Shootout at the Linux Corral: AudioLinux vs Euphony

[lpost]

My I2S card only handles PCM so the DSD settings in HQPe are not relevant.  Set filters and dither to none and it won't upsample.

[luisma]

8 hours ago, lpost said:

My I2S card only handles PCM so the DSD settings in HQPe are not relevant.  Set filters and dither to none and it won't upsample.

Thanks 

[bobfa]

On 5/23/2019 at 9:11 PM, bobfa said:

This thread has turned to something different for me.  It has turned into a  mission to standardize my setup for simplicity.

 

 I have been refining my systems design over the past few weeks and both AL and Euphony have been providing regular updates. Everything is now working as I want it to. Here is a quick outline of the system cleanup and testing I have done to prepare for the shootout.

My home network is now on all Ubiquity Unfi hardware.  The 24 port switch has two fiber ports.  One port connects to my server, and the other connects to my endpoint.  The FMC devices are connected powered linear power supplies.  I have upgraded Ethernet cables from the FMC to the end devices.  Both AL and Euphony can now use WiFi and appear stable.  I have a dedicated SSID for the system that only broadcasts on 5Hgz.

 

• I have conducted multiple tests over the last couple of weeks, and I cannot hear a difference from my server when it is bridged and directly connected to the endpoint.  I think that the switch may have made a difference.
• I had previously tested my NAS vs. External HD over USB. After an evening at @johnseye and listening to his system with the USB upgrades, I see that there is a possibility here to improve further.  As of now, the NAS is where I am keeping my music.  I have not been able to hear a difference in my system.
• I have both Fiber and Ethernet cables running from the basement to the living room.

 

I can boot AL from USB and run in RAM on my i7 NUC endpoint. The 32GB Optane card in the endpoint is where I have Euphony installed.  The endpoint can swap simply between OS and audio software.

 

The server is “fixed,” it is running Euphony and Roon server.  The server has bridging turned on as needed. 

 

I have a set of three tracks I am using to evaluate these configurations

 

 

 

I went through all the tracks multiple times with each setup.  I can swap systems designs using a Web Browser on my iPad in the listening chair. The average swap time is about 40 seconds. 

 

The Tiebreaker for me is between dual system Roon w/StylusEP on the endpoint and single box Stylus. Today after about three hours of listening pointing me to the dual system as best I found that taking a 2-hour break and then coming back to listen to track "Artifacture" by BT on his untitled album.  Just listen to the segment from 6:00 to about 7:00 a couple of times.

 

 

If I had never built the server and had been running a single box system I could have been satisfied.  I have not gotten to endpoint re-clocking, I am not using my TLS switch.  I am sure that I could optimize AudioLinux further.

 

Here is my current status:

 

SW for endpoint	Notes	"Results"
AudioLinux	V 1.3- menu 118 ramroot, boot extreme,operate extreme
Roon Bridge		I used to think this was great
SL Default		This told me that Roon Bridge is not as good as it could be
SL w Large Buffer	Not tested yet
Euphony	Running 3.0 the 0520 release
Roon Bridge		This is better than the AL setup. more space, depth.
SL Default		This takes us up a couple of notches, the depth and "blackness" improve.
SL w Large Buffer	Not tested yet
Stylus EP		The stage opens up the music is expansive and there is much more detail. This is still the best I have done BUT have not spent enough time on the standalone setup.
Stand Alone Testing	I do not pay attention to this to start with... I should have.
Roon	Skipping for now.
Roon +SL	Skipping for now.
Roon +SL large Buffer	Skipping for now.
Roon + Stylus EP		I am having a very hard time telling the difference between this and Stylus!
Stylus		This is very close to the dual box, but I still fall behind on expanse and depth.

I want to add some notes to this sheet. I am also putting a pin in the map and moving along.

 

I have been running the Xeon as a single box system and I have added and Element H USB card to the machine powered by the HDPLEX400.  I am now to the point that I hardly hear the differences between Stylus and Roon with Stylus EP.   Things have gotten very subtle Very Subtle.

 

I am sure I could improve the SQ more with clocking changes or MB or any of the other ways we have found to make incremental improvements.  

 

I have another direction I am testing that I will report on when the work is done.  

 

Bob

 

[Gavin1977]

Thanks Bob - I have been following your posts (and music) really helpful.

 

What are SL and SL + Large buffer configs and their effect?  I forget what most of the acronyms mean :-)

 

Does the default playback engine (not HQPlayer) in Stylus EP automatically upscale 16/44 and maps it to 16/388 (or whatever your DAC will accept)  like it does in the full version of Euphony?

 

I think I'm also starting to conclude the Roon (on a server elsewhere) plus a Stylus or AL endpoint is the best option - mainly from functionality purposes as Roon is a lot easier / consistent, it just a shame that Roon is comparatively expensive.

