules set up properly? In this HOWTO LinuxOrbit shows you how to compile and install the ALSA kernel modules, and then setup things using the ALSA Debian script so that modules are automatically loaded and unloaded, and your mixer levels are saved and restored on boot up.
Installing and Configuring ALSA Sound Modules in Debian
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I’m not going to even try this. Until configuring sound in Linux (ALSA or not) is plug-and-play, I am not going to touch it again. I have fine tunes in BeOS and Windoz, thank you.
Yes, I used to do this kind of stuff, but that doesn’t make it any more pleasant. It was more a challenge to prove that it can be done.
Shame, though, with all the OpenSource power behind it, Linux sound support still badly lags behind the other two platforms I mentioned.
I have to agree. I tried configuring ALSA a few months ago on my Red Hat 7.2 back then, and it was an awfull experience. Nothing worked right. The sound chip was (semi-)supported (whatever that means) and *sometimes* I would get sound on Gnome, sometimes not… really weird stuff. Since then, the new ALSA fixed support for that VIA chip, and on Red Hat 8 it supports it out of the box. But if I had to build it and configure it by myself for RH8, it would again be a big hassle.
ps. On BeOS indeed, if your hardware is supported by installed drivers, you have to do nothing. The OS will autodetect the card and load the right driver on the fly. And if you have to install a new driver, most of the time, all you have to do is unzip the driver you just downloaded on /boot and either reboot, or type on a terminal: “rescan name-of-the-driver” and that’s it. Simple stuff.
I have been looking everywhere for a good How-to on Alsa for debian…. I realy can not wait until I can compile a 2.6 kernel though…built in Alsa…how easy will that be 🙂
I am currently running a SuSE 8.1 pro, and I want to put two sound cards into my server box so that I have two outputs (headphones and amplifier) for a party. This works just fine. The card is recognized and configured automatically by the suse hardware detection. But I can not find any kde programs that support multiple sound cards. What I would like is an application that lets me mix several sound streams and pan them between the two soundcards (like a mixer for DJs). Is there a KDE program that supports this????
regards,
tuttle
realy…if you are running stock woody or stock bf for XFS support, it is an apt-get away and 2 line edits.
ALSA on QNX configures itself very nicely
Stay with the kernel OSS drivers.
Yes, Alsa is a apt-get -f install on most Redhat versions. You might have to edit one file add a few lines with a cut and paste from a zillion examples on the web and then at the most run a couple of commands to get the old modules out and modprobe the new module. It is not that awfull but then again it is not brain-dead easy and ALSA should just come with ALL distros in my opinion. But turned on by default?
Hell no.
I get better sound out of my maestro 3 card with the plain OSS support the ALSA stuff always sounded tinny as hell to me and that was on SuSE where ALSA is the default.
sndconfig for Redhat and Yast2 for SuSE and every other commercial has its own sound configuration tool and most sound detection happens at install. Sound support has not been a real issue for me if I accept what the install gives me. Its when I go mucking with it by hand that things get messed up.
It makes no sense to complain about ALSA being hard to install. ALSA is still development code, so it’s not supposed to be easy to install. It’s been merged into kernel 2.5, however, so when 2.6 comes out, using ALSA will be as simple as using the OSS sound system is today (modprobe <snd-driver> for most Linux distros, auto-detect for distros like Mandrake). Now, about the whole “Linux lagging on the sound” issue. Linux’s sound system is quite good, thank you. ALSA is a very nice audio layer, and certain drivers (like the Creative Labs ones) are quite full featured and powerful. And, thanks to preempt, audio latency is great as well. Being hard to use just means it lags behind in the ease of use department, not necessarily in other areas. Ease of use is just one factor in the quality of a piece of software, not the only factor.
get use to it. Linus likes what ALSA has and he wants ALSA to be the sound layer from here on. it makes it easier for hardware companies to write drivers, and the layer is a hell of a lot cleaner. OSS only has good sound support right now becasue OSS is much older than ALSA…that does not mean it is better than ALSA.
