Jump to content

New Linux Based Server Software - Snakeoil!


Recommended Posts

@@gypr1961 go to the hardware tab and on the page choose your card. then i think you have click on the Add Alsa Device to be able to use it.

 

then on LMS look in the top RH corner and make sure it is there.

Thanks, not at home at the moment to check but the dac is seen in the hardware tab but there was nothing to choose in the top right, will have a look again later.

Link to comment
Share on other sites

On hardware tab, make sure DAC is shown under "Configuration" and not only under "Information" ...and that Card and Device numbers are correct.

 

Can't believe I dare to think I might be capable of helping anyone with this stuff.  :unsure:

  • Like 1
Link to comment
Share on other sites

On hardware tab, make sure DAC is shown under "Configuration" and not only under "Information" ...and that Card and Device numbers are correct.

Can't believe I dare to think I might be capable of helping anyone with this stuff. :unsure:

You did well there!

I was informed not long ago, there is an update to Snakeoil. This update provides the latest MPD player to 19.13 and git (a beta version has yet released publicly, version 20.xx). Damn it - someone is hard at work here! [emoji6]

Sent from my iPhone using Tapatalk

Link to comment
Share on other sites

I cant get it to load on my main pc (fails on kenel load) so Ive loaded it to a nuc.  I can say its much better than any other linux disto Ive tried, theyve all had great soundstaging but have been very tiring to listen to, tis is much more relaxed.  I know its not a great test to have it running on 2 different computers but for me didnt match up to server 2012 and AO on my main pc. The main thing that stood out was the bass, it was very lightweight compared to server2012/AO. 

 

I'll keep a lookout for the next release and hopefully I can get it to load on my main pc

Link to comment
Share on other sites

Last time I tried Windows server / AO/ JRiver I thought there was a mid bass emphasis which sounded quite pleasant. I would not call the sound accurate though but when it comes to software, it's a case of personal tastes and how it reacts with the hardware. One thing we have noticed , for example, is that a motherboard like Intel 2500cc produces a lot more bass from regular materials than other boards using the same software. It's all a question of getting your favourite balance right

Link to comment
Share on other sites
Guest rmpfyf

@@Tasso - what does that mean? Digital hardware shouldn't be contributing any analogue quality with frequency content.... it's just a question of accuracy. Which is more accurate?

Link to comment
Share on other sites

I gave up on accurate a while ago, I'm just aiming to get to something not attainable :)  Comparing AO to snakeoil and windows 10  for me its a darker richer sound, more relaxing so I listen longer

Link to comment
Share on other sites

@@Tasso - what does that mean? Digital hardware shouldn't be contributing any analogue quality with frequency content.... it's just a question of accuracy. Which is more accurate?

Actually there are no digits being transmitted, only voltage pulses which are interpreted as digits by the DAC. Extraneous noise in that signal does not help but I would agree that much of the hardware /chipset differences relate to timing.

Personally I think obvious bass enhancement is not heading toward greater accuracy but I have noticed at GTG and in discussions on the topic, most people tend to prefer the sound that has a smoother timbre and strong bass, and not necessarily the most accurate. I don't think it's an accident that commercial software packages like JRiver are strong in these two areas.

Sent from my iPad using Tapatalk

Edited by Tasso
Link to comment
Share on other sites

Guest rmpfyf

@@Tasso Linux is a significant operating system (even with a custom kernel and a lot stripped out) and what it takes to get audio to a DAC via USB typically offers very little by way of buffering. More fundamental than timing differences is the propensity to have any data go missing. Then, yes, there's a keeping a bunch of analogue pulses happy that are interpreted as digital bits at the receiver end. But a PC and transport doesn't have a 'sound', it just offers a complete bitstream at the DAC or not, with varying degrees of the latter.

 

Given the way DACs typically work, one would think results (i.e. digital accuracy) might be audible as a slight increase in attenuation and greater detail. Maybe. That's been synonymous with my journey, and mine probably isn't everyone's. 

 

I'm still a little concerned at Chanh's numbers for system latency under load for music - not the last word in SQ but those numbers were extreme. How do you fare?

 

This weekend, time permitting, I'm keen to try Snakeoil. 

Link to comment
Share on other sites

@@Tasso Linux is a significant operating system (even with a custom kernel and a lot stripped out) and what it takes to get audio to a DAC via USB typically offers very little by way of buffering. More fundamental than timing differences is the propensity to have any data go missing. Then, yes, there's a keeping a bunch of analogue pulses happy that are interpreted as digital bits at the receiver end. But a PC and transport doesn't have a 'sound', it just offers a complete bitstream at the DAC or not, with varying degrees of the latter.

 

Given the way DACs typically work, one would think results (i.e. digital accuracy) might be audible as a slight increase in attenuation and greater detail. Maybe. That's been synonymous with my journey, and mine probably isn't everyone's. 

 

I'm still a little concerned at Chanh's numbers for system latency under load for music - not the last word in SQ but those numbers were extreme. How do you fare?

