[KimLaroux]

Guys please, don't derail this thread on yet another sound science war. This is a technical thread on how to setup audio on Linux, not a debate on which setup is best.

 

[Snoozles101]

Hey, hopefully this is the right place to ask this... does anyone know of a way to get finer control over a system-wide EQ, such as the PulseAudio Multiband EQ?
 
It *only* has 15 bands by default, but ideally I'd prefer much finer control to tweak specific frequencies to my liking. Is there anything for linux that has this capability on the OS level?

 

[TwinQY]

Jack has jamin, which allows for parametric EQ, 30-band, everything. Most (worthwhile) players will output to jack. I remember back when I used moc, it could do jack as well. I've slowly transitioned most of my music servers to jack and love the customization. 
 
If you want to stick with just ALSA, fil-plugins and a lot of LADSPA plugins running off of ALSA, but I never got it fully working. Not anywhere near what you can do with jack though.15-band? Forget about it....
Although, keep in mind, I've not bothered setting jack up on a daily-usage machine like my work laptop. There's no need for that hassle.

 

[mindbomb]

   I wouldn't mind it to work like WASAPI in shared mode on Windows Vista and 7 (bit perfect until another application needs to play audio). 

 
wasapi shared mode is not bitperfect, it will convert the audio from integer to floating point. you need to use wasapi exclusive mode.

 

[zhunter]

What is wasapi exclusive mode? Are they push and event mode?

 

[gevorg]

What is wasapi exclusive mode? Are they push and event mode?

WASAPI is Microsoft’s own ASIO, it talks straight to the soundcard if set to exclusive mode. For more info see this and this.

Now back to audiophile playback on Linux.

 

[yay101]

Alsa output with no mixer will be bit exact unless you break something to with bit depth or resampling.

 

[emueyes]

I'll read all 12 pages of this eventually, but just to jump to the exciting bit, are any of these players able to be controlled via DLNA ? I realise using the words 'control' and 'DLNA' in the same sentence is hilariously oxymoronic but it's worth a try. Use case: Linux box full of songs needing to feed a receiver via optical, coax or HDMI, and be controlled by a tablet.

 

[TwinQY]

XBMC can do server/client. So can MediaTomb. I'm not too sure about your usage case though - probably can't help you out. More details?

 

[emueyes]

I'm using XBMC on an i3 based machine ATM, and feeding my new Denon receiver audio via HDMI, The XBMC server is very nicely controlled with Yatse, running on an Android tablet. Yatse does beautiful cover art, remote control, has never crashed. All in all, I don't think I can better this. It sounds excellent. The one area is replacement of the i3 machine. It's a few years old, runs Windows 7, and behaves perfectly. But I have a Snapdragon 4 based board, and there are plenty of other ARM boards out there -  CubieBoard, BeagleBone, Odroid, that run on, say, 6W, compared to 60W from the i3 machine. They'll run Linux of course, or Android, which I'm desperately trying to avoid. (Android, that is, not Linux). My sardonic remark about DLNA being unable to control itself was a bit unjust, Yatse / XBMC work great together.. XBMC could be coerced to run on one of these SBCs, running Linux, and that'd suit me just fine, but I'm wondering if there are better alternatives that will fit into the small ARM computer running Linux scenario. 
 
The SBC computer probably won't output sound over HDMI either I realise, so I'm looking at USB to S/PDIF converters in my spare time

 

[TwinQY]

If only Alpine could port to ARM I'd be doing so many things right now instead of being unproductive on a forum...sigh...
 
ReadyMedia (MiniDNLA) is very barebones and might be suited to the task - my previous experience with it left me thinking that it was perhaps a bit too barebones for things I had wanted to do. You'd have to stick with a web client all the way as well.

 

[emueyes]

Thanks TwinQY, I had a look at miniDLNA too, I kinda like it because of that simplicity; the alternative being the labyrinth of XML files in XBMC that I think were written by the same person who wrote Rogue, on a bad day, with a toothache and dentists visit coming up. It's probably got my keystrokes now and is just waiting for me to watch a movie.
 
The OS is the weak point of these boards. To have that beautiful Snapdragon just sitting there because nothing will run on it is sad. I hope you get Alpine, and that someone tells you that you are very friendly and helpful.

 

[stv014]

WASAPI is Microsoft’s own ASIO, it talks straight to the soundcard if set to exclusive mode. For more info see this and this.

Now back to audiophile playback on Linux.

 
WASAPI actually has some relevance on Linux, because it can be used in WINE to bypass the (not bit perfect) DirectSound emulation, for example in foobar2000.

 

[Deac0n]

Here is a site I came across while I was looking for more info, on this thread.

W w w dot ap-linux dot com

Linux distro with just a music shell. Based on mint . Trying it out, shortly on Virtualbox .

I hope this helps

 

[julian67]

It's hocus pocus.

Question: if you already have bit perfect audio then how do the bits get more perfect?

Apparently the "answer" is

The unnecessary services and daemons, included in standard Linux distributions, have been removed and their negative influence on audio playback made impossible. Printing and some other network services running in background of the standard Linux distributions, use the CPU and memory and are completely useless for audio reproduction.

This is kind of idiotic superstition. It's also not really what they are doing because the distribution runs a full graphical desktop with X and the Xfce desktop environment. I'd love to know how a network printer (for example) might have a mysterious (and unquantified) "negative influence" on audio decoding or on the signal passing through a toslink cable and if these are indeed "a bad thing" then how is it that a full X server/client running is conveniently benign? The authors also seem to believe that a real time kernel is need for simple audio playback. It's a lot of bs. Or to put it another way, it's as credible as the measurements they offer to support their claims.