Installed Linux Mint-Assistance Required

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

Sometimes ago I wrote on this forum that I like to run Linux Mint as my OS and then operate my music server using this platform. I am not very familiar with Linux even though I have been using Windows OS in various forms for many years. If you suggest any changes I would appreciate you can explain clearly.

My HW system:

T41 IBM Thinkpad
1G Memory
1.7GHz Centrino

What I like to achieve IS to stream a "clean" music signal via the USB port where they are picked up by my DAC. This signal can compare to those generated by Windows 7 in terms of sound reproduction.

The Linux Mint is up and running. The Internet and remote network access features work. My initial inquiries are:

a. Which music server to use? Is Music Player Daemon (mpd) music server a good choice? If not
please suggest.

b. Any changes to the Linux OS? Please explain, in some detail please.

These are my initial queries, any suggestions are welcomed.
 
A few questions:

* What kind of USB soundcard do you have?

* Can the USB soundcard play only the music you choose in your player, or are you planning on listening to other sources too (streaming radio stations, youtube in the browser, movies, etc.)? I ask to find out whether your setup will have to mix streams or not.

* Open the terminal, type

cat /proc/interrupts

and paste here the result (just highlight with mouse and paste using the middle button=wheel push).

* Which audio formats does your audio library contain? Sampling frequency/bits, formats.
 
A few questions:

* What kind of USB soundcard do you have?
Does this matter? I will be using an external DAC(stand-alone) which accepts both co-axial or USB switchable.

* Can the USB soundcard play only the music you choose in your player, or are you planning on listening to other sources too (streaming radio stations, youtube in the browser, movies, etc.)? I ask to find out whether your setup will have to mix streams or not.

I am streaming Audio(format see below) only no radio.

* Open the terminal, type

cat /proc/interrupts

and paste here the result (just highlight with mouse and paste using the middle button=wheel push).

* Which audio formats does your audio library contain? Sampling frequency/bits, formats.
ling

Playing MP3, Flac, APE at mainly 44kHz and later higher bit rates later.


Looks like no one wants to input any hints/suggestions so that I can update/modify my Linux, what player should use?
 
?? Did you answer my questions first? Interrupts are pretty important if you want to finetune your system for clean playback on USB sound card. Its type is important too - are you sure your chosen card is properly supported in linux?

Pardon my ignorance, does the sound card needs to be supported by Linux, I assume the data stream coming out from my Thinkpad USB port are common to both Windows and Linux, hence any USB DAC can play from this data stream.
I am using a off the shelf DAC with USB port, if you want to know the chipset used I can tell you.
 
Pardon my ignorance, does the sound card needs to be supported by Linux, I assume the data stream coming out from my Thinkpad USB port are common to both Windows and Linux, hence any USB DAC can play from this data stream.
I am using a off the shelf DAC with USB port, if you want to know the chipset used I can tell you.

If the sound card adheres to USB audio specifications, it will play out of the box on any modern OS without additional drivers. There are USB sound cards with proprietary communication protocol though. Most probably the DAC will be fully supported.

You are saying you will play higher sample rates - does your USB DAC support them? Please post the DAC type.

/proc/interrrupts will tell us which interrupts are shared. You do not want to use the one USB controller which shares interrupts with your graphics card nor ATA/SATA controller - pretty common on laptops.
 
If the sound card adheres to USB audio specifications, it will play out of the box on any modern OS without additional drivers. There are USB sound cards with proprietary communication protocol though. Most probably the DAC will be fully supported.

You are saying you will play higher sample rates - does your USB DAC support them? Please post the DAC type.

My DAC supports up to 96kbps, initially we can set up for 44kbps, I will upgrade later, it should be possible right.

/proc/interrrupts will tell us which interrupts are shared. You do not want to use the one USB controller which shares interrupts with your graphics card nor ATA/SATA controller - pretty common on laptops.
 
I dunno, I never tried JRiver, but Amarok is pretty awesome on linux. :)

Do not worry, all correctly written applications will produce the same sound, when properly setup. I suggest to fine-tune the system for playback first. Then you can choose any application you enjoy using and we can set it up to fit your system best, if you want to.

How about these steps:

1) USB - choosing the one USB controller available in your notebook with least shared interrupts. That minimizes chances of xruns (underruns) and feeds the adaptive USB at a steady fixed pace. Perhaps playing with usb-audio module parameters to keep IRQs at bay.

2) Disabling pulseaudio, configuring plain alsa with maximum buffers and minimum conversion plugin called plug

3) Checking out multiple playback applications, choosing your favourite one

4) Configuring this application or framework (gstreamer, xine, etc.) to use our pre-configured device and increase the buffers to maximum available to minimize the threat of underruns

That's it.
 
