Shootout at the Linux Corral: AudioLinux vs Euphony

[Downtheline]

3 hours ago, Peter Avgeris said:

Supermicro board, Intel Xeon CPU, 12Gb ECC ram for the server and 6Gb ECC ram for the player. They are connected directly without network switch (bridged NICs). The NAS is a Supermicro 10 core 20 threads Intel Xeon CPU with 128Gb Reg ECC RAM,

I'd love to hear more about your builds. Are you using fanless cases? How do you power the player and the I2s card?

[Holzohr]

16 hours ago, davide256 said:

For Roon, I have found that when using it the HQplayer option for Stylus endpoint does improve sound,you might want to try that also.

 

11 hours ago, Peter Avgeris said:

I am not using HQplayer because I never liked this program. HQP is basically a DSD interface.

 

 

@Peter Avgeris No need to use the HQPlayer.. just use the HQPlayer 'path' on Roon for StylusEP. A very nice tweak from the Euphony team.

 

About USB interfaces. I was testing Euphony the very first time with a Gigabyte H110N board and an Intel Core i7-7700T one year ago and yes I was that 'blown away' that I purchased a licence. And that even by listening through the native USB ports. (I had the Windows Server 2019 core running on that system before and AO, Fidelizer, Jplay...). 

However I thought it is a good idea to have a more powerful system (due to HQPlayer) and bought an i7-8700 cpu and an ASRock Z390M-itx/ac board. And the irritations began... The Euphony sound through the native mainboard USB ports didn't blow me away anymore. A Matrix Element H card made the things better but the problem is (imho) Euphony or AL gives realtime priority to the native USB ports by default. Well I went back to Windows Server 2019 (HQPlayer, JPlay, Roon Bridge) again for this Intel/ASRock system and used Euphony (Stylus, Roon Core, StylusEP) on a NUC7I7DNKE in Akasa case for a couple of months.

 

Yesterday my NUC8I7BEH arrived (interestingly enough it gives me a 'Summus' option in the settings. For enabling the 'exclusive features' you have to contact the support.). Well, running Stylus or the Roon Core on the NUC8I7BEH and StylusEP on the NUC7I7DNKE now but for sure I will try the NUC8I7BEH as stand-alone device, too. I like to play/test and like to have the possibilities to play around though it gives me a 'headache' now and then. I was still wondering why the Gigabyte H110N/Intel i7-7700T system gave me such a great sound with Euphony (and even more wondering why I have sold that system!). Though the Gigabyte board doesn't have these DAC-UP USB ports that Jussi (HQPlayer) likes so much. After reading a couple of days ago that Emile from Taiko Audio had a preference for the Intel Core i7-7700K I was confused even more. It seems I already had a very nice sounding system without to know about the possible hardware reasons for.

 

Sorry for my English.

 

Cheers,

Mario

 

 

The NUC overkill: NUC8I7BEH and NUC7I7DNKE (below).

[Peter Avgeris]

12 hours ago, Downtheline said:

I'd love to hear more about your builds. Are you using fanless cases? How do you power the player and the I2s card?

Hello!

Thanks for your message.

I have visited more than 100 audio systems, the majority of which are in excess of 100k of value. None of them (And I need to underline in bold) was tolerable to my ears in their digital domain. You know, we, the analog guys, we are searching tor connection to the inner meaning of music, musical talent, musical expression, thrill and excitement. I don't think that it would be wise to name some of the digital gear that I have come across with because they are businesses that make their living out of it, they hire people to work and earn their livings as well. I am a pro in the field and it is not to my mentality and attitude to write flames. I strongly prefer to compete in the real arena and not through the keyboard. But on the other hand, to my taste and feeling, nothing in the commercial arena of digital audio reproduction comes close to what we, the vinyl guys, are addicted to. Nothing.

 

The best digital gear that I have listened to, is strictly DIY. Some friends have made very good sounding digital gear, without necessarily spending a fortune. Not of course up to the level of top analog machines, but this is another story. We love music, not the hardware. As soon as something is reproduced nicely, then let's go listen. Right? So I will start writing about my journey here, as a hobby only. I am not involved professionally into digital streaming nor will I intend to do so.

