During the 2.5 development cycle, ALSA, the Advanced Linux Sound Architecture was merged into the Linux kernel. At that time, support for OSS, the Open Sound System was marked as deprecated. A recent thread on the lkml discussed the future of OSS, questioning if it is worth fixing issues with now deprecated OSS drivers. Read the full discussion at KernelTrap.
The Future Of The Linux Open Sound System
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Seriously, if ALSA is the future, then why dilute developer resources working on two competing implementations?
Alsa still has some issues with older hardware, esp. laptops. I’m glad the OSS support for my soundcard was put back in to 2.6.4. Someone took in out earlier and I had all kinds of annoyances with suspending and resuming.
…some drivers still need some work. For example, the ALSA EMU10K1 driver sound like shit (especially the bass) when you compare it to its OSS counterpart.
ALSA is still far from perfect. OSS works much better on some hardware and I think it should not be deprecated (at least not yet). But I agree that ALSA is more modern and maybe it’s the future…
But I agree that ALSA is more modern and maybe it’s the future…
I’m only saying that it appears as though the big kernel guys have made up their minds that ALSA is the future. I have no doubts that OSS could be ‘fixed’ but if they’re going to be replacing OSS with ALSA anyway, they might as well go all out, and only keep the OSS stuff that currently has no equivelent in ALSA.
Kinda offtopic, but it’s one of those things that bugs me, people making a well thought out, reasoned choice and not fully committing to it. Things like Red Hat’s preference for GNOME. You can’t even get their version of KDE installed alone on their distribution because all the configuration tools and such are GNOME apps (along with their GNOME lib dependancies), so you’ve got duplicated functionality eating away at your HDD regardless of your own personal preference.
For crying out loud! Who cares if some people prefer KDE (I do), they’ve chosen GNOME, and should only provide QT and KDElibs for those of us who want/need KDE programs. Not two entire DEs.
It seems that the Linux guys have chosen ALSA, so why waste their time maintaining two complete sound systems instead of fixing tthe one they prefer?
What they should do is, as one poster mentioned, deprecate it. It should simply be put into maintanance mode, no new features added, simply fixing bugs. During this time, all NEW drivers should only be written for ALSA (just as FreeBSD now has all new drivers written for BusDMA) and OSS drivers should be gradually migrated over with a small team of “bug hunters” working through the bugzilla.
Each month, the bugzilla should stop taking new bug reports, and for a small team to work through, remove duplicates, mark the bugs that need to be fixed and seperate the features out. Fix the bugs, release an update then unfreeze the bugzilla only allowing bug reports from the new version of ALSA to be lodged.
“What they should do is, as one poster mentioned, deprecate it.”
Already happened.
I happen to have a SB Live!#% and have no complaints about the ALSA EMU10K1 driver (the card sucks, but that’s a different story).
What i think they need, before they decide, is input from users. Users submit bugs or distorted sounds. If there are no complaints about a certain ALSA driver, remove the OSS version from 2.7 versions… till no driver remains in it.
My only complaint is that somehow, some program (i think discover) in my bootscripts since 2.6.3 choses to load the OSS EMU10K1 driver. I don’t want that! I remove these modules using a quick hack in the init.d scripts: unloading it after everything else has been done (along with a fix for the mouse) so that both work fine after the computer has been booted up.
For some reason when I use ALSA, my line-in quits working. And I have new hardware… ASUS A7N8X Deluxe (rev 2.0)
So I still have to use OSS because I listen to my FM Radio Tuner at least 2 or 3 hours a day (and it goes through my line-in)
> For example, the ALSA EMU10K1 driver sound like shit
> (especially the bass) when you compare it to its OSS
> counterpart.
This is an ALSA bug if you have surround speakers enabled (which I don’t):
From alsa wiki:
http://alsa.opensrc.org/index.php?page=emu10k1
Somewhere in Note 2:
I had issues with the bass being too low on my SBLive Value with the OSS emu10k1 drivers. The issue, apparently, is that some SBLive cards have their analog front signals inverted, which cancels out the bass on some speaker systems (Cambridge Soundworks four-point systems, for example). The emu-tools [1] fixed that by offering an option to invert the rear speaker signals as well. The emu-tools don’t work with ALSA though, so the only way to get acceptable bass sound is to disable the rear speakers (Wave Surround to 0, as mentioned above). Just something to keep in mind if you have a similar setup.
Nah, I don’t have surround speakers. The bass just sound crappy, especially when you rise it. It’s like the signal becomes heavily distorted. I don’t have this problem with the OSS sponsored Creative not with their drivers under MS Windows. Then again, I have a SB Audigy so it might be a bug or feature with this specific card…
I must admit they have improved their drivers considerably since last year. The sound was simply terrible. Now it’s quite acceptable. Anyway, I keep using the ALSA drivers because “it’s the future”.
Well, if it’s Audigy, I agree with you; it sounds somewhat differently from SB Live with the same snd-emu10k1 driver (especially in the MIDI department), but it’s not really terrible in my opinion.
Midi on Audigy and SB Live is same, there is no diffrence.
Well, the function is the same; but if you pay attention, you will hear the difference in quality – of course the choice of soundfont will also affect that difference. My ears certainly can tell whether it is a SB Live or Audigy playing a MIDI file with the same soundfont, for eg. 8MBGMSFX.SF2 from SB Live CD.
When you not use Bass, Treble on ALSA and OSS quality must be same. I think, than recording from Line In will be better on ALSA.