 

This weekend, time permitting, I'm keen to try Snakeoil. 

 

 

I dont like to place limitations on the causes for differences in SQ, it has tended to restrict progress based upon past experiences. On the subject of timing, I can  say that a PSU designed specifically to cope with the rapidly fluctuating current demands of computer based servers ( such as the JS-2) can result in significantly better timing  and accuracy than PSU's built to general DIY or analog audio standards.  this would probably be the biggest single contributor to narrowing the gap between Accuphase  state of the art HS-Link and USB CA. Differences still remain, but  I don't feel that I'm slumming it as much as I used to with CA.  I would suggest  that a DIY i2s solution may not be better than " USB done right"   at this point in time.  Until this PSU is actually tried, people will never know how good the timing for USB based CA can be. The mods we are doing to it do not improve timing,  but they do  improve the overall  timbre and tonality.  the big Hashimoto transformer also produces harder hitting dynamics and more extended bass. 

 

 Another piece of hardware that makes a noticeable difference to most USB setups is this particular USB receptacle .  It was mentioned by another SNA member a while ago and I finally got around to testing it the other day. A definite and worthwhile improvement when fitted to my SOTM card.

http://au.element14.com/wurth-elektronik/8492121/conn-usb-2-0-a-w-esd-emi-filter/dp/2442757?CMP=KNC-GOO-SHOPPING-2442757&CAGPSPN=pla&gross_price=true&gclid=CKXOtrDZwssCFcSOaAoduqECzw&mckv=sDRv2KEEE%7Cpcrid%7C59148083448&CAWELAID=120185580000404932

 

This numbers that Chanh posted were not his  machine.  To give you an idea of what the atom server looks like, this is a reading i just took:

# /dev/cpu_dma_latency set to 0us
T: 0 ( 2400) P:80 I:10000 C:  10000 Min:      7 Act:   10 Avg:   10 Max:      27
T: 1 ( 2401) P:80 I:10500 C:   9526 Min:      7 Act:   11 Avg:    9 Max:      20
Edited by Tasso
Link to comment
Share on other sites

I have now listened very briefly to Squeezelite on Snake Oil O/S and it was very pleasing. Even got a couple of goosebumps, always a sign things are on the right track. And that was using the motherboard SPDIF output !!

Initially it did not pick up my USB connected Accuphase DP-550. Pretty sure that is a punishable offence over here :-)

Anyway, a few emails back and forth with Agent Kith and it was sorted yesterday morning. So hopefully a proper listening session soon.

I also want to acknowledge the easy setup, web pages, documentation, etc that Agent Kith has produced. All for our benefit. Well done mate.

I don't think this has been highlighted enough. It is not just an optimized operating system, it is a flexible environment to try Linux based music players. Pick one, click save and a couple of seconds later your new player software is loaded. No more creating boot disks, re-connecting share drives, rebooting over and over.

Again, well done Agent Kith. I was inspired enough to pimp my server :-)

post-122592-0-97405300-1458048942_thumb.

  • Like 4
Link to comment
Share on other sites

Guest rmpfyf

 

I dont like to place limitations on the causes for differences in SQ, it has tended to restrict progress based upon past experiences. On the subject of timing, I can  say that a PSU designed specifically to cope with the rapidly fluctuating current demands of computer based servers ( such as the JS-2) can result in significantly better timing  and accuracy than PSU's built to general DIY or analog audio standards.  this would probably be the biggest single contributor to narrowing the gap between Accuphase  state of the art HS-Link and USB CA. Differences still remain, but  I don't feel that I'm slumming it as much as I used to with CA.  I would suggest  that a DIY i2s solution may not be better than " USB done right"   at this point in time.  Until this PSU is actually tried, people will never know how good the timing for USB based CA can be. The mods we are doing to it do not improve timing,  but they do  improve the overall  timbre and tonality.  the big Hashimoto transformer also produces harder hitting dynamics and more extended bass. 

 

 Another piece of hardware that makes a noticeable difference to most USB setups is this particular USB receptacle .  It was mentioned by another SNA member a while ago and I finally got around to testing it the other day. A definite and worthwhile improvement when fitted to my SOTM card.

http://au.element14.com/wurth-elektronik/8492121/conn-usb-2-0-a-w-esd-emi-filter/dp/2442757?CMP=KNC-GOO-SHOPPING-2442757&CAGPSPN=pla&gross_price=true&gclid=CKXOtrDZwssCFcSOaAoduqECzw&mckv=sDRv2KEEE%7Cpcrid%7C59148083448&CAWELAID=120185580000404932

 

This numbers that Chanh posted were not his  machine.  To give you an idea of what the atom server looks like, this is a reading i just took:

# /dev/cpu_dma_latency set to 0us
T: 0 ( 2400) P:80 I:10000 C:  10000 Min:      7 Act:   10 Avg:   10 Max:      27
T: 1 ( 2401) P:80 I:10500 C:   9526 Min:      7 Act:   11 Avg:    9 Max:      20

 

See, this is where it gets interesting for me - those stats are high, and things are audibly different for me when they're lower.