 

I have not yet finished my streamers but my findings are quite clear I think. First of all, you do not need to buy exotic motherboards and extreme OCXOs, neither spend a fortune in hardware. You can find gems in the second hand market that cost almost nothing. I have also verified that when you feel obsessed with super clocks, extreme cards, etc, and you constantly buy and buy new things, in my opinion this can be considered as a point of dissatisfaction with what you currently have. You won't make things better this way.

 

I have evaluated more than 10 CPUs, another 10 motherboards, several RAM chips, etc. During my journey I had the chance of evaluating several USB converters. Power supplies (switching of course), cables, etc. None of them succeeded in persuading me that it can really connect me to the musical event in a way that this is the thing I was searching for, as the end of my digital adventure. I don't know if you get my meaning. This has nothing to do with comparison issues. Of course differences exist, there is always a winner and a loser in this procedure. I try to judge this part or accessory not in relative terms, if it is better or worse, but try to evaluate if it really makes a sparkling difference in connecting me to the musical event. Not in relative manner but in absolute terms.

 

Of all things I have lived with in my digital life, no other gear besides one (and only one) has ever blended all those musical elements together and served me a digital stream that is not only tolerable but lovely! I don't care if it does not sound like exotic vinyl systems, I don't care if it will never sound like vinyl because I don't care of A-B comparisons. I just want to listen to music at least in pleasant and joyful way, if not enthusiastic (if possible). And I have too many albums to enjoy in the next years. So, this accessory is my Pink Faun I2s bridge interface, through which you get I2s stream directly to your DAC input. Personally I can't live without this, so my whole setup is designed and built around this thing.

 

Besides this card, there are of course Linear power supplies, network switches, etc. A lot of stuff. I am not currently using linear power supply, I know what this thing does. I am not in a hurry, I will build it pretty soon. I will be using a fanless case because it is so nice to be calm around. I will write again about my findngs regarding processors, motherboards, RAM chips etc. Stay tuned for more.

[davide256]

Its possible to go down a "false trail" with digital audio, I did this with JPlay and AO which do a great job on transients and detail... but as I improved electronics I realized

that midrange was being "bleached", tone colors were weak and over shadowed by transients, kind of like having a band where the drummer plays over everyone else.

One of the very clear differences I can hear between Roon and Stylus server with Euphony is that midrange instruments have a natural decay with Stylus whereas

with Roon notes decay, drop to inaudibility too fast, very similar to what I experienced with JPlay and AO.

[Peter Avgeris]

5 hours ago, Holzohr said:

 

 

@Peter Avgeris No need to use the HQPlayer.. just use the HQPlayer 'path' on Roon for StylusEP. A very nice tweak from the Euphony team.

 

About USB interfaces. I was testing Euphony the very first time with a Gigabyte H110N board and an Intel Core i7-7700T one year ago and yes I was that 'blown away' that I purchased a licence. And that even by listening through the native USB ports. (I had the Windows Server 2019 core running on that system before and AO, Fidelizer, Jplay...). 

However I thought it is a good idea to have a more powerful system (due to HQPlayer) and bought an i7-8700 cpu and an ASRock Z390M-itx/ac board. And the irritations began... The Euphony sound through the native mainboard USB ports didn't blow me away anymore. A Matrix Element H card made the things better but the problem is (imho) Euphony or AL gives realtime priority to the native USB ports by default. Well I went back to Windows Server 2019 (HQPlayer, JPlay, Roon Bridge) again for this Intel/ASRock system and used Euphony (Stylus, Roon Core, StylusEP) on a NUC7I7DNKE in Akasa case for a couple of months.

 

Yesterday my NUC8I7BEH arrived (interestingly enough it gives me a 'Summus' option in the settings. For enabling the 'exclusive features' you have to contact the support.). Well, running Stylus or the Roon Core on the NUC8I7BEH and StylusEP on the NUC7I7DNKE now but for sure I will try the NUC8I7BEH as stand-alone device, too. I like to play/test and like to have the possibilities to play around though it gives me a 'headache' now and then. I was still wondering why the Gigabyte H110N/Intel i7-7700T system gave me such a great sound with Euphony (and even more wondering why I have sold that system!). Though the Gigabyte board doesn't have these DAC-UP USB ports that Jussi (HQPlayer) likes so much. After reading a couple of days ago that Emile from Taiko Audio had a preference for the Intel Core i7-7700K I was confused even more. It seems I already had a very nice sounding system without to know about the possible hardware reasons for.