Zero, I never played a MIDI file on Linux. I made my comparisons with MP3, OGG and WAV files. Like I said before, the driver _is_ getting better but I still think the OSS driver sounds the best. I really don’t understand why as the ALSA driver was developed from the source of the OSS one (or so I’ve heard). They _should_ sound the same unless the developer used an archaic version of the OSS driver.
pzad, I know they should but they don’t here. That said, the ALSA driver is probably the best because of the vastly superior architecture and that’s why I support it… The OSS driver is just not obsolete yet because it sounds better. IMO, it would be a mistake to remove OSS from the kernel without fixing this.
I have to agree here. The bass is copletely distorted since I switched from OSS to ALSA. There is definately something wrong with that driver. The card sounded fine with OSS and “That other Operating System TM” with the same cheap speakerset (no surround).
My second machine with a Fortissimo II on the other hand sounds perfect with ALSA.
Just put OSS in maintenance mode until there is an equally good ALSA alternative. The crysstal Sound drivers in ALSA from my experience are quite good. These could be left unmaintenaid and eventually be removed from OSS.
If I am listening to an mp3 and I use say an instant messenger, and get a message, the sound from the message is apparently put into a que to be played after the mp3 is finished. is this normal ? SUSE 9.0 is what I have tried it on.
I used to have that problem (at least, the symptons are the same) too in early 0.9.x versions of ALSA, with the SB Live!#%. But IIRC it was better with OSS emu in ALSA. Anyway, it passed away in later 0.9.x versions and still is okay in 1.0.x versions. You might try a new versions, ie. one which comes with 2.6.x
i mean, wanting to play more than one sound at once is just selfish
I think that the optimal solution would be to stop adding new features to OSS, and limit work on it to fixing serious bugs only. But keep it in the kernel tree, because – why not? It’s only a few hundred KBs of additional source to download, and all those who compile the kernel- be it distro maintainers, users, whoever – can decide for themselves.
This way most of the programming effort could be shifted to ALSA, while there would be no regression in functionality.
My personal experiences with both systems: I started using OSS in 1999 . I remember it was a real pain to set it up on my laptop, but once I did it, it was working like a charm.
Recently I tried switching to ALSA, and I couldn’t make it work in 1 hour. So I simply gave up. I think there are many more users like me out there, who are simply satisfied with OSS and dont want to spend several hours RTFMing to switch to ALSA.
Hiya. I can’t seem to recreate this with the EMU101k1 driver on a couple of cheap creative cards and the latest alsa drivers…
When you say the bass is distorted, what do you mean? Is it bass light, too bass heavy for tiny speakers, or is there actual clipping distortion going on (which would be audible as related harmonics in the high frequencys)? Is the problem still there if you lower the pcm input level in the mixer?
Perhaps the default mixer settings are very wrong (bad initialisation for bass+treble controls)…
Have you tried…
amixer sset Tone on
and then adjusting the bass tone control?
The emu10k1 cards are never going to pass a signal very cleanly at the best of times, but it would be nice to nail this problem if it’s still in the recent drivers.
Cheezwog: I’m using the alsa drivers that are in kernel 2.6.3 with Andrew Mortons patches. I have two systems that were nearly identical save the soundcard (the other has a CS based card) (same KT133-chipset/Athlon900 Geforce2 MX HD etc.)
The same soundfile plays perfect over the CS card. And the bass on the SBLive causes the speaker to crack as well as not being clear. This is not the same as the SB problem on windows it definately sounds different. It also only happens when the volume is over 3/4. But at 1/2 the max volume the volume is about the same. Bass/trebble are the same on both systems in alsa-mixer. And on lower volumesthe SB doesn’t have notably more bass. It does seem though that the SB becomes moer “bassy” when the volume increases compared to the CS card. So it could be simply to much base at volumes over 3/4.
I have the same distortion problem with SB Live! here. The most ironic thing is that I was thinking it was a Mandrake bug
By saying distorted, well imagine poor and cheap speakers. What happens when you turn up the volume? Song sounds like “bogl, bogl, bogl”
This is the same. When I turn buss up more than the half of a bar it sounds very bad, even with SB 4400 sound speakers. Hmm, gonna try OSS right now.
Hiya.
It sounds like the emu10k1 internal mixer is clipping somewhere. Most of the energy in music is in the bass region, so when a kick drum hits you get a big peak in the waveform, and if a digital mixer has run out of headroom you will get a ‘click’, ‘splat’, or just distortion. The music also starts to sound more bassy as transients can no longer get through. Clipping digital eqs sound even nastier….
After a little googling,,,,
http://lkml.org/lkml/2004/1/20/127
It appears the problem is in the tone controls. The OSS driver uploads some DSP code to the emu10k1, but the ALSA driver uses software eqs, which don’t appear to work as well.
I have the tone controls disabled by default, which is why I probably have not heard the problem. I’ll give it another go, and ask the ALSA folk if they could make that the default if they are the problem.
Could you try ‘amixer sset tone off’ and see if that sorts it? If not, lowering the PCM volume in the mixer until it stops clipping may be a tempory workaround, and should give you the same general output level. Leave the master volume at max, as the clipping is probably happening before it gets to it.
I you have sound distortion try to lower PCM volume in alsamixer and up master volume. Also check that the “Bass” Slider is not too high. An play with the “Wave *” sliders.
I you mute the “tone” slider , it disable bass and treble and put them at 50% which should be the default.
Yep. The problem might be that the bass is way too loud for small speakers… The sound is muffled and there’s a bit of clipping distortion when you crank it beyond 60%. However, I don’t have this problem with the OSS drivers nor with Windows. To be fair, the problem is pretty much unnoticealbe when I don’t play with the Tone setting but I generally raise the bass a bit. I wonder if there’s a way to upload some DSP code on the card? A problem with the software eq would make sense.
FYI, I’ve set all my volumes at 80% (or 81% (25/31), depending on which control).