CRAS - Chrome Audio Server

[Boost Hardware]

Hi,

I have just found out about the proposed CRAS 

https://sites.google.com/a/chromium.org/dev/chromium-os/chromiumos-design-docs/cras-chromeos-audio-server

Can you explain the reasoning behind this?

http://boosthardware.com/images/pulse-chrome.png

We already have pulseaudio and JACK so it seems that implementing a complete rewrite and effectively bypassing both of these projects is going to cause a lot of pain. This was already attempted with AudioFlinger in Android resulting in an audio stack that is effectively unusable for pro audio and the android audio team is still working on latency issues. Considering this stuff was solved with JACK over 10 years ago and that pulseaudio is a very powerful and robust system too which is now well accepted and understood by the general global community of Linux users having another audio server just for ChromeOS is a big letdown.

http://linux-audio.com/images/linux-audio-stack.png

Currently chrome can be plugged into JACK through pulseaudio. Why do we need another audio server to replace the role of pulseaudio and JACK?

Considering all the effort that has been put into making Pulseaudio and JACK work together it is a big deal for chromeOS to implement a new version of exactly the same system. IMO the effort would be better spent on fixing any integration issues between pulseaudio and JACK to enhance the existing infrastructure. We already went through several years of pain while the Ubuntu team figured out how to properly configure pulseaudio and we can't write professional audio applications for Android due to latency issues so making Linux users suffer again for an unknown amount of time seems like a very shortsighted and unhelpful method of solving any issues that chrome may have with latency and audio routing.

--

Patrick Shirkey

Boost Hardware Ltd

[Trever]

Chrome OS is a browser with a keyboard.  One does not install applications as with Android.  What applications are you talking about?

cras is for chrome os, not chrome generally.  Assumes Linux, etc.

[Patrick Shirkey]

Hi,

If CRAS is implemented it will inevitably end up competing with PulseAudio and JACK causing more confusion for normal users interested in working with audio tools in a Linux environment. After the amount of pain and confusion from the past 5 - 10 years while the Linux Audio Stack matured and stabilised it seems like a very extreme solution to audio latency and routing issues with ChromeOS.

If you are seeking a lowlatency bluetooth audio router then pulse and jack can be made to work well together.   

If pulseaudio is too much weight to maintain taking some of the existing functionality and making it into a jack "app" for chromeOS could be a viable alternative.

If CRAS is built to replace both PulseAudio and JACK users will loose the flexibility to route a chromebook through the network API's that both provide.

In addition the AudioFlinger experience on Android in regards to latency and exclusion of both Pulseaudio and JACK support for professional and consumer Linux Audio solutions should be a warning that it will take a lot of effort to build out CRAS from scratch compared to integrating and fixing the holes that pulseaudio and JACK have for the ChromeOS UX.

Surely it is in everyones best interest to work with Pulseaudio and JACK communities than to replicate the existing system?

--

Patrick Shirkey

Boost Hardware Ltd

--
Chromium OS discuss mailing list: chromium-...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-os-discuss?hl=en

[Mike Frysinger]

CRAS is already implemented & deployed in CrOS and has been for a
while at this point
-mike

[Trever]

I gave up on Linux on the desktop about 7 years ago, so I can't judge the audio situation because don't know it anymore.  And I don't want to unless someone is paying me to.  When I left the Linux desktop world, there wasn't even a driver for my DAC, and the ALSA stuff was a mess, as I recall, and as you seem to point out.  Never mind audio, I couldn't even login to my 401k because of Flash and other ridiculous issues.  Half of the world of computing was shut out to a Linux desktop user.  It was ridiculous.  The time I spent trying to make things work, compiling in hacked drivers for new hardware that wasn't supported upstream yet, etc..  I wasn't learning anything anymore and it was just pure nonsense and a time waste.  I felt like an idiot trying to make it work, frankly.  The whole thing was an insult to common sense.  And lo- along came Mac OS X which just worked and there was a reasonable enough semblance of UNIX there too.

Nowadays the Linux desktop is a better proposition, in part thanks to the support Google gives it, though also Canonical (whom Google has been paying/supporting too).  But I don't think cras will inevitably wind up competing with what's typical for Linux desktops.  Unless it's clearly better, in which case, everyone wins.  But cras seems a relatively Chrome OS specific mini server.  The auto switching for example.

And this is not the place for it but I really don't see Linux on the desktop unless a vendor is behind it in a mainstream way (Canonical, the Dell sputnick project- these aren't going to go mainstream).  Hence, Chrome OS, which "Works For Me".  (Mac OS X going off into the weeds these days... I've switched yet again.)  If you like Linux, it seems to me you should be excited about Chrome OS.  I am.  Look- it's even open source.  I don't see the problem here.

What I particularly don't understand is your basic argument.  Maybe that's just because I'm out of the loop (what I said above).  But it seems to me you make many references to latency.  The cras design doc is heavily concerned with *power useage*.   At least that's how I read it.  You don't address that in your concerns?  It would appear that Google had to implement a new solution (cras) because the mess which is or was Linux desktop audio wasn't going to work for the kinds of devices Google wanted to build (Chromebooks).  To say they should not have done a new solution, given an engineering need for one and the problems with the existing solutions, doesn't seem like a compelling argument.

I have no interest in taking a side here, just saying.

[Trever]

Addendum:

I can buy a laptop and/or desktop machine at Best Buy, that's running Linux and it's open source.  See?  That's huge.  Number one selling at Amazon (which only means so much, but it's an advance).   What's not to like about this?  This is not an ISO "easy install" (cough!) download OS.  It comes on the machine, like Windoze.  It's not a phone or a tablet.  Doesn't this strike you as a huge advance of just about any Linux cause?  "Everyone's interest..."

There was apparently an engineering need for a new audio server (which is open source, BTW) to make this happen.

I guess I am taking sides.  I don't understand why a thing like cras is upsetting in the world of Linux desktop stuff and all of it's history and all the work people did, etc..  I don't get it.  I don't see where the existence of cras is not good, I don't see how a think like cras undermines, whatever.  You invoke "everyone's best interest"... where did that get trampled on by the fact that cras was implemented?

I'll give up here but I'm not getting it.  

[Patrick Shirkey]

So now we have two Google funded Audio layers that completely bypass the work that was done on Pulse and JACK if not making it damn near impossible to integrate?

Nice one guys!

Not only do you not bother to participate in general Linux Audio Development by even notifying anyone of your plans for Google domination of the Audio System but you seem insistent that your way is the best way at the expense of everyone else who is working to make something happen for Linux Audio.

In case you didn't notice Pulseaudio worked just fine for PalmOS.

At this point I can officially say that Google and Linux Audio Development just don't work (tm). I will give up on Google and professional audio now. You guys can stay in your lofty towers getting paid handsomely to bypass the hard graft that has already been done and the rest of us will just get on with fighting in the trenches.

--

Patrick Shirkey

Boost Hardware Ltd