Jump to content
IGNORED

Shootout at the Linux Corral: AudioLinux vs Euphony


Recommended Posts

3 hours ago, flkin said:

Here's another data point AL vs E

 

Managed to boot my Pink Faun 2.16x streamer with the demo Euphony/Stylus USB dongle. it was pretty straightforward requiring only a minor BIOS change to point to the boot USB dongle.

 

Setting up was easy too, enter the IP address into my browser and Stylus launched immediately. Log into Tidal and I could play music in five minutes.

 

Settings were:

  • Cache songs
  • No volume control, no delay
  • Buffer 100% before playing

 

Stylus playback from browser result :

  • Not bad but not at the PF-AL level. Much smaller soundstage and less engaging. Less energy to the musi

RoonCore+StylusEP via Squeezebox result:

 

Slightly better soundstage but still smaller than AL soundstage

  • Engagement level still not the same as with AL
  • Sound less forward and generally not as real sounding
  • Treble a touch harsher especially with loud female vocals

 

My AL settings are:

  • Ramroot enabled
  • Boot mode standard
  • Priority extreme
  • No cores isolated
  • Kernel 4.19.37-rt19-1-rt-bfq
  • Audiolinux 2.3.2

My findings - booting from USB demo dongle, Euphony/Stylus both for Stylus playback or StylusEP end point, isn't better than the optimised PF-AL in my PF 2.16x machine. Not by quite a bit

 

What is your player in PF? HQPe?

 

Link to comment
6 hours ago, flkin said:

Here's another data point AL vs E

 

Managed to boot my Pink Faun 2.16x streamer with the demo Euphony/Stylus USB dongle. it was pretty straightforward requiring only a minor BIOS change to point to the boot USB dongle.

 

Setting up was easy too, enter the IP address into my browser and Stylus launched immediately. Log into Tidal and I could play music in five minutes.

 

Settings were:

  • Cache songs
  • No volume control, no delay
  • Buffer 100% before playing

 

Stylus playback from browser result :

  • Not bad but not at the PF-AL level. Much smaller soundstage and less engaging. Less energy to the musi

RoonCore+StylusEP via Squeezebox result:

 

Slightly better soundstage but still smaller than AL soundstage

  • Engagement level still not the same as with AL
  • Sound less forward and generally not as real sounding
  • Treble a touch harsher especially with loud female vocals

 

My AL settings are:

  • Ramroot enabled
  • Boot mode standard
  • Priority extreme
  • No cores isolated
  • Kernel 4.19.37-rt19-1-rt-bfq
  • Audiolinux 2.3.2

My findings - booting from USB demo dongle, Euphony/Stylus both for Stylus playback or StylusEP end point, isn't better than the optimised PF-AL in my PF 2.16x machine. Not by quite a bit

Thank you for confirming.

Link to comment
7 hours ago, flkin said:

Here's another data point AL vs E

 

Managed to boot my Pink Faun 2.16x streamer with the demo Euphony/Stylus USB dongle. it was pretty straightforward requiring only a minor BIOS change to point to the boot USB dongle.

 

Setting up was easy too, enter the IP address into my browser and Stylus launched immediately. Log into Tidal and I could play music in five minutes.

 

Settings were:

  • Cache songs
  • No volume control, no delay
  • Buffer 100% before playing

 

Stylus playback from browser result :

  • Not bad but not at the PF-AL level. Much smaller soundstage and less engaging. Less energy to the musi

RoonCore+StylusEP via Squeezebox result:

 

Slightly better soundstage but still smaller than AL soundstage

  • Engagement level still not the same as with AL
  • Sound less forward and generally not as real sounding
  • Treble a touch harsher especially with loud female vocals

 

My AL settings are:

  • Ramroot enabled
  • Boot mode standard
  • Priority extreme
  • No cores isolated
  • Kernel 4.19.37-rt19-1-rt-bfq
  • Audiolinux 2.3.2

My findings - booting from USB demo dongle, Euphony/Stylus both for Stylus playback or StylusEP end point, isn't better than the optimised PF-AL in my PF 2.16x machine. Not by quite a bit

 

Thanks so much for taking the time to do this experiment. Every system is different, and ultimately our ears are the final arbiter, so I respect your findings.