Do not worry, all correctly written applications will produce the same sound

You'll never learn it. I'm pretty sure about that. ;)

Apps don't produce sound!!

Apps deliver bits or better they manage the delivery of bits towards the operating systems sound layer.
The way a bit is finally delivered to the sound device in respect to timing variations and the ability of the sound device to cope with these timing variations affects the soundquality, assuming that all bits are delivered bit-perfect. Bit-perfect does not mean time-perfect!
Timing variations ( we're talking pico/nano seconds) can be caused by numerous sources such as power fluctuations and instabilities , timer fluctuations, load fluctuations, RFI/EMI interferences, clock interferences, inefficient codeing, coding errors, inefficient device drivers, poor interface terminations asf asf. A PC is usually full of these.

Depending how an app gets integrated into a system, it can have more or less impact on those timing variations.
Those timing variations are usually of cumulative nature.
As long as you won't be able to slave a PC audio interface to a master clock (e.g. asynchronous USB does not seem to be sufficient), to galvanic isolate and to avoid clock interferences you'll have a problem with above mentioned sources for timing variations.

It's that easy. :D
 
Last edited:
Do not worry, all correctly written applications will produce the same sound, when properly setup. I suggest to fine-tune the system for playback first. Then you can choose any application you enjoy using and we can set it up to fit your system best, if you want to.
Beyond soundcheck's comments on apps affecting timing errors, there are also apps and drivers that do things to audio streams that audiophiles avoid. That's something that ttan should look for after the basic audio playback is up and running. Examples:

Pulseaudio: it actually works ok on ubuntu (off the shelf, easy to control, switch inputs or outputs), but for serious audio, it has a problem: all audio to a given device defaults to a given sample rate and format. Non-matching audio is resampled, by default using a resampler that's optimised more for speed than quality. The highest quality resampler is very CPU hungry. On the occasions I use a laptop for audio, I actually leave pulseaudio running (idle), but direct my output to an alsa plug device that supports multiple sample rates and formats.

Players: I noticed that some players have 'features' / plugins that work on 16-bit format only. Even if the audio is 24-bit, it truncates to 16-bit. IIRC, some s/w volume controls on linux players do that. In this sense, not all players are equal. Sadly even with volume = 100%, the audio is processed in 16 bits, so the solution for was to disable volume control altogether. Sorry I can't recall which linux player this was.
 
Soundcheck, I am not going again into this meaningless discussion about your unsupported claims. When you present a credible way of consistently and repeatedly analyzing the impact of audio application on the bunch of technical terms you have piled without any deep understanding, I will consider your claims. And perhaps that would make you eligible for the nobel prize then. But taking into account your attitude towards learning how things really work, not how you assume they do, I very much doubt you ever will.
 
Beyond soundcheck's comments on apps affecting timing errors

Please be specific, how do properly written applications introduce timing errors in the audio chain? Do you know how the pulseaudio-free alsa chain operates? Have you studied any of these "timing errors", their causes?

there are also apps and drivers that do things to audio streams that audiophiles avoid. That's something that ttan should look for after the basic audio playback is up and running.

Yes, and that is what the source code is for. Plus it is rather simple to check the application for bit-perfection, I can do it for the thread author upon his request.

Pulseaudio...

See step 2) in my post. Pulseaudio should be disabled since runs at a fixed fs and resamples all the others. Or dmix, the same functionality. That is why I asked about the need for mixing streams.

Players: I noticed that some players have 'features' / plugins that work on 16-bit format only. Even if the audio is 24-bit, it truncates to 16-bit. IIRC, some s/w volume controls on linux players do that. In this sense, not all players are equal. Sadly even with volume = 100%, the audio is processed in 16 bits, so the solution for was to disable volume control altogether. Sorry I can't recall which linux player this was.

Please be specific. I wrote "all correctly written applications sound the same". If a player handles input 24bit data in 16 bits internally, even though the output device offers full 24bit resolution, such an application is not "correctly written" IMO. And it is pretty easy to check in the source code - see above.

Audio in linux is a pretty deterministic issue due to availability of source code for the whole chain. No need to make voodoo of it. But I understand the inherent mysteriousness of voodoo attracts a lot of followers, logically solely those who do not know any better. Nothing against you, it is just a state of fact.
 
Last edited:
I have got Amarok to work properly on my Linux Mint, using internal sound card. Play Flac it worked OK.

How do I get it to work by streaming out the data stream on USB port? Any suggestion, using ALSA s/w package, how do configure it to work with Mint? Any idea?

Amarok is using Xine as audio engine, which interacts somehow with Pulseaudio, which finally feeds Alsa.

You better don't worry about Alsa for now,it might get a bit complex... Just enjoy what you hear.


Trust phofman. He has looked up the code. Everything will be alright. :D
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.