It saves time. I don’t think many people use a debian stock kernel (why use Linux if you’re going to use a gimpy stock kernel, especially one that debian provides :-p ?) so the best option is to just compile them into your kernel. I wasted massive amounts of time trying to find other ways to do it. Not worth the effort, Oh and I don’t think the stock 2.4.19 kernel that debian provides has a nice alsa-modules .deb for it yet. Roll your own! 🙂
deb-man:
You don’t have to wait for 2.6 if your scared of 2.5. Alsa has been backported to 2.4 in a patch which is present in the WOLK (Working Overloaded Linux Kernel) http://sourceforge.net/projects/wolk . Download it and apply it to a stock 2.4.18 kernel. Fun fun 🙂
it failed to compile
I did not want to waist my time on trying to get it to work since I messed up 3 times on the config (Xconfig and menuconfig were broken) and when I finaly got everything I though I needed, it did not compile…soo…I did not want to even spend another second on it….as far as the patch, I use the bf kernel so I can get the .deb for it 🙂
it failed to compile
I did not want to waist my time on trying to get it to work since I messed up 3 times on the config (Xconfig and menuconfig were broken) and when I finaly got everything I though I needed, it did not ompile…soo…I did not want to even spend another second on it…
Compiling 2.5.44 as I type this. Basically the bleeding edge kernels needs massaging until they compile. I just turned off (ISA) PnP support and parallel printer support, both non-essential things. Also needed to add an extra include to ipv4.c, but things are compiling fine now. So if you are impatient or you don’t consider kernel compiles R&R 2.5.x is not yet for you 🙂
As for ALSA being difficult to setup. As Rayinder states, ALSA is still development code. It is not ready for primetime, meaning an average Linux user will not be able to install it, not without reading some docs at least. ALSA brings excellent pro-style audio support to Linux. I personally prefer coding against the JACK (http://jackit.sf.net) audio API instead of talking directly to ALSA. All I can say is that Linux audio is already much better than BeOS audio ever was. For the BeOS coders, JACK is a callback based audio API, much like the BeOS Media API. It supports very low latency operations (low ms on consumer grade hardware). Whoops, 2.5.44 is done compiling (4min.20sec!) … reboot!
-fooks
>> I use the bf kernel so I can get the .deb for it 🙂
I’ll have to check that out.
It makes no sense to complain about ALSA being hard to install. ALSA is still development code, so it’s not supposed to be easy to install. It’s been merged into kernel 2.5, however, so when 2.6 comes out, using ALSA will be as simple as using the OSS sound system is today (modprobe <snd-driver> for most Linux distros, auto-detect for distros like Mandrake). Now, about the whole “Linux lagging on the sound” issue. Linux’s sound system is quite good, thank you. ALSA is a very nice audio layer, and certain drivers (like the Creative Labs ones) are quite full featured and powerful. And, thanks to preempt, audio latency is great as well. Being hard to use just means it lags behind in the ease of use department, not necessarily in other areas. Ease of use is just one factor in the quality of a piece of software, not the only factor.
Let me start from the beginning, as it’s the proper thing to do:
– In fact, I wasnt complaining about ALSA, I was complaining about the sad state of sound support in Linux in general. With all the opensource pumping behind Linux, one would expet it’s able to read your musical thoughts and play them in polyphony with 0 latency.
– modprobe <driver> is not the end of the story. You also have to edit the modules.conf file, or else you’re bound to modprobe every time you reboot. And no, modprobe and sh*t is not an acceptable thing, it’s primitive. If you tell me that come Linux 2.6 I’ll have the same “luxury” as what I have now with Linux, well, that’s really stupid.
– The Creative ALSA drivers are a joke. Can I load soundfonts into the card’s memory? Can I use MIDI? Ooohh, so it’s not all -that- powerful after all. But surely by Linux 5.4 they’ll get it right. Right…
– Being hard to use is the single worst feature. You’re winning no musician’s heart if you have to dick on the command line. And the fact that the drivers stink is ot making it any sweeter.
You can preach as much as you want and can, you can shout on the top of your lungs, it all won’t imrove the sad state of the Linux sound, ALSA or OSS.