 

I'm trying to reconcile your findings with the others. One question I have is this: does PF-AL have proprietary tunings that the public AL does not? If yes, I wonder if it significantly outperforms "vanilla" AL on SQ? 

 

Of course, to test this, you'd have to compare PF-AL vs. vanilla AL, and I'm not sure what appetite you have for such a comparison!

Link to comment
17 minutes ago, austinpop said:

 

Thanks so much for taking the time to do this experiment. Every system is different, and ultimately our ears are the final arbiter, so I respect your findings.

 

I'm trying to reconcile your findings with the others. One question I have is this: does PF-AL have proprietary tunings that the public AL does not? If yes, I wonder if it significantly outperforms "vanilla" AL on SQ? 

 

Of course, to test this, you'd have to compare PF-AL vs. vanilla AL, and I'm not sure what appetite you have for such a comparison!

just an other datapoint: I am not running a PF streamer but do have PF-AL and in my system E outperforms PF-AL

 

Pink Faun Streamer —>  Pink Faun DAC --> Ayre AX5 --> Paradigm S8 

Link to comment
7 hours ago, austinpop said:

 

Thanks so much for taking the time to do this experiment. Every system is different, and ultimately our ears are the final arbiter, so I respect your findings.

 

I'm trying to reconcile your findings with the others. One question I have is this: does PF-AL have proprietary tunings that the public AL does not? If yes, I wonder if it significantly outperforms "vanilla" AL on SQ? 

 

Of course, to test this, you'd have to compare PF-AL vs. vanilla AL, and I'm not sure what appetite you have for such a comparison!

 

I was told that the AL-PF was customised for the 8-core Rizen chip. How? I don't know. I suspect it's not a better version just adapted carefully to the hardware in the 2.16x. Otherwise why wouldn't Piero incorporate all his advances into the latest vanilla AL software?

 

All the AP-PF menu options seem to be the same as the vanilla AL so there must be changes outside the menu setting that have been made. Perhaps similar to the Euphony's Black List items where some less used Linux functions are switched off or replaced with a more suitable version for the PF hardware?

 

There are definitely some adjustments outside of menu. Two I've been playing with under console mode:

  • Switching off SAMBA seems to make a slight improvement in sound. Inconvenient for some as it’s necessary for moving music in and out of the server but for those with external music sources, you don’t need samba on to access AL from a browser.

systemctl status smb - (to check status)

systemctl disable smb

systemctl disable nmb

save system if in RAMroot

reboot

  • You can check whether the network card drivers you are using might have software enhancements or improvements. Under Piero's directions, I used lspci -v | grep Ethernet to discover a Kernel/Ethernet driver bug for my hardware.

The solution was to install ethtool and run ethtool -K interface tso off (substitute interface with my own interface given by networkctl). This adjustment didn’t bring about any perceivable sound improvement but it’s all cumulative in small steps and any network errors wouldn’t be good for the sound anyway.

 

You've got me wondering what difference the latest generic AL would sound like compared to the hardware optimised PF-AL on the 2.16x. 😄 I'll try to get this done this soon. Ran out of dongles…

Link to comment
1 hour ago, flkin said:

You've got me wondering what difference the latest generic AL would sound like compared to the hardware optimised PF-AL on the 2.16x. 😄 I'll try to get this done this soon. Ran out of dongles…

 

Mission accomplished!

