A novel way to massively improve the SQ of computer audio streaming

[austinpop]

5 hours ago, R1200CL said:

Need some advice. I have 2 of LPS-1 and 2 of LPS-1.2. (The LPS-1 is not in use). 
 

Will adding a DXP-1A5DSC to either of the UpTone LPS give same result if I’m after 5V ?
This is the main question, as then I probably can achieve a lot using my present power supplies adding Alexey’s modules. 

 

I need a better way to power my opticalrendu than my present LPS-1.2. After upgrade to 2.8 I got problems. 

My first taught was the Farad supply, but maybe parallel two LPS-1.2 with the correct combination of @[email protected] modules will add the extra amp I’m after. I understand limit is 1.5A, so maybe still a bit low, but should be ok or ? (Just noticed Sonore PS output 1.5A, so I guess 1.5 is OK then). 

If so, what (boxed) items need to purchased ? And at what voltages ?

 

Another thought is using 4 LPS-1 in serial and parallel with Alexey’s boxes, if I’m able to get 2 more LPS-1 at a very reasonable price.

 

I have the etherRegen powered by Uptone stock SMPS, and an unused opticalModule.
 

The second LPS-1.2 is powering the 5V to my modded SU1. 
 

So good advices how to best put all this together at the most reasonable price (and best SQ) would be very helpful. 

I think my opticalrendu sounds a bit better at 9V than 7V. What’s other people’s experience ?

 

I suppose if Sonore sold their 7V supplies in 9V versions, then adding the DXP-1A5DSC would be equal to using Farad ?
That said, I think Sonore’s PS is equal to some very cheap Chinese PS. At least I’ve seen some that looks equal. If I’m able to find these Chinese PS again, and adding the DXP-1A5DSC, is this a way to go ?

 

Some key points:

1. the DXP-1A5DSC units can and do improve regulation, so can improve the SQ of an existing PSU. But the driving PSU's quality still matters. So, using a DXP-1A5DSC with an el cheapo Chinese LPS will not be the equal of the DXP-1A5DSC with a higher grade PSU. Sorry - I don't have direct experience comparing LPS-1 and LPS-1.2, both with 
2. The DXP-1A5DSC cannot help you with current capacity, and the innate ability of a PSU to deliver current. So with LPS-1 or 1.2, you're still limited by the 1.1 amp capacity. I would go further and say I prefer LPS-1.2 only when they are run below capacity, at < 1a.
3. The DXP-1A5DSC will limit you to 1.5A, irrespective of the current capacity of the PSU. You also need to be careful about the voltage drop from input to output - Alexey's chassis is rated to dissipate 5W max. I believe. Check with him before you buy.
4. you cannot run LPS-1 (or 1.2) in parallel. @Superdad has posted about this quite often.
5. I don't recommend running 2x LPS-1 in series. It can be done, and it works, but SQ suffers — presumably because the output impedance of both supplies get added in this configuration.

I would say get a Farad Super 3 or Paul Hynes SR-4T - either would make excellent PSUs for the oR. Then use your 2nd LPS-1.2 to power the ER.

[austinpop]

2 minutes ago, ASRMichael said:

Effectively this is what I will have.

 

Photo - The LPS in the middle

 

1st - SJ -  LPS DC3 5a (7 rails) (Mundorf options)

 

- ATX

- EtherRegen

- JCAT Ethernet card

- Pink Faun Ultra USB

- TLS NUC (which I use as a NAS)

 

2nd - SJ -  LPS DC3 10a (4 rails) (Mundorf options)

 

- CPU 10A

- Edge Router

- Sonore OM

- TLS NUC OCXO Board

 

11 rails in total

 

More detail on Topology PDF attached on how all setup. Shows Voltage and Amps

 

 

 

My Hifi.pdf 457.64 kB · 0 downloads

 

Nice!

 

I was actually more curious about the rear of the SJ PSUs. How is Sean is delivering ATX on the output of the DC3? Are 3.3, 5, and 12v on a single output, or separate outputs? If single - a single GX16-4 connector? Or a full 24-pin ATX out, like the HDPlex?

[austinpop]

6 minutes ago, ASRMichael said:

I’ve edited my post above with photo of the back

Perfect, thanks.
 

I love the look and lighting in your setup!

[austinpop]

9 minutes ago, auricgoldfinger said:

I recently added a modified Melco/Dela S100 switch to my system.  My experience follows.