 

Sorry for my English.

 

Cheers,

Mario

 

 

The NUC overkill: NUC8I7BEH and NUC7I7DNKE (below).

 

@Holzohr

I can understand your frustration with all this unwanted adventure and failed experiments.

I have used very old MoBos with CPUs and got  sound that the newer CPUs simply didn't. I will write more about this.

 

[ted_b]

Peter,

Thanks for the informative posts, and welcome to AS.  Two things:

1) has Pink Faun ever said why their I2S bridge doesn't support DSD (nor higher rate PCM than 192k)?  My Singxer SU-1 does not have these techincal limitations (although not commenting on SQ of course)

2) although I understand why you dubbed HQplayer a DSD product, I have used it to also (in addition to upsampling DSD to DSD512) upsample PCM to 352/384k+ and found it to be wonderfully musical, with very good PCM filtering/dithering options.

 

Sorry for keeping this a bit OT.  PM me if you'd rather.

[Peter Avgeris]

1 hour ago, ted_b said:

Peter,

Thanks for the informative posts, and welcome to AS.  Two things:

1) has Pink Faun ever said why their I2S bridge doesn't support DSD (nor higher rate PCM than 192k)?  My Singxer SU-1 does not have these techincal limitations (although not commenting on SQ of course)

2) although I understand why you dubbed HQplayer a DSD product, I have used it to also (in addition to upsampling DSD to DSD512) upsample PCM to 352/384k+ and found it to be wonderfully musical, with very good PCM filtering/dithering options.

 

Sorry for keeping this a bit OT.  PM me if you'd rather.

1. Go to
https://www.cmedia.com.tw/products/PCI_PCIe_AUDIO/CM8888

The chip does not support DSD streams. Studio mastering is not made in DSD domain but only in PCM.

 

2. It has never happened so far (for me always) to experience anything PCM upsampled and be tolerable to my ears. Either software upsampling or hardware upsampling. Not one exception. Two months ago I had listened to the very good Holo Audio Spring 2 DAC. It was so damn good in NOS mode. Switched to OS and good music was gone. No more than 5 seconds of testing.

 

You may laugh here, of my crazy findings, but OK, you can always give a try.

The best non-discrete DAC I have ever listened to, and by a great margin, was the first Philips TDA1540 in dual chip configuration, as these are mono DAC chips. It blew away TDA1541 in seconds, with all else being equal of course. 14 bits vs 20 or 24 of modern chips. In musical terms everything was clearly & undeniably better. This little thing together with Pink Faun is a miracle. It outperforms 99% of all other USB/DAC/R2R combinations experienced.

 

Pink Faun adapter has one very serious disadvantage. I have tried more than 6 motherboards, modern ones, that all failed to perform under Euphony Linux. Only two of them play fine with Pink Faun. Both these two are very old models with CPUs of ~10 years of age. While Pink Faun works fantastic under windows, it does not work under Linux in most Intel configurations. I do not know why does this happen. Nobody does. Maybe the CM8888 chip is related to specific CPU instructions that are not properly executed in various MoBo/CPU combos in Linux kernels. If interested I can let you know of the proper combinations of CPU and MoBo.

 

I am not interested in trying more USB adapters. All sound more or less the same. With Pink Faun attached preferably on a PCIe lane that talks directly to the CPU, you can drive directly a R2R ladder DAC with 32bit right justified I2s stream bypassing all this stuff of chips that generate this intolerable digital signature. DSD plays fantastically in Euphony through this card, by converting DSD stream to PCM.

 

Another alternative solution that was also very interesting was to take the DSD stream through a USB2I2S bridge like JLsounds, not feeding a DAC but a pair of step up transformers. This was an idea that came into my mind some years ago. So, I winded a pair of transformers at 40x step up ratio and the result was indeed very interesting. DACless digital to analog conversion. This is also very good idea but need to try it with more focus and attention to detail.