Just have a look at how BeOS does it. That little OS has yet so much to teach to these Linux programmers. But will they ever learn? After all these years??
could it be they had a script that wrote the lines needed into config files? I think so….what is keeping Linux from that? nothing…all of your complaints are crap….the drivers are not full featured because the companies are not helping the driver writers….and I hope you know that when a new driver for ALSA comes out, you will not need to have new ALSA code in the kernel…infact all you will need is the snd-xxx module…that is the same damn thing that Windows uses (loadable drivers)
the infrastructure is all complete, now it is trivial to put it all together to make a nice coherent system for the sound system to work well.
could it be they had a script that wrote the lines needed into config files? I think so.
This shows that you shold leave the thinking to the people who can: do you have any idea how a PnP operating system works? Do you really think BeOS needs some stupid entry in a config file in order to load the correct sound (or video or LAN) driver? Sorry mate, that’s how primitive operating systems do it.
….what is keeping Linux from that? nothing…all of your complaints are crap
A self-accusation and a slander all in one breath, ha! So, you are basically saying Linux developers are imbeciles that were not able to come up with a really simple impementation yet. And you think one shouldn’t complain about this?
5 years ago I was still very enthusiastic about Linux, but slowly this enthusiasm started to decrease as I noticed that most fellow Linux users were completely blind to the weaknesses of Linux. Many promises were never fulfilled, further discouraged me. Some zealots would quite plainly lie about Linux, to make it appear better: there was this Linux dude from my home country who told me how RedHat 5.2 has a “graphical installer”. Yeah right.
Today I just want Linux to own up to all the hype. I treat Linux just as a Unix-like OS with it’s own peculiarities, which can be useful in my everyday worklife (I’m a seniour Unix designer) for some tasks. Not a desktop or multimedia OS, not something that could or should go mainstream.
modprobe <driver> is not the end of the story. You also have to edit the modules.conf file, or else you’re bound to modprobe every time you reboot. And no, modprobe and sh*t is not an acceptable thing, it’s primitive.
>>>>>>>.
If you’re not using ALSA, then you don’t need to do *anything*. Your distro (I’m assuming you’re using a newbie friendly distro like Mandrake, or why else would you be complaining?) will autodetect your sound card for you. And before you give the example of how it didn’t work for your third uncle’s $5 radio shack brand souncard, let me point out that it’s worked for me on three different pieces of name brand sound cards, as well has for a lot of other people. OTOH, if you are using ALSA, then you deserve whatever you get!
If you tell me that come Linux 2.6 I’ll have the same “luxury” as what I have now with Linux, well, that’s really stupid.
>>>>>>
I presume you mean to say “Windows” or “OS X” instead of the second Linux. Linux has only recently been targeted at the whole media area. Just because it isn’t quite there now doesn’t mean it won’t get there real soon. Even Be didn’t get the media kit right until 4.x, so what’s the point trashing Linux for it’s media capabilities, which haven’t been in development for long at all?
The Creative ALSA drivers are a joke. Can I load soundfonts into the card’s memory?
>>>>>
Yes, man sfxload.
Can I use MIDI?
>>>>>>
Yes.
Maybe you were thinking of Creative’s emu10k1 drivers instead? Unless you need the hardware MIDI support (you might not, computers are fast enough to do real good software MIDI these days) you should probably use that instead. And since you’re such an audio know-it-all, go help the sourceforge project;
http://sourceforge.net/projects/emu10k1
Being hard to use is the single worst feature.
>>>>>
Maybe for you. But a lot of people (particularly the Linux users this article is aimed at!) care more about, oh, actual features!
As for professional musicians, where do they come into the picture? Did the article say anything about using Linux in pro-level audio? Did I say anything about using Linux in pro-level audio? I just said that ALSA was a technologically a very nice audio layer.
Additional ranting deleted, because it contained not a single example or piece of supporting evidence…
this shows how much you do not know…
whether it is a binary file or a simple text file, the config
gets placed in a file that is read by the system.
config files are not primitive, they are just older.
I like config files as they are more flexable and any
application has the ability read from it with out some
stupid API call.
you don’t need any file, be it textual or binary. A PnP OS will build a device tree in memory and load the necessary drivers, during bootup. In fact, Linux does this (somewhat), too, but alas, not very well. Here Mr. Hashem has a point, too bad the implementation still isn’t up to scratch.