 

 

Background

 

My first audiophile switch was a modified consumer-grade Linksys SE3008.  SOtM changed capacitors, replaced the switching regulator with a linear regulator, and enabled the switch to be re-clocked by the sCLK-EX board in my tX-USBultra.  Later, the sCLK-EX was re-clocked by their sCLK-OCX10 master clock.  As others have noted, the result was a nice improvement in sound quality.

 

At some point, we were told that a particular Netgear switch could shunt upstream noise to ground, so I added one before the modified SE3008.  Although I could never hear any difference, I left it in place.  Much later, I tried a Cisco Catalyst 2960.  I heard no change, so I removed it.

 

I purchased the SOTM sNH-10G audiophile switch a few months after it was released.  To be honest, I was put off by the price and was skeptical of its value.  I knew from prior experience with my tX-USBultra and sMS-200ultra that I could get much better sound quality using 20AWG solid core Neotech silver wire with JSSG 360 shielding rather than the generic 22AWG stranded silver offered as an upgrade by SOtM.  The Neotech upgrade increases openness, transparency, and detail.  Importantly, it also adds body to the sound signature.  Consequently, my sNH-10G does not suffer much from the thinness that others have noted.

 

Along the way, I acquired a Mutec REF 10 to re-clock the tX-USBultra and sNG-10G.  The REF 10 is a marked improvement over the sCLK-OCX10.

 

I purchased an EtherREGEN soon after its release to chain with the sNH-10G.  During this time, I removed the Netgear switch from my network and noticed an uptick in sound quality.  I ran this configuration for a few months before deciding to part with the EtherREGEN and try a re-clocked Melco/Dela S100 instead. 

 

The S100 modifications were made by Pink Faun.  They replaced the 25MHz oscillator on the S100 board with their Ultra OCXO clock module.  They also removed the internal power connectors and soldered those wires directly to the board.  Finally, and perhaps most importantly, they created a 5v input to SEPARATELY power the clock module.  In other words, the modified S100 requires two power supply rails:  a 12v rail for the switch board and a 5v rail for the clock module.

 

I would like to give credit to @seeteeyou for the idea of modifying an S100 and to @Nenon for giving me his thoughts and suggestions during my due diligence process.

 

Topology

 

A simplified representation of my system excluding grounding and power is as follows (apologies for the lack of an easy-to-read flow chart):

 

NUC7i7DNKE Euphony/Roon Server > Viablue EP-7 Cat 6a > SOtM sNH-10G SE

Synology DS218j > Viablue EP-7 Cat 6a > SOtM sNH-10G SE

Linksys WRT 1900 ACS router > Viablue EP-7 Cat 6a > SOtM sNH-10G SE

 

SOtM sNH-10G SE* > Single Mode Fiber > Modified Melco S100** > Synergistic Research Galileo UEF Ethernet > NUC7i7DNHE StylusEP > Habst USB Ultra III > tX-USBultra SE* > Synergistic Research Galileo UEF USB > Chord Hugo M-Scaler > Synergistic Research Galileo SX Digital BNC (2) > Chord DAVE > 12AWG 16-wire UniCrystal silver custom headphone cable > HE1000se headphones

 

* Re-clocked by Mutec REF 10 (Habst Digital BNC)

 

** Powered by a Farad3 12v and a JCAT Optimo Duo 5v rail.  A 100Mbps port is used to connect to the NUC7i7DNHE endpoint.

 

(Note:  A particular Finisar SFP recently has become popular for its added transparency and air.  In my system, I found the perceived improvement to be the result of a recessed midrange.  I went back and forth several times to confirm this impression.  In my system, for this particular application, the Planet Technology SFP is clearly superior.)

 

Listening Impressions

 

In a nutshell, the modified S100 has profoundly affected my listening experience.  I have implemented many changes during the past few years, upgrading sources, all types of cables, signal regeneration and re-clocking, power, grounding, and RAM.  In a few instances, a tweak or modification resulted in the perception of having acquired an upgraded version of a particular piece of gear.  Others were less dramatic but provided welcome improvements in sound quality.  None had the transformative effect of the modified S100.  Replacing my DNA Stratus headphone amp and Auralic VEGA DAC with the Chord DAVE was the most significant upgrade that I have made to my audio system.  The modified S100 is probably the second most important.