[ted_b]

1 hour ago, Peter Avgeris said:

Studio mastering is not made in DSD domain but only in PCM

Well, not really.  For example, our NativeDSD.com site is chock full of examples of music that were mastered (edited, analog-to-DSD transferred, etc) in DSD.  Hell, early Sony DSD stuff was ALL DSD mastering.  Even Pyramix edited multi-mike stuff can be 99.99% DSD (with microseconds of DXD PCM in the edits).  The results do not need to be listened to in PCM, but rather DSD.   If you define "mastering" as sweetening, well then we have a different issue, yes.

 

1 hour ago, Peter Avgeris said:

Two months ago I had listened to the very good Holo Audio Spring 2 DAC. It was so damn good in NOS mode. Switched to OS and good music was gone. No more than 5 seconds of testing.

That is merely an example of the majority of dacs' $5 oversampling electronics.  My now-years-old review of the initial Holo Spring dismissed this OS option in the first paragraghs.  However...upsampling and modulating via powerful sophisticated algorithms like HQPlayer, offsite (meaning cpu-based, then sent to the NOS setting on the dac) is a whole different kettle of fish.   But anyway,  I continue to stray OT.  My bad. 

[Downtheline]

1 hour ago, Peter Avgeris said:

Pink Faun adapter has one very serious disadvantage. I have tried more than 6 motherboards, modern ones, that all failed to perform under Euphony Linux.

Pink faun seems to have a new version of their I2S bridge, I recall linux compatibility being an issue previously. Might that have changed?

[AudioDoctor]

Guys, I went through this thread and I didnt see this answered here...

 

Can Euphony Stylus be installed on a Mac mini and will that mini retain its ability to use the ethernet port to output audio over the network and recognize an external drive with music.

 

edit: I understand MacOS would be completely erased. I am okay with that. It is a 2011 Mac mini that currently runs Roon Server and HQPlayer.

[lpost]

On 4/26/2020 at 1:53 PM, Downtheline said:

Pink faun seems to have a new version of their I2S bridge, I recall linux compatibility being an issue previously. Might that have changed?

It's Intel compatibility that's the issue and why they use AMD for their builds.  The I2S bridge on their website appears to the same as the one I bought and still use 2 years ago (in an AMD system).

[BCRich]

On 4/26/2020 at 5:02 PM, AudioDoctor said:

Guys, I went through this thread and I didnt see this answered here...

 

Can Euphony Stylus be installed on a Mac mini and will that mini retain its ability to use the ethernet port to output audio over the network and recognize an external drive with music.

 

edit: I understand MacOS would be completely erased. I am okay with that. It is a 2011 Mac mini that currently runs Roon Server and HQPlayer.

Hi @AudioDoctor,

Before I decided to buy the Euphony Summus Server I did a trial of Euphony on my 2012 Mac Mini. I did the USB stick and also did a install to my SSD Drive, I’m pretty sure it creates another partition on the Disc, I don’t think it overwrites the entire drive. I remember having some external hard drive issues and going into disc utility and seeing a small partition. The only thing is I did not remove the USB Stick at any point in the trial so I am not 100% sure. Hopefully some one else can chime in. You will have the ability to add music via a network share and I am using an OpticalRendu as my endpoint. I am sharing two FireWire Hard Drives on the Network to the Summus and also have a Lacie Thunderbolt Drive attached directly to it as well.

Have you trialed it yet? I love it.....

Good Luck and stay safe and healthy!

Mike

[AudioDoctor]

2 hours ago, BCRich said:

Hi @AudioDoctor,

Before I decided to buy the Euphony Summus Server I did a trial of Euphony on my 2012 Mac Mini. I did the USB stick and also did a install to my SSD Drive, I’m pretty sure it creates another partition on the Disc, I don’t think it overwrites the entire drive. I remember having some external hard drive issues and going into disc utility and seeing a small partition. The only thing is I did not remove the USB Stick at any point in the trial so I am not 100% sure. Hopefully some one else can chime in. You will have the ability to add music via a network share and I am using an OpticalRendu as my endpoint. I am sharing two FireWire Hard Drives on the Network to the Summus and also have a Lacie Thunderbolt Drive attached directly to it as well.

Have you trialed it yet? I love it.....

Good Luck and stay safe and healthy!