Link to comment

As it have been said before all systems are different, although I am very conservative and I see changes to my system very pragmatically I discovered in my system the sound of stylus player alone no upsampling, no roon, no HQPE, just pure bare flac through my denafrips dac sounds better to my ears than AL, custom Ubuntu, Miskas image NAA and such 

I know, it doesn't make sense, I need to replace my ears next or jump out of this crazy train :)

 

Kidding aside I just replaced my speakers 3 weeks ago and that alone got me entirely out of reference which I just realize now I will have to start this road again 

 

Link to comment
3 hours ago, luisma said:

I discovered in my system the sound of stylus player alone no upsampling, no roon, no HQPE, just pure bare flac through my denafrips dac sounds better to my ears than AL, custom Ubuntu, Miskas image NAA and such 

I know, it doesn't make sense, I need to replace my ears next or jump out of this crazy train :)

 

You're not crazy. Different system, but your results are similar to mine.

Link to comment
3 hours ago, luisma said:

As it have been said before all systems are different, although I am very conservative and I see changes to my system very pragmatically I discovered in my system the sound of stylus player alone no upsampling, no roon, no HQPE, just pure bare flac through my denafrips dac sounds better to my ears than AL, custom Ubuntu, Miskas image NAA and such 

I know, it doesn't make sense, I need to replace my ears next or jump out of this crazy train :)

 

 

I agree. I am using the same DAC and Stylus sounds exceptionally good compared to all other s/w (and OS) I have tried earlier. I just pass native resolution and turn on the OS mode. The good part is with Dual AES/EBU support with the new DSP board, a M-Scaler can take things further without having to build high power server for up-sampling purpose.

Link to comment
2 hours ago, austinpop said:

 

You're not crazy. Different system, but your results are similar to mine.

Thank you for supporting the I'm not crazy statement lol 

But it really defies any logic 

I was listening to Within yesterday (which I found through a playlist you published some time ago) the intro piano with the decaying notes with slowing tempo, same song I have heard many times but so sweet and simple now 

 

Link to comment
2 hours ago, Dev said:

 

I agree. I am using the same DAC and Stylus sounds exceptionally good compared to all other s/w (and OS) I have tried earlier. I just pass native resolution and turn on the OS mode. The good part is with Dual AES/EBU support with the new DSP board, a M-Scaler can take things further without having to build high power server for up-sampling purpose.

Have to try OS mode 

Link to comment
30 minutes ago, bobfa said:

Luis,

I started this thread with the hope that others would be able to validate what I was hearing!   I cannot explain why these different OS variants and software can sound so different!  It makes me happy to see others enjoying their systems more, no matter what system it is.

 

Bob

Hi Bob, and I am very grateful for your contribution. 

 

 

Link to comment
On 5/31/2019 at 11:12 AM, flkin said:

No, just Roon without upsampling. Also NUMA disabled. And no hyper threading.

 

Although I've extensively used it in the past, I'm preferring Roon to HQPe currently.

I have the 2.16 and have been running HQPe for months.  I switched back to Roon straight for a few days and decided I prefer HQPe, in both cases no sampling being performed.

 

The kernel with NUMA disabled gave me a noticeable increase in dynamics.

 

I've settled on much the same as you, no isolated cores but I still have boot set to extreme.  I should try with it on standard.  

Link to comment
17 hours ago, lpost said:

I have the 2.16 and have been running HQPe for months.  I switched back to Roon straight for a few days and decided I prefer HQPe, in both cases no sampling being performed.

 

The kernel with NUMA disabled gave me a noticeable increase in dynamics.

 

I've settled on much the same as you, no isolated cores but I still have boot set to extreme.  I should try with it on standard.  

 

Please do try with Normal mode with Roon. I'm finding this combination works well. Especially with the new kernel. Extreme for me has more detail but is harder and less emotional. Prior to the new kernel I preferred Extreme1 setting.

 

I wonder if the sound of Roon with clocks helping focus is similar in presentation as HQPE which tends to focus sounds?

Link to comment
On 5/29/2019 at 1:17 AM, Nenon said:

1. How many amps do we need to sufficiently power up a CPU with TDP of 95 Watts for Euphony?

 

IMHO it's tricky to define what's actually "sufficient" when we've gotta take this into consideration

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/542/?tab=comments#comment-941760

On 3/25/2019 at 4:50 AM, romaz said:

First, I wanted as much headroom as possible.  After speaking with Sean Jacobs, he was very much in favor of over-provisioning any ATX PSU he would design for me to avoid core saturation.  In fact, his design incorporated a 300VA transformer even though I told him I was expecting my server to only consume about 30-35 watts.