 

I am not a professional reviewer, and finding the words to properly describe what I am hearing has been challenging.  To begin, the background is very black.  Low level detail is jaw dropping.  I hear much more subtle elements in voices and instruments.  Reverberation and decay are more apparent.  The sound signature is smooth with an extended high frequency response that is not the least bit fatiguing.  Images and soundstage are more focused and coherent.  The acoustical space is more easily discerned.  Tonality, density, and resolution are outstanding.

 

I consider the sNH-10G to be my baseline switch.  Adding the ER modified the sound, primarily adding body while sacrificing some of the transparency, detail, and spatiality which are the hallmarks of the sNH-10G.  The net result was to tune my system rather than take a major step forward.  Replacing the ER with the modified S100 is an entirely different experience.  As I have attempted to describe above, the modified S100 elevates the sound quality to dramatic new heights.

 

 

Fantastic report! Your self-deprecating remarks notwithstanding, your listening impressions are wonderful, and really make me want to hear this mod.

 

It’s also interesting to read your remarks about the Finisar transceivers, as I am also finding that I like the Planet Tech ones better. To be fair, I’ve only had them in my system only a few days, so these are just preliminary.

[austinpop]

52 minutes ago, auricgoldfinger said:

 

The new server is essentially a replica of the first build in @Nenon's DIY thread.  The main difference is I will be using the HDPLEX 800W DC-ATX converter to power the ATX rather than separate PSU rails as he did.  I will have a 19v/10A SR7t rail powering the HDPLEX and a second 12v/12A SR7t rail directly powering the EPS.  I'm sure you are a generation or two ahead of me there.  😏

 

Yeah, that should be an outstanding-sounding server. As you know, I've been vocal in considering the PHD SR7TMR2 19v/10a+12v/12a  with HDPlex 800W DC-ATX as a particularly good PSU solution for a music server, but I'm curious to get first hand reports of it.

 

Remind me again if this build was with AMD 3700X or Intel i9-9900K? And I'm assuming JCAT USB XE and Net Card Femto?

[austinpop]

51 minutes ago, Peter Avgeris said:

FYI, I have been notified by the greatest network guru in Greece that the audiophiles *MUST* purchase two net switches and *NEVER* one.

 

LOL ... you are "preaching to the choir!" If you look back on this thread over the last 3+ years, you'll see this has been explored extensively.

[austinpop]

Dual-BNC HFC CT-2 Cables - My experience with the Schroeder method

 

Several people have noticed this in my latest system diagrams and asked, and I realized I've not really talked about my current dual-BNC High Fidelity Cables CT-2 cables between the M Scaler and DAVE in my system. Since Chord uses BNC ports on the HMS and the DAVE, you can use 75Ω BNC cables for this application. But which ones? And do the same cables that work well for 10MHz clocking also work well for this application? My experience says yes. 

 

For a while, I had been running Blaxius^2 cables between my HMS and TT2. When I upgraded to the DAVE late last year, I decided to revisit my choice of cables. I tried a pair of SOtM dCBL-BNC-75, and then the WAVE STORM Reference cables. Each was a significant step up in SQ, and I was on the verge of buying the STORMs.

 

Then I talked to Roy.

 

He and I had in the past talked about using the Schroeder method, but in the context of analog interconnects. You can read about this method in the link provided, and there is an active Audiogon thread about this as well. Very briefly, this method uses 2 physical interconnects in parallel, in conjunction with Y-connectors, for a logical interconnect. You can read the thread about reported SQ gains.

 

Roy told me he was getting outstanding SQ using a Schroeder configuration for S/PDIF based on HFC CT-2 cables. A little detail about what this entails:

• 4x 1.5m HFC CT-2 S/PDIF cables, RCA-RCA. Note, 2 of these together are used per connection.
• 4x 1BNC-2RCA adapters, sourced from Steve @ Audio Sensibility
• Furutech Nano contact fluid at the Y-connectors
• JSSG360 shielding on all 4 cables.

Here's what the resultant Schroeder-CT2 cable looks like. Note - the BNC's are hidden by the black protectors in the picture.

 

 

I was able to find used CT-2's at a significant discount, but even so, these end up being a fair bit more expensive than the Habst or WAVE STORM Reference. Is there a notable increase in SQ?  