 

This said - system latency is not the same as audio latency, it's completely possible that Snakeoil's is excellent. I'm looking forwards to trying this out. It's author has been at this far longer than I have.

 

As for the big iron trans on a PC - I'm all for the choke but oversized iron is overkill for audio. I don't doubt it sounds better and I don't doubt it can sound even better than what you've got, but it's a lot of work to keep USB as happy as can be (the rest of a PC can be made to run at very low load during playback with very little that's transient) - I'd put money on the I2S setup, with some decent money on the power supply for your oscillator (particularly if an OCXO). 

Link to comment
Share on other sites

I'd put money on the I2S setup, with some decent money on the power supply for your oscillator (particularly if an OCXO).

If you are ever get into this route, Salas Shunts worth check out for powering the clock or those critical path in digital?! You must have a decent raw feeding into it, a choke input of 10-50mH @2A is sufficient. I have exhausted lots of funding in quest for better psus, those were relatively to better sq. A journey is of fun doing but financially burden. If this can be shared for a head start without killing time/wallet. :)

Sent from my iPhone using Tapatalk

Link to comment
Share on other sites

- I'd put money on the I2S setup, with some decent money on the power supply for your oscillator (particularly if an OCXO).

You should audition the PSU before placing a bet. I had no idea whatsoever it could make such a difference to timing, but John Swenson obviously did.

Sent from my iPad using Tapatalk

Link to comment
Share on other sites

I have now listened very briefly to Squeezelite on Snake Oil O/S and it was very pleasing. Even got a couple of goosebumps, always a sign things are on the right track. And that was using the motherboard SPDIF output !!

Initially it did not pick up my USB connected Accuphase DP-550. Pretty sure that is a punishable offence over here :-)

Anyway, a few emails back and forth with Agent Kith and it was sorted yesterday morning. So hopefully a proper listening session soon.

I also want to acknowledge the easy setup, web pages, documentation, etc that Agent Kith has produced. All for our benefit. Well done mate.

I don't think this has been highlighted enough. It is not just an optimized operating system, it is a flexible environment to try Linux based music players. Pick one, click save and a couple of seconds later your new player software is loaded. No more creating boot disks, re-connecting share drives, rebooting over and over.

Again, well done Agent Kith. I was inspired enough to pimp my server :-)

Well done Don, looks awesome mate!

Cant believe you have cut the ties with ITunes finally :thumb:

Link to comment
Share on other sites
Guest rmpfyf

You should audition the PSU before placing a bet. I had no idea whatsoever it could make such a difference to timing, but John Swenson obviously did.

Sent from my iPad using Tapatalk

 

Auditioned. It's a good effect. But what's it actually changing? the on-board oscillator and soft timers in your OS don't change... so it's a power issue somewhere else... suggest USB...

Link to comment
Share on other sites

Well done Don, looks awesome mate!

Cant believe you have cut the ties with ITunes finally :thumb:

Well, not totally. I still need that for AirPlay around the house. I will probably pull all my (CD) music from a shared folder on the Mac Mini and keep higher resolution stuff on the SOS.

Also, JRiver seemed better behaved on Linux than on OS-X. But that also was a very brief test.

Link to comment
Share on other sites

Auditioned. It's a good effect. But what's it actually changing? the on-board oscillator and soft timers in your OS don't change... so it's a power issue somewhere else... suggest USB...

 

Of course the timing has changed!!   Just listen and you will be in no doubt.   There is no way that anyone can satisfactorily demonstrate the sonic effect without listening.  Fiddling with software and settings provides very small differences  compared to the gains that can be  had with the best PSU.  

Link to comment
Share on other sites
Guest rmpfyf

Yes but @@Tasso, where is the timing changing? Clue: once your computational loads are optimized (as flat as possible) the majority of the effect isn't in getting bits to the USB port - those are well-timed even with a switch-mode PSU.

 

Mission-critical real-time machines aren't running Swenson-designed linear PSUs. (They're not running USB either - think about it).

Link to comment
Share on other sites

Yes but @@Tasso, where is the timing changing? Clue: once your computational loads are optimized (as flat as possible) the majority of the effect isn't in getting bits to the USB port - those are well-timed even with a switch-mode PSU.

 

Mission-critical real-time machines aren't running Swenson-designed linear PSUs. (They're not running USB either - think about it).

 

Mission critical real time computers for industrial applications  have no real relevance to audio . With audio, we are not only concerned that all the analog voltage pulses (representing digits) that arrive at the DAC in the best order possible, but also that the  electrical signal carrying the voltage pulses that will enter the DAC and flow through it will  be as clean as possible and have the least possible distortion  to minimise adverse sonic effects.  A server may be made of computer parts but it is an audio device and all the factors affecting audio must be taken into account.   If this is going to degenerate into an alternate "digits is digits" debate, then count me out. That type of thinking  has set CA back years before people like Swenson showed the way.

Link to comment
Share on other sites
 Share





×
×
  • Create New...
To Top