Mike

 

 

I have not trialed it yet, in the researching stage right now.

 

So you were able to boot into both MacOS and Euphony at different times, of course? My Mini is a 2011, so not that different from yours and you have given me hope.

[cat6man]

started discussing this in the generic super-thread and am moving the discussion here.

 

regarding audio-linux v. euphony

 

i see that many folks have moved from audio-linux to euphony over the past months with reports that euphony sounds better...............so i got a trial of euphony to check it out...........and the results are very interesting.

 

common baseline:

running HQPplayer embedded in server to NAA running on opticalRendu, much network optimization.

HQP running pcm only, no upscaling, no dither, just bit exact (theory being that having renderer at server and minimal processing at NAA is a good thing to do--i find my totaldac sounds best with no filtering).  music on HDD attached via usb to NUC running music server

 

audio-linux running in ramroot, realtime set to extreme [music on usb-HDD attached to NUC]

euphony running trial in ramroot (default expert settings) [music copied/cached on euphony ram drive]

 

note:  i am comparing the two only as music servers streaming to a networked NAA.  while i always suggest YMMV, this is even more true if you are running a single box PC incorporating both server/NAA.

 

tests-subjective impressions:

audio-linux 2.5 -- meh

euphony 20200114v3 -- meh

audio-linux 1.4 -- wow!

 

in my system, the old build of audio-linux is vastly superior to the newer builds of both audio-linux and euphony and it isn't close.  i have not chosen to compare the first two as who cares if mehA is superior to mehB when you have wow as an option.

 

i'd love to hear from other audio-linux and euphony users.  try comparing your current servers with the old audio-linux 1.4 and let us know what you hear

 

something is going on here but damned if i know what............scheduling/queuing algorithms in the realtime scheduler, arch-linux build variations/issues, ?????.....................but whatever it is, it isn't minor.

 

i can't listen to the first two for more than a few minutes before the difference is obvious and i'm back to the audio-linux 1.4 build.

 

 

[luisma]

16 hours ago, cat6man said:

something is going on here but damned if i know what............scheduling/queuing algorithms in the realtime scheduler, arch-linux build variations/issues, ?????.....................but whatever it is, it isn't minor.

Nice comparison thank you

I don't think you mentioned if you are doing any optimizations, which kernel on audiolinux you using 4.19? are you using realtime rtapp and rtirq processes?

thanks again

 

[cat6man]

56 minutes ago, luisma said:

Nice comparison thank you

I don't think you mentioned if you are doing any optimizations, which kernel on audiolinux you using 4.19? are you using realtime rtapp and rtirq processes?

thanks again

 

 

other than priority set to extreme, no other optimizations done. 

basic configuration shown in attached image.

if there are changes to rtapp and rtirq that help streaming to a NAA (versus direct usb to DAC), i would be very pleased if you could elaborate what you suggest might help, as i haven't quite figured out what to do there.

[cat6man]

other config/setup details:

NUC running music server using external usb HDD, both powered by hdplex200.

 

planned tests: 

1.  hdplex200 to be replaced by hdplex300 (end of week?), hdplex200 moves to basement to replace a bunch of wallwarts on my sagetv video server

2.  monoprice slimrun usb has arrived.  to be inserted between HDD and NUC and powered separately by hdplex

 

edit:  adding RTparameters info.  i assume that i want priority to be given to the ethernet connection(to stream audio to NAA) and to the USB (to get music from the HDD)..........should ethernet streaming to NAA be a higher priority than USB receiving music from HDD?  i assume that irq#126 is the usb.  which IRQ is for the ethernet that streams to the NAA? is it irq#127?

 

RTparams.txt

[cat6man]

On 7/7/2020 at 10:34 AM, luisma said:

Nice comparison thank you

I don't think you mentioned if you are doing any optimizations, which kernel on audiolinux you using 4.19? are you using realtime rtapp and rtirq processes?

thanks again

 

 

just found updated euphony trial download package...........i'll retest with latest euphony to see if any conclusions are changed

[beautiful music]

16 hours ago, cat6man said:

 

just found updated euphony trial download package...........i'll retest with latest euphony to see if any conclusions are changed

Looking forward for your experiment.