 

Heck yeah! Prior to deploying these cables, I had just gone through the steps from Blaxius^2 to SOtM to WAVE STORM. The STORM already sets a high bar, providing a spacious dynamic sound stage, tremendous density and bass slam, and more focus and coherence than the lower-priced cables. Even as standard cables, the CT-2 go a step beyond the STORM in coherence and resolution, with a dead neutral tonality.

 

What the Schroeder-CT2 reveals is there is even more refinement and coherence to be had. As hard as it was to believe, the Schroeder-CT2's brought instruments into even clearer focus than the regular CT-2, with an utterly natural tonality. 

 

One question you may have, which I certainly did, is: does the use of parallel cables change the 75Ω characteristic impedance of the resultant pair? From my very, very rusty recollection of transmission lines from my EE undergraduate days, the characteristic impedance of a cable is related to the inductance and capacitance per unit length, and is engineered to the desired value by the separation between conductors, and the material and thickness of dielectric. Just to be sure, I reached out to Rick Schutz of HFC and asked him. He confirmed, saying:

 

...the characteristic impedance is still the same, the only way to change impedance characteristic is by altering the cable by spacing the ground closer of further from signal. I think most would argue that the length of the cable would be more relevant to reflections than the characteristic impedance as reflections are time domain.

 

So why am I talking about this? It's certainly not to suggest that people go out and buy expensive cables. I suspect the Schroeder method will give improvements across the price spectrum. Has anyone tried this for S/PDIF or other digital cables? If not, this is a call to experiment and report back. I'm curious to know what people find.

[austinpop]

31 minutes ago, beautiful music said:

Just out of curiosity Rajiv who is made the JSSG360 for you, is this Audio Sensibility or Rick

 

No, homemade.

 

31 minutes ago, beautiful music said:

also curious why you cover the connectors with black protectors.

 

Those were just for shipping. I took the picture with them on at the time.

[austinpop]

I knew you'd approve Larry.

 

Between this and Schrödinger's Cat, seems like there are some alternative approaches to try! 

[austinpop]

 

3 hours ago, Nenon said:

Ugh, absolute and generalizing statements using the word "ALWAYS" in caps in audio?

 

Thanks @Nenon for stepping in... I get tired of moderating this thread sometimes!

 

5 hours ago, Peter Avgeris said:

I would just like to add my two cents here...

 

Please do not consider my writings as an offense but as a starting point for further experimentation and critical thought. I need to double-underline this, as I have no intention to 'treat' one's mind or to persuade anybody in favor of my thoughts.

 

1 hour ago, sunny_time_99 said:

Thank you Peter,

I agree 100% with you. I've been doing this hobby for over 30 years and I've spent 25 of them swapping cables. For beeing my hobby, i build several hundred cabled during the last years. Connectors, Solder, Shielding, Material... i checked that all. Today I can say with a clear conscience that the big gamechanging differences do not come from the cables.

 

Hello Peter, Hello Sunny,

 

You will not find much traction in this thread by coming in and making sweeping generalizations. I urge you to read the posting guidelines, and the index in the first post of the thread.

 

Had you searched back, you would have found that many of us had also approached the idea of using silver in digital and DC cables with skepticism, having experienced the brightness from using analog cables. It was only from direct experience that we found what we did.

 

This is why sweeping generalizations are pointless here. This thread is for, and populated by, people who want to try things for themselves, and are open to hearing others' experiences. And this means specific experiments, not "over 30 years."

[austinpop]

1 hour ago, rickca said:

Yes, exactly.  And if one sounds better than the other, why?

 

I acknowledge @StreamFidelity's point that an i7-10700 may just be a lower binned i9-10900K that didn't test well enough to meet i9-10900K specs.

 

I spent nearly a year in Burlington, VT testing the PS3 chip at an IBM manufacturing facility.  The amount of test data collected about each chip is mind boggling, way beyond any published specs. 

 

Rick,

 

I too am having trouble discerning what your question is. Does this recent post of mine, detailing the differences I heard between i7-8700T and i7-8700K, help in any way? Or are you asking a different set of questions?

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/?do=findComment&comment=1044662

 

[austinpop]

42 minutes ago, ray-dude said:

Alas, tuning a mechanical system can be an iterative process.

 

You sayin' Daiz'an art to it?  

 

Doesn't Emile sell Daiza panzerholtz footers to enable rigid coupling between the component and the Daiza platform? Or did you find ceramic tiles to sound better?