 


 

On 5/29/2019 at 1:17 AM, Nenon said:

2. Can we power up the EPS connectors with multiple independent DC rails on this motherboard? The keyword is "independent", because if each +12V on the connector is in parallel with the others, we are essentially paralleling multiple power supplies, and we know by experiments with paralleling multiple LPS2.1's that unless they are closely matched, this is far than ideal (if it even works). 

 

Are we going for that Euphony + Stylus combo while gaming motherboards and processors are preferred? If so, forget about LPS-1.2 when we're looking at stuff like an i9-9900K with 95W TDP etc.

 

Let's say we're shooting for the stars here, maybe having 16 direct power phases to the CPU alone with Smart Power Stages that are rated at 70A might do the trick?

 

https://www.gigabyte.com/Motherboard/X570-AORUS-XTREME-rev-10

 

Of course the most important part here should be having two EPS connectors instead of one, 105W TDP for 12-core Ryzen 3 3900X while we don't necessary know the actually power consumption of base clock versus boost clock.

 

Since the availability of SR7 rails should be comparable to that of hen's teeth at the moment, maybe combining two 12A rails (i.e. 144 + 144 = 288W) as quoted below could be our best bet as long as the price is right?

 

http://www.faradpowersupplies.com

Quote

Farad power supplies incorporate the latest of capacitor technology, EDLC super capacitors. The Farad power supply is available in one version: a 3A output (1-3F capacity). A 12A (15-50F capacity) version is currently on the test bank.

 

For other high-current 12V rails that are powered by linear transformers, here's what they're claiming in China

 

12V/24.5A

https://item.taobao.com/item.htm?id=35319195801

https://item.taobao.com/item.htm?id=35604325201

https://item.taobao.com/item.htm?id=41531479282

https://item.taobao.com/item.htm?id=590818543597

https://item.taobao.com/item.htm?id=591457233653

https://item.taobao.com/item.htm?id=591457661424

https://item.taobao.com/item.htm?id=596102786830

 

12V/15A

https://item.taobao.com/item.htm?id=584109188945

https://item.taobao.com/item.htm?id=591222416438

https://item.taobao.com/item.htm?id=591457145630

https://item.taobao.com/item.htm?id=592312329878

https://item.taobao.com/item.htm?id=592640771609

https://item.taobao.com/item.htm?id=596093950210

 

12V/7A from Hong Kong

 

https://www.clonesaudio.com/powerstation

Link to comment
  • 2 weeks later...
On 5/24/2019 at 8:43 PM, romaz said:

 

I never thought I would leave Roon having paid for a lifetime membership a number of years back.  It remains second to none with respect to library management and overall user experience but I have found the Stylus player to sound TOO GOOD to ignore and so I have made the switch. 

 

Roon by itself has a very good bloom and liquidity to it but it is at the expense of control resulting in considerable overhang and smearing of details, at least to my ears.  For vocals, Roon sounds more than just acceptably good but for orchestral music, it really is a mess.  Roon + SqueezeLite provides more of this control and much needed precision resulting in better damping and cleaner transients.  It's easier to hear when notes start and stop giving you the perception that the noise floor is lower but it also has a tendency to sound dry and mechanical.  While tolerable with orchestral music, with vocals and solo instruments, there is a sterility and thinness to SL (even with large buffers) that has always left me wanting.  The brain tells me all is well but the heart tells me differently.  Stylus is exceptional in it's ability to provide the bloom, liquidity, and tonal richness of Roon but also the the timing precision of SL.  In fact, to my ears, Stylus actually does all of these qualities better.