[bobfa]

6 hours ago, Gavin1977 said:

Thanks Bob - I have been following your posts (and music) really helpful.

 

What are SL and SL + Large buffer configs and their effect?  I forget what most of the acronyms mean 🙂

 

Does the default playback engine (not HQPlayer) in Stylus EP automatically upscale 16/44 and maps it to 16/388 (or whatever your DAC will accept)  like it does in the full version of Euphony?

 

I think I'm also starting to conclude the Roon (on a server elsewhere) plus a Stylus or AL endpoint is the best option - mainly from functionality purposes as Roon is a lot easier / consistent, it just a shame that Roon is comparatively expensive.

SL stands for Squeezelite.  

There are settings in Squeezelite that allow you to adjust the buffering in the software.  Tuning  them can change the sound.

 

I would have to look to see if StylusEP does that.  I do not remember and I am running different hardware in my endpoint right now.  I may be testing again ia a few days.  Maybe someone else can look.

[luisma]

On 5/28/2019 at 9:16 AM, tapatrick said:

Down the road I'm going to try out a Ciunas Audio Supercap PSU - but that's a whole different topic

Hi did you ever tried the Ciunas?

[tapatrick]

7 hours ago, luisma said:

Hi did you ever tried the Ciunas?

I have ordered a 7.5v version and its on it's way. I tried out John Kenny's 5v version powering my Aqvox switch (it is on tour over at Pinkfish forum) and it is very good, having a calming effect on audio running through it. He is developing a variable version 12-25v.

[tapatrick]

10 hours ago, tapatrick said:

I have ordered a 7.5v version and its on it's way. I tried out John Kenny's 5v version powering my Aqvox switch (it is on tour over at Pinkfish forum) and it is very good, having a calming effect on audio running through it. He is developing a variable version 12-25v.

 

18 hours ago, luisma said:

Hi did you ever tried the Ciunas?

..in fact the Ciunas at 5v compared favourably to the PH SR4 as I couldn't distinguish the difference between the 2..

[lpost]

On 6/4/2019 at 5:09 AM, flkin said:

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

 

I had to reboot this morning and once again, 4 times in the past year, HPQe started up with a new fingerprint.  So, I decided to run Roon straight for a bit and I'm liking what I hear.  Perhaps I'll stop fussing with HQPe.

 

Apparently, AL isn't supported so it presents a bit of a problem with the key...

[Crom]

7 hours ago, lpost said:

 

I had to reboot this morning and once again, 4 times in the past year, HPQe started up with a new fingerprint.  So, I decided to run Roon straight for a bit and I'm liking what I hear.  Perhaps I'll stop fussing with HQPe.

 

Apparently, AL isn't supported so it presents a bit of a problem with the key...

I'm saddened ... but comforted ... to see that others are having the same problem as I. For whatever reason I have been completely unable to obtain a stable fingerprint on boot up (although it has now remained the same for a week!). My distro is audiolinux too. I wish Jussi and Piero would put their heads together and come up with a workable solution to this. Both are great bits of software and I have been using both in complete harmony for years...until recently. I understand that with software it is often very difficult to pinpoint the issue when there is more than one system in play, but us poor users are the ones who end up suffering.

 

I am going to wait a few more days until I see if this fingerprint chooses to hold but in the meantime, I have had to listen to 30 minutes and then reboot. Not much fun :-(

[flkin]

11 hours ago, lpost said:

 

I had to reboot this morning and once again, 4 times in the past year, HPQe started up with a new fingerprint.  So, I decided to run Roon straight for a bit and I'm liking what I hear.  Perhaps I'll stop fussing with HQPe.

 

Apparently, AL isn't supported so it presents a bit of a problem with the key...

 

Ah I take it those are recent issues with the fingerprint. I haven't been using HPQ for a while now so didn't know.

 

I made some recent experiments in the Bios inspired by your frequency adjustments that I'm finding quite nice. I'll post my findings in the Pink Faun thread.

[lpost]

I've had the fingerprint change on me 4 times during the past year.  So not new but I went months in between changes, with many reboots along the way.  It is seemingly random when it decides to gen a new one.

 

Piero says he's communicated everything he knows to to Jussi but it appears that ramroot is the main culprit regarding the fingerprint changing.  Jussi told me that 4.10.2+ has adjusted code so the fingerprint should remain more stable and in fact on my system through numerous reboots it has now remained the same, unfortunately I'm waiting for Jussi to respond with a new key.  I hate to bother him as it's got to be an annoyance for him as well.

 

That said, I'm going to give Roon a good run for a few weeks.  It's sounding every bit as good as with HQPe with the last few months of improvements in AL tuning thanks to Piero.  I like ramroot for the improvement in Roon db responsiveness.  Super snappy.