[austinpop]

3 hours ago, lmitche said:

 

 

 

 

In my industry, Financial Services, we have Chief Compliance Officers that ensure that an institution has processes and procedures that "comply" to government regulations. These people are not flexible in any way. Strangely the term regulatory compliance is lately being replaced with regulatory adherence, which is closer to the meaning here.

 

Anyway, obviously I missed the "adherence" memo from Emile so thanks for filling me in. Based your input above, a maple block was placed between my DAC and the Daiza platform. Lastly the feet of my endpoint on the mid-sized platform were removed.  Weights are placed on both boxes. Coincidentally both boxes contain clocks.

 

The impact was immediate, more coherence across the spectrum, punchier bass, and the removal of congestion on some previously problematic test tracks.

 

The drums and bass in Patti Smith's new single "Jubilee" has real punch now.

 

On the last third of Cecile McLoren Salvant's, "What a Little Moonlight Can Do", on the Woman Child album, with the frenetic bass and drum play it can be difficult to separate the instruments. That is greatly improved now.

 

The last Third of Dixie Chicks new single "Julianna Calm Down" is much more transparent as well. It can be horribly congested when  things are setup wrong.

 

So many thanks to all of you for this Fathers Day gift. Enjoy the day.

 

Larry

 

OK one more Dad joke... where Daiza will, Daiza way!

 

Glad to hear about your good results, Larry!

[austinpop]

4 hours ago, rickca said:

Quite simply, other things being equal, would a 65W i7-10700 sound the same as a 125W i7-10700K or would one of them sound better?

 

I think the hardest thing to get right is sizing the PSU for both machines. This is complicated by the fact that the actual power demand is so dependent on energy required to:

• simply POST the machine, given current BIOS settings
• boot into the OS, given the energy management policy of the OS - e.g. frequency settings and governors in AS and Euphony
• run the music server workload. We're assuming vanishingly low CPU utilization for native sample-rate playback
• run other non-music workloads
  • system and application updates
  • library scans
• Other applications.

What my experience showed me is that the 8700K sounded better than the 8700T at the same operating frequency:

• IF the PSU is big enough to easily energize both CPUs for all the factors above, and
• when using the performance governor (default) in Euphony that sets and holds the CPU frequency to the max value specified in the settings.

Does this prove anything? Nope. And if asking on any other OS than Euphony, without knowing how the OS will control CPU frequency dynamically, I wouldn't be able to say. But certainly for Euphony, my inclination would be to go with the K CPU, if you're confident of your PSU's capacity.

[austinpop]

1 hour ago, BigAlMc said:

The Buffalo

 

Isn't the word "Buffalo" warning enough? 😎 

[austinpop]

Reminder from OP 

 

Hi folks,

 

Your friendly OP here. Several people have PMed me, and I see some posts here as well, that suggests the thread has gotten a bit off track. I will admit that I have been letting the discussion flow, because it has been civil, for which I thank everyone. But I do see now that we have gotten bogged down in a mix of generalities, Extreme-envy, Roon-bashing, and system-building DIY nitty-gritty.

 

Now, I know some will chuckle at the very idea that this massive meandering thread had a track to start with! 😎 But it does. And that is to keep postings focused on direct listening reports. While the scope of this thread is the entire digital streaming chain, let's get our focus back on listening reports, please. Detailed DIY discussion, especially around system builds make more sense in the excellent DIY build threads already out there, like @Nenon's https://audiophilestyle.com/forums/topic/58164-building-a-diy-music-server/ 

 

And as for how great the Extreme is — I think it's been said enough times! Further discussion needs to be in the context of listening comparisons, or results of experiments inspired by this fine server.

 

For participants who haven't already, please take a look at the top of the page at the posting guidelines, as well a link to the index, which will help you understand the history and spirit of the thread, as well as help you find already-discussed material.

 

[austinpop]

Removed a couple of OT posts. Please stay on topic. Thanks.

[austinpop]

On 6/30/2020 at 9:41 AM, flkin said:

I have a particularly good recording of the ubiquitous Jazz at the Pawnshop by Naxos and at the beginning of the “High Life” track around the 25 seconds mark, I’m able to clearly hear an order and the bartender’s reply for bottled water. This is with the entire tinned copper loom so it shows that details are resolved just fine. For a bit of fun, can anyone say which brand of water it is? 😄

I can hear the exchange clearly, but I don’t recognize the brand. Not surprising, as I’m not an expert on bottled waters available in Sweden in 1976! 
 