 

When I was doing my tests on AudioLinux, I found that ramping up CPU frequency resulted in better dynamics and an overall more muscular sound.  On an i7-8700K, this meant CPU frequencies as high as 4.7GHz and while this improved dynamics was very pleasing, it also came at the expense of harshness and an inability to convey delicacy and nuance.  As I set CPU frequency to the other end of the spectrum to a fixed 800MHz (and even 400MHz), delicacy and nuance was there in spades with no apparent harshness but the sound signature was thin and anemic sounding in comparison, similar to what I hear with an sMS-200ultra or ultraRendu.

 

Somehow, with the right CPU, Euphony is capable of providing both the benefits of high and low CPU frequency where it can do dynamics and expansive sound stage but also subtlety and nuance while never sounding harsh.  In hindsight, probably one of the worst features ever developed for AudioLinux is the Extreme2 mode because it forced you to a single frequency.  I think it's best to allow the CPU the flexibility to scale to whatever frequency is called for by the track.   While there is more to Stylus' magic than that, it is Stylus' ability to be both muscular and delicate that has forced me to rethink my digital front end once again.

 

While in Munich, I had a discussion with Jord Groen of Pink Faun about his decision to go with an AMD 1800X CPU.  It was his opinion based on listening tests that with his version of AL, CPU frequency was nowhere as important as the number of cores and so he never felt it necessary to move to a more powerful CPU with higher CPU frequency capability.  My experience is that CPU frequency definitely adds something but ultimately, harshness was the biggest trade off and so I presume this is what Jord meant.

 

With AudioLinux, I was forced to cap an 8700K at 3.8GHz because any speeds beyond this sounded harsh and so this supports Jord's comment.  But with Euphony and Stylus, I let the 8700K run with no cap whatsoever and notice that CPU frequency will typically reach 4.4-4.5GHz consistently, however, I get no harshness at all.  In fact, the higher the frequency a CPU is capable of, the better, and so I imagine the ultimate CPU at this time would be something like an i9-9900K that provides 8-cores and a max frequency of 5GHz and a TDP of only 95w.  I hope to be able to test this soon but what I will say is that with Euphony + Stylus on a single box 8700K machine, my reclocked i7 NUC driven by a 19V SR7 sounds absolutely puny.  Even with the 8700K server powered by an HDPlex 400W ATX LPSU and with no special clocking outside of my tX-USBultra, I prefer the big server to the i7 NUC but as I have figured out a way to independently power both the 8700K CPU and ATX motherboard (using a DC-ATX converter) with SR7 rails, the i7 NUC has now been officially retired.

 

Do I think that a powerful single box server running Euphony + Stylus is better than a dual box machine running Roon + StylusEP?  At this time, yes, no question.  Stylus is that good.  StylusEP contains a subset of Stylus but they aren't the same.  To my ears, Stylus sounds smoother and richer and better textured.  It is also ultra stable and Željko has done a wonderful job enhancing its feature set.  In fact, expect further enhancements in the coming days.

 

I have now also tested Roon vs. Euphony and came to the same conclusion.  I will test vs AudioLinux shortly.

 

Link to comment

Hi Gavin, thanks for posting your results.

 

When testing Euphony with the default config of the Stylus player I was very surprised by the SQ as well, I communicated with Euphony back and forth but at some point I opened a trouble ticket (still on trial and haven't purchased the software yet) with a few questions and never got a response back, I will follow up with them as I'm sure the response was somehow lost or something. Because of this I went back and tested AL on a server NUC core i7 with Roon Server streaming to an HQP endpoint running Audiolinux as well. SQ is very good and it beats the single server solution on my system. I have a Denafrips terminator connected via USB in NOS mode (old Amanero board) whether possible include that design on your testing. I am prioritizing the IRQ for the network interface and Roon only on the server side and I am prioritizing the USB, network IRQ and networkaudiod on the NAA side and that made some of a difference compared with the out of the box AL.

 

I am just waiting for the zen2 models to come out to build a fanless server for audio and then do the testing with single / dual boxes all over again. Also waiting for the new board for my DAC which might make a difference as well because of the new specs. 

 

So at this point I'm kind of enjoying music without worrying too much about other changes until July or so when I get my zen2 and do this all over again :)

 

 

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...