Mr. Hashem: And before you give the example of how it didn’t work for your third uncle’s $5 radio shack brand souncard It’s a SB PCI 128. You seem to use a strange rhetoric attack: “you are a wortless being ecause you probably use cheap soundcards, and therefore what you say is not important”. Besides, I have no intention on using software such as Timidity; it’s a real sucker. Besides, I wasn’t just thinking of “hardware” General MIDI synthesis, that would go without saying. I was talking about the MIDI interface on the soundcard.
Dear Linux (ab)users, to your concerted attacks I can only say: I was a Linux supporter once. Now I just wait and watch. If I’m not impressed I’ll say it.
Linux can do that…it is called modules….at boot you have a program like discover list the hardware, then the system modprobes for the modules that match that hardware…oooooo hard. it is still a config file..it just resides in memory and is generated every time you boot the machine.
– The Creative ALSA drivers are a joke. Can I load soundfonts into the card’s memory? Can I use MIDI? Ooohh, so it’s not all -that- powerful after all. But surely by Linux 5.4 they’ll get it right. Right…
This just goes to show that you don’t know what you’re talking about.
Creative has never put out ALSA drivers. They have released OSS drivers.
The ALSA drivers for the SB Live do support MIDI and soundfonts.
Learn a thing or two before you post next time.
Adam
That’s not the only soundcard in existence, I think.
I am with you on this Adam. The PNP OS features can be better. I think it is getting there though.
I remember when I changed out a network card on my app server and the OS recognized it on bootup and prompted me to configure it. In SuSE on of my programmers unplugged his mouse and then turned around and plugged in a USB Logitech trackball mouse and SuSE prompted him that new hardware was found and he was immediately back in business. But I know it can get better. I have a bigger beef.
All distros have sound configuration tools. However, whether or not it is “officially” in the kernel all distros should comes with ALSA. It is the future right? Still, I think the user should get a quick prompt alerting them that OSS kernel drivers and ALSA drivers are both present for their hardware and have the choice on which to choose or whether if they are a newbie to let the system choose. I say this because there are cases where the OSS drivers give you better sound.
But that is just my opinion. To Mr. Hashem, there are precious few soundcards that are not supported by linux but they do exist. If your distro detects your soundcard, sets it up and it works, then most users will never, ever have to go through these instructions. I am surprised no one mentions this except me. Yes, it is complex because it is not the official standard, yet and as other people have said it is coming.
Here is a idea then: Don’t run linux. I am not asking you to. Nobody here is. We don’t need you, we don’t want you, we don’t care what the hell you run. Your not some special butterfly. Your nobody. Run windows. Go away.
i have a cambridge soundworks with 4 points + a bass, so for sure i want to get my rear speaker to work, anyone know how to get it work with alsa driver? oh yes, my sound card is sblive! value..
thx before :>
brussel:~# apt-cache search alsa-modules
alsa-modules-2.4.18-386 – Advanced Linux Sound Architecture (drivers)
alsa-modules-2.4.18-586tsc – Advanced Linux Sound Architecture (drivers)
alsa-modules-2.4.18-686 – Advanced Linux Sound Architecture (drivers)
alsa-modules-2.4.18-686-smp – Advanced Linux Sound Architecture (drivers)
alsa-modules-2.4.18-k6 – Advanced Linux Sound Architecture (drivers)
alsa-modules-2.4.18-k7 – Advanced Linux Sound Architecture (drivers)
apt-get the one you need.
C.
the HP notebooks that take advantage of the VIA sound chip will
not work with sound until you do one of 3 things.
1) hack the source code under drivers/acpi/tables/tbconvert.c
then go to about line 213 or so. you will see:
“if (acpi_gbl_FADT->header.revision >= FADT2_REVISION_ID) {”
make this like evaluate to false always by adding “0 &&” to
the begining of the condition, before the “acpi_gbl_FADT”
2) add the ACPI patch to the kernel and compile
3)run kernel 2.5.x
for some odd reason sound works after you do one of these things
have fun.