Pray tell. 

[austinpop]

26 minutes ago, HeeBroG said:

 

I believe it's a well renowned one beginning with E.

Have another listen and it may be clearer now🤓

Yes, I hear it now. 😉

[austinpop]

Experiments with Vibration...

 

...Control. Minds out of the gutter, please! 😎

 

I have not played with vibration control much in my system for a couple of reasons. First — due to my reviewing activities, things are always in a state of flux. Secondly, I've subconsciously dismissed this aspect as less important in a headphone system, even an admittedly high-end one like mine.

 

Well, Daiza time and place for everything, and all this Daiza talk piqued my curiosity, so it seemed like a good time to try something new. I wanted to do a proof of concept, so I did not want to spend $$$ on this experiment. I recalled the good results reported with IKEA bamboo butcher blocks as captured in this article: The Best Things in Life Are (Almost) Free: IKEA's Aptitlig Chopping Boards, so I tried this path. 

 

I bought 4 of these APTITLIG butcher blocks:

 

 

and for rigid footers, I found these hardwood birch 1" cubes on Amazon:

 

 

Pro tip regarding contactless shopping at IKEA in the US and Canada. Home delivery is backed up for weeks, but you can do contactless curbside "Click and Collect" at your local store. If — that is — you can find a slot! That's where this nifty site comes in: https://ikea-status.dong.st/. Yep, we may be the laughing stock of the world when it comes to tackling the pandemic, but we're still the best and brightest at shopping, baby.

 

My system is a hodge-podge spread over 2 vertical racks. I have my PSUs in one rack, and then my digital gear rests on a giant Tranquility Base UEF XL on top of another rack. I deployed the Ikea boards and birch footers under the following:

• power conditioner,
• 3 big PSUs.

    

 

My digital stack was unchanged, and looks like this. I have MiG 2.0 footers under my REF10 SE120. I can't get the DAVE to balance on MiG 2.0's, due to the weight of the cables, so I need to work on that. 

 

 

I realize I still have a ways to go to do this comprehensively across all my gear, but I thought this was a good start to evaluate if it made a difference.

 

Listening Results

 

For a total outlay of under USD 100, I'll say it makes a difference! It's not earth-shattering, but it is noticeable. Keep in mind that it took a good 10-15 minutes to put the blocks and footers in place, so rapid A-B was not possible. But what I am hearing is similar to what Larry @lmitche described a few pages ago with his Daiza platform. Transients are crisper, there's a general deblurring, and upper bass lines in particular are better defined and easier to follow.

 

While I'm sure something like a Daiza platform with Panzerholtz or Stillpoints footers will take this to the next level, I'm happy for now with this inexpensive hack!

 

[austinpop]

12 minutes ago, Confused said:

Are the LED’s on your REF10 too bright for you?

Yes, but I’ve dimmed all my LEDs with these stickers:

 

https://www.amazon.com/dp/B009WT7PYO

[austinpop]

56 minutes ago, BigAlMc said:

Where does it end....... 🤔

 

I know you were being mostly facetious, but it’s a great question. Trust me, I’ve asked myself that many times, and of late I’m convinced it’s “whenever you decide.” There IS no end, but we all can and should stop or pause and enjoy our systems as they are.

 

Regarding prioritizing where to apply the vibration control, I looked at what the goal was. We’re trying to shunt vibrations from the component to the platform via rigid couplers, and relying on the properties of the platform to dissipate this energy. So I attacked components as follows;

• Those with transformers, as these are generators of vibration. Hence PSUs
• those with clocks as these are very sensitive to vibrations. Hence master clocks and DAC.

That still leaves a bunch of spaghetti in my system unaddressed, so I’m in the same boat as you. Maybe I’ll go further... or not. TBD.

[austinpop]

Roon Update Build 571 on Euphony OS

 

As many of you know, my primary music player is Stylus on Euphony OS. I play music both from local files stored on a NAS, and from Qobuz. I do from time to time use Roon, especially when I am in discovery mode, as one of my favorite activities is grazing in the Qobuz new classical releases section, and Roon is so much better as a user experience.