 

My 2.16 is just over a year old last week and continues to impress me with it's ability to produce a fantastic signal for my dac and ultimately for my ears. 

[davide256]

I will be making the change from AL/Rooncore to Euphony/Rooncore. Choral harmonics ( and others) have been nagging me with "subliminal" harmonic blurriness

for the past few months, so much so that I was starting to think a DAC  upgrade was needed. Thankfully the combination of Euphony /Roon server NUC

takes care of about 90% of the issue and when I replace AL on endpoint with Euphony/ Stylus EP the remaining 10% is taken care of. 

 

While  the 1 box Euphony Stylus setup is nice, I find with my system that it loses transients, sounds compressed vs  Euphony/Rooncore 1 box solution.

[luisma]

On 7/11/2019 at 8:20 PM, lpost said:

 

I had to reboot this morning and once again, 4 times in the past year, HPQe started up with a new fingerprint.  So, I decided to run Roon straight for a bit and I'm liking what I hear.  Perhaps I'll stop fussing with HQPe.

 

Apparently, AL isn't supported so it presents a bit of a problem with the key...

I had no problems moving my HQPE from Ubuntu to AL, as long as the hardware is the same you should be fine. This may sound obvious but be sure you are not ramroot when entering the license for HQPE as it won't be saved. Mine is working great. 

[luisma]

On 7/12/2019 at 4:18 AM, Crom said:

I'm saddened ... but comforted ... to see that others are having the same problem as I. For whatever reason I have been completely unable to obtain a stable fingerprint on boot up (although it has now remained the same for a week!). My distro is audiolinux too. I wish Jussi and Piero would put their heads together and come up with a workable solution to this. Both are great bits of software and I have been using both in complete harmony for years...until recently. I understand that with software it is often very difficult to pinpoint the issue when there is more than one system in play, but us poor users are the ones who end up suffering.

 

I am going to wait a few more days until I see if this fingerprint chooses to hold but in the meantime, I have had to listen to 30 minutes and then reboot. Not much fun :-(

This is weird, I have an Intel nuc, you guys have full PCs? Mine hasn't changed and I update server and NAA every two days or so and running latest versions on everything. 

[luisma]

On 7/12/2019 at 11:59 AM, lpost said:

It's sounding every bit as good as with HQPe

Wait till you try the new EC modulators in HQPe, 

[luisma]

10 hours ago, davide256 said:

While  the 1 box Euphony Stylus setup is nice, I find with my system that it loses transients, sounds compressed vs  Euphony/Rooncore 1 box solution.

I found stylus sonically better than rooncore, of course library handling and interface not even close to Roon but you report that roon sounds better, are you playing dsd or PCM? What dac do you have, I'm just curious 

[luisma]

On 7/12/2019 at 11:59 AM, lpost said:

but it appears that ramroot is the main culprit regarding the fingerprint changing

Never mind, I'm only ramroot on the naa not the server so your post explains it, sorry to hear. 

[davide256]

57 minutes ago, luisma said:

I found stylus sonically better than rooncore, of course library handling and interface not even close to Roon but you report that roon sounds better, are you playing dsd or PCM? What dac do you have, I'm just curious 

Most of my library is PCM and I tend to play a lot of movie soundtracks. The DAC is Metrum Octave I.  Its possible that if I were to spend over $2K on a newer DAC that

my preference might change, will have to "soak" using Euphony/Roon for a few months to see if any new "subliminal" issues rise to the surface. Have reached the point with my system

that few things are obviously broke and I follow the "if its not broke, don't fix it" mantra.

[luisma]

55 minutes ago, davide256 said:

55 minutes ago, davide256 said:

few things are obviously broke and I follow the "if its not broke, don't fix it" mantra.

 

Agreed

56 minutes ago, davide256 said:

Most of my library is PCM and I tend to play a lot of movie soundtracks

Well I noticed that playing PCM and even upsampling PCM as little as 44.1 to 48 in Stylus creates a difference, more IMO than going from 44.1 to 384 (with my DAC and in proportion) for example from 44.1 to 48 80% improvement from 44.1 to 384 100% to use some numbering. DSD is a different story and recently found that the only time I prefer DSD over PCM is with the new HQPe modulators with DSD128 which is the max I can reach with ASDM7EC.

 

I have been warned to not enter into PCM / DSD discussions on other forums, I hope @bobfa doesn't consider this to be OT as I believe it is directly related to the whole AL / Euphony comparison.

 

And I don't want to start that discussion either just posting my opinion because with my Denafrips Terminator (old Amanero board) it does makes a night and day difference to me and what software / player I prefer over another.

