Linux Audio Is Being Further Modernized With The 4.1 Kernel

[phoronix]

Phoronix: Linux Audio Is Being Further Modernized With The 4.1 Kernel

Takashi Iwai sent in his sound driver updates for Linux 4.1, which includes major modernization with the standard bus for ALSA in the sequencer core and HD-audio code...

http://www.phoronix.com/scan.php?pag...und-4.1-Kernel

[birdie]

I wonder if we ever get kernel based audio mixing in Linux or a user space solution which does work.

It's quite possible though that in 10 years most Linux applications will drop direct ALSA support and switch to PulseAudio, but PA is not exactly suitable to low latency audio.

Damn, why is it so difficult to fix at least the audio subsystem in Linux instead of rewriting the code for the Nth time (which usually means more bugs and regressions)?

Android got it right on the first try. Linux/GNU hasn't been able it right for the past 24 years. :-(

[froyo]

Android got it right on the first try. Linux/GNU hasn't been able it right for the past 24 years. :-(

I am not an expert on audio, but today I came across this:

http://superpowered.com/androidaudiopathlatency/

[Azpegath]

I wonder if we ever get kernel based audio mixing in Linux or a user space solution which does work.

It's quite possible though that in 10 years most Linux applications will drop direct ALSA support and switch to PulseAudio, but PA is not exactly suitable to low latency audio.

Damn, why is it so difficult to fix at least the audio subsystem in Linux instead of rewriting the code for the Nth time (which usually means more bugs and regressions)?

Android got it right on the first try. Linux/GNU hasn't been able it right for the past 24 years. :-(

You can still use a real-time, low-latency kernel coplued with JACK underneath PulseAudio and then it would be pretty OK, wouldn't it?

[xeekei]

I do not hate PulseAudio like some people do, but I agree that only code your app for PA rather than ALSA is unwise. Linux audio is such a mess.

[Ancurio]

I wonder if we ever get kernel based audio mixing in Linux or a user space solution which does work.

It's quite possible though that in 10 years most Linux applications will drop direct ALSA support and switch to PulseAudio, but PA is not exactly suitable to low latency audio.

Damn, why is it so difficult to fix at least the audio subsystem in Linux instead of rewriting the code for the Nth time (which usually means more bugs and regressions)?

Android got it right on the first try. Linux/GNU hasn't been able it right for the past 24 years. :-(

You apparently have never written a single line of audio code on Android. Please stop spreading bullshit.

[NSLW]

Takashi does great job since long time, although functionality of my sound driver surprises me with every release, especially support for 5.1 channel

[rabcor]

Linux audio is so messy, I hate pulse, everytime I install it it makes my audio go from bad to worse (not to mention fucking latency issues, who the hell thought it was a good idea to play games with pulse audio? Competitive shooters on high latency audio = bad idea)

I like alsa in terms of quality, but hate it in terms of difficulty to configure and lack of support for some devices, I think it wouldn't be that hard to fix though if it just had a proper GUI front-end, and some automated configuration functions (like they've got for the audio drivers in Windows, you just plug your speakers in and it doesn't matter if it's HDMI, S/PDIF, 2 channel, 5.1, 7.1, or 8.1, as soon as you plug that shit in, it works, and all the channels work, and the most complex thing you may ever have to do manually is setting the default playback device, it's pretty much the same situation for recording (i.e. u plug it in, and it just works)

Meanwhile in Linux land we have to scratch our heads for hours to figure out how ot specify the default playback and recording devices if we don't want what is the default, we have to manually set up dmixing for any digital outputs for whatever stupid reason (thank god dmixing is actually enabled by default now on analogue speakers!) pulseaudio tries, and succeeds in bringing linux a little bit closer to the ease of configuration for sound they have over at Windows and OS X, but it's still not really there, and the cost for this is way too high (not to mention I absolutely detest per application mixing, I don't mind having separate volume sliders for every running application, and I think every good mixer should do it, but there should always be a master channel and all other channels should be slave to that master channel, if the master channel (I guess PCM is the right word) volume is lowered, the volume for all other applications should also be lowered. It's nice to be able to mute one application in your system in a mixer, but not if it's done as damn badly as it is in pulseaudio)

What I want, the only thing I want for sound in Linux is a better alsa, an easy alsa, a good alsa, crappy programs like pulseaudio should be there only for people who specifically need them for whichever specific reasons (like pulseaudio's audio sharing over network feature, that is a pretty useful function, but not to everyone) needless to say, I never use pulse myself, since like I said it only makes any audio problems I have with alsa even worse.

But this is definitely one of these issues that underline how Linux is not ready to be a desktop platform, the audio systems are just way too messy.

[CaptainBlood]

Maybe it will bind pcspeaker to recent laptop chip, cause as of today my new laptop with migrated HD doesn't show it.
Thks 4 ur attention.

[soulsource]

What I'd like to know, as people here tend to call dmix a "dirty hackery":
What is so bad about dmix (compared to pulseaudio)?