 

Of late, I've been feeling that the SQ with Roon streaming Qobuz has sounded very satisfying, so it got me wondering if this was just a feeling or whether there was something going on. I did a quick check with this album, which is identical on Qobuz and my local copy on NAS:

 

In all cases, I was booted into Euphony OS, version 20200430, in RAM root mode. I compared the SQ of the following:

• Roon streaming Qobuz
• Roon from NAS
• Stylus (with cache, buffer queue to RAM) from Qobuz
• Stylus (with cache, buffer queue to RAM) from NAS.

To my surprise, Roon from Qobuz sounded significantly better than Roon from NAS! Overall, the SQ progression was:

• Roon from NAS < Roon from Qobuz < Stylus from Qobuz < Stylus from NAS.

For the math-challenged: "<" refers to "less than" or really, "worse sounding than."

 

Also interesting was that the difference in SQ between Roon from Qobuz and Stylus from Qobuz was not that big. So I was not imagining what I heard with Roon from Qobuz.

 

Analyzing Roon network behavior

 

To investigate further, I looked at the network behavior in the Roon cases, using the nifty network monitor in Euphony (click on the word REGISTERED on the bottom right of the web UI). Here's what I observed.

 

Roon build 571's network behavior is different based on source:

• When does preload start: when you hit play
• Qobuz/streaming sources: 
  • download current track at network speed
  • appears to preload enough of next track to maintain gapless, but full track download appears to only happen when the track becomes current. This is different from Stylus (below).
• Local network files from a NAS or SMB share:
  • no preload!!
  • Appears to be doing normal file reads, which results in some read-ahead, but not a full download.
• Local files on SSD
  • did not test, but not relevant to network traffic.

In contrast, let's look at Euphony/Stylus. The behavior here is also different based on source:

• When does preload start: 
  • Either when you hit play, OR
  • Click on "Buffer Queue to RAM." This allows you to wait for network activity to quiesce before you hit play.
• Tracks from Qobuz/streaming sources: 
  • download current and next track at network speed
  • when next track becomes current, download the following track at network speed, and so on
  • In other words, at most 2 tracks (current and next) are buffered
• Local network files from a NAS or SMB share:
  • download all tracks at network speed, up to limit of RAM buffer (tunable option).
• Local files on SSD
  • did not test, but not relevant to network traffic.

More experiments with Roon file playback

 

To see how much of the Roon file playback degradation was due to my files being on NAS, and the Core's not preloading the entire file like it was doing on Qobuz, I compared the same track, and varied the file location:

• original NAS location
• local SSD
• in RAM.

To achieve the above, I first fired up Stylus and cached and buffered the album. Then switching back to Roon, I added the following data locations in the Storage settings:

 

Cached files are stored on my boot Optane SSD in /data/Music/E_CACHE, and the files buffered in RAM actually live in /tmp/EUPHONY_BUFFERED. The reason that the /tmp files are considered in RAM is because /tmp is a filesystem of type tmpfs in Linux, which is a flavor of RAM disk.

 

Listening results

 

As expected, the SQ was worst from the NAS location, and best in RAM. However, what is interesting is I still felt the playback from Qobuz sounded better. So in SQ terms:

• Roon from NAS < Roon from local SSD < Roon from in RAM < Roon from Qobuz.

My best guess is that even in the "in RAM" case, the Core is still making filesystem calls to access the file, whereas in the Qobuz case, the entire file is cached within the Core process (note: this is a guess, not confirmed).

 

Relating this back to the Taiko Extreme Roon "degradation" experience

 

The above finding of disappointing results with Roon file playback at least loosely correlates now to the complaints of Roon SQ degradation reported on the Extreme forum.

 

Emile is claiming, and Extreme owners are reporting, that his latest tunings mitigate the problem, and that Roon now sounds better than ever. This to me suggests that there are OS tunings possible that can close this gap. I have some ideas, but this is one of the huge challenges with Roon.

[austinpop]

9 minutes ago, kennyb123 said:

Did you rip it from CD yourself?  If so you could eliminate source file differences by purchasing this album directly from Qobuz.  But - I doubt that's the issue and it has more to do with what you reported.  

No CD involved. This is a very recent 24/48 release. I bought the download files from the label. Can’t imagine they would be different on Qobuz, as this is too new to possibly have other masterings in the wild.

[austinpop]

9 hours ago, altsouza said:

I have noticed the same buffering from Tidal in the recent days. Do tou think this applies also to Tidal?

 

I don't use Tidal anymore, but I suspect the same applies.