 

 

[luisma]

1 hour ago, davide256 said:

PCM and I tend to play a lot of movie soundtracks

And since you mentioned the movies you may prefer either one because of bass reproduction is very good in both cases in PCM, upsampling DSD sounds a lot cleaner and open at the expense of bass presentation moving back IMO and OMSS (on my specific system)

 

[bobfa]

43 minutes ago, luisma said:

Agreed

Well I noticed that playing PCM and even upsampling PCM as little as 44.1 to 48 in Stylus creates a difference, more IMO than going from 44.1 to 384 (with my DAC and in proportion) for example from 44.1 to 48 80% improvement from 44.1 to 384 100% to use some numbering. DSD is a different story and recently found that the only time I prefer DSD over PCM is with the new HQPe modulators with DSD128 which is the max I can reach with ASDM7EC.

 

I have been warned to not enter into PCM / DSD discussions on other forums, I hope @bobfa doesn't consider this to be OT as I believe it is directly related to the whole AL / Euphony comparison.

 

And I don't want to start that discussion either just posting my opinion because with my Denafrips Terminator (old Amanero board) it does makes a night and day difference to me and what software / player I prefer over another.

 

 

I am always interested in how folks perceive the differences in systems, formats, etc.  I have not tried to A/B PCM/DSD differences between AL and Euphony. It would interesting to see some testing in that space.

[luisma]

Hi Bob, I am waiting as DHL just left a notice for delivering my new DAC board which upgrades my DAC to support DSD1024 and PCM 1536kHz as well as other improvements like new FPGA code, new 5 LDO voltage regulators, improved quality, which pretty much is going to send me back to zero and to the drawing board of testing , funny feeling I'm excited for what it could bring but I already got used to like the existing board (now obsolete). I have to do a test to see if Euphony works fine with this board, the old one Zeljko had to do specific adjustments to support it, and so on so it will be fun. Will report back.

 

[TheAttorney]

On 7/14/2019 at 3:43 AM, davide256 said:

While  the 1 box Euphony Stylus setup is nice, I find with my system that it loses transients, sounds compressed vs  Euphony/Rooncore 1 box solution.

Could you elaborate a bit more about this please?

 

You say you use PCM mostly. Does that include standard redbook?

Any DSP involved?

So Roon is effectively headless and controlled only by a remote tablet/PC/phone?

And direct USB cable connection from server to DAC?

With any USB cleaner/reclocker in the path?

 

I have a lifetime roon license that is currently gathering dust. So I'm hoping that roon will one day catch up on SQ, but I'm not holding my breath - with the roon team seemingly not accepting that there's a problem that needs fixing, and a very vocal, audiophile-bashing, bits-is-bits brigade who jump in at every opportunity on the roon forums.

[davide256]

6 hours ago, TheAttorney said:

Could you elaborate a bit more about this please?

 

You say you use PCM mostly. Does that include standard redbook?

Any DSP involved?

So Roon is effectively headless and controlled only by a remote tablet/PC/phone?

And direct USB cable connection from server to DAC?

With any USB cleaner/reclocker in the path?

 

I have a lifetime roon license that is currently gathering dust. So I'm hoping that roon will one day catch up on SQ, but I'm not holding my breath - with the roon team seemingly not accepting that there's a problem that needs fixing, and a very vocal, audiophile-bashing, bits-is-bits brigade who jump in at every opportunity on the roon forums.

I used the latest Euphony image with trial registration on a NUC8i3BEH, powered at 19v by an HDPLEX 200W supply

Software setup was Euphony/Stylus vs Euphony/Rooncore+Stylus EP, 1 box setup

No  DSP used, controlled via iPhone/iPad using http:\\{ IP address} for Euphony and Roon client for Roon

USB device is Schiit Eitr with Lush 2 USB cable and AQ Eagle Eye coax to DAC. I had an ISO Regen and sold it as detrimental after switching to AL

 

Primarily I was testing with the Sony 192/24  recording of Williams -Spielberg collection disc III and random samplings from other recordings. My collection is largely CD

quality, some DSD from San Francisco Symphony that is converted to PCM but most sound track recordings are available in CD quality only.

 

I found that Euphony was very pleasing in a KEF sort of way, great tonality and separation between instruments, no offensive edges. But using

Roon via same setup, I was hearing more "air",  sharper edges on percussion sounds and dynamics were bigger, in the way that  an Air Force

band or Harry James like band performance  behaves. Both fixed 90% of the issue I was having with AL, where complex choral harmony sounded frayed, words lost

coherency

 

It is possible that my Roon preference is because it overcompensates for deficits of Metrum Octave I DAC.  But I feel covered for that with Euphony if I trial a newer DAC

and find that to be so. I am surprised that the 1 box setup does so well, that has me wondering if a better 19v power supply purchase would eliminate any need

for the LPS 1.2 powered Stylus endpoint

 

In the FWIW column, after registration I used the install to SSD option... that was disappointing. Back to booting from a USB3 stick until I can sort through

an Optane SSD purchase.