[cat6man]

does anyone know how to contact the euphony team?

i've tried messaging on their web site but it doesn't seem to be working.

[walloon]

12 hours ago, cat6man said:

does anyone know how to contact the euphony team?

i've tried messaging on their web site but it doesn't seem to be working.

 That's strange because they are usually very responsive. Alternatively, you can submit a ticket:

 

https://euphony-audio.com/hesk/

[cat6man]

8 hours ago, walloon said:

 That's strange because they are usually very responsive. Alternatively, you can submit a ticket:

 

https://euphony-audio.com/hesk/

 

thanks

 

nothing to do with their responsiveness, its a web issue...........the contact form on their website kept hanging when i submitted my enquiry.  i'll try again today.

 

update: yesterday it just hung.  today i get an error message saying to try again later.

[cat6man]

On 8/14/2020 at 10:27 AM, cat6man said:

 

thanks

 

nothing to do with their responsiveness, its a web issue...........the contact form on their website kept hanging when i submitted my enquiry.  i'll try again today.

 

update: yesterday it just hung.  today i get an error message saying to try again later.

 

web site contact working this afternoon

[cat6man]

ok now..........i have the trial euphony running stylus in ramroot and audio-linux running ramroot with the latest v2.6, and LMS/Squeezelite

 

both options use a HDD on usb off of a 7i7 NUC, slimrun USB from NUC to DAC, all powered by hdplex300

 

first a question:

i'm running the trial euphony, with ramroot enabled, but without the 'complete installation' (actually, i can't even find the complete installation button anymore).  will this have any impact on SQ?

 

assuming the answer to the above is no, i have done comparative listening on some of my favorite test tracks

Guantanamera (Weavers)

At the Break of Day (John Renbourn)

It's All Over Now Baby Blue (Joan Baez)

Angel from Montgomery (Bonnie Raitt)

Concerto for Cootie (Ellington)

Things Ain't What They Used to Be (Keith Jarrett trio at Blue Note)

West Side Story (Bernstein)

La Misa Criolla (Carreras)

 

The most obvious difference I hear is that audio-linux has a much better presentation of leading edges.

For example, with Renbourn's guitar I hear the fingers on the strings and the attack of each note, which are muffled by comparison with Euphony.  On the other hand, Euphony seems to present a little more space and a smoother, more contiguous soundstage, but at the expense of subduing the presence of each instrument/voice. With AL, I am more able to follow multiple instrumental lines simultaneously.

 

So, I see benefits to both but cannot give up the presence and increased detail that audio-linux gives me right now.

Any suggestions for parameter or configurations tweaks?  The loss of detail with euphony almost reminds me of the sound i would get when using HQP with various filters, upsamplilng and dither.  The soundstage would improve, but always at the expense of nuance/detail/presence such that I always preferred to configure HQP with no signal processing and ran bit exact into my R2R dac.  At this point, I find that the "NUC + slimrun USB to the DAC" solution beats out my previous setup of "NUC streaming to an opticalRendu (with highly tweaked networking)".

 

Next:  I'm going to try LMS on a separate PC with euphony and audiolinux operating just as an endpoint.

 

I've got 4-5 days left on my euphony trial and don't want to leave something important on the table that might impact my decision/conclusion.

 

Thanks

[austinpop]

Some Euphony tunings:

1. Do not constrain the BIOS. Leave it wide open. By this I mean, turn on Turbo boost, HT, and don't run in low power mode.
   1. Then, play with max frequency in Advanced settings, starting at the max turbo freq and stepping down in small steps until any trace of harshness is gone. 
   2. With a really good PSU, you may not need to do this. 
   3. This is a big contributor to transient accuracy.
2. Apply core isolation. I'm assuming you have HT on — if not, turn it on. Try something like this:
   1. 0-1 stylus 2-3 gstp 4-7
   2. If others have better suggestions for NUC7i7 (t's been a long time for me), please chime in.
3. Turn on caching (Use cache flag) and buffering (buffer before play 100%, and buffer album)
4. Before playback on Stylus (local music, not Tidal/Qobuz):
   1. Add tacks to queue
   2. Hit buffer queue to RAM button on the top
   3. Wait for all the track status buttons (little round circles with i) to turn blue.
5. Now hit play