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

[austinpop]

1 hour ago, Always.Learning said:

It seems the three of us are in agreement, then. Have others with similar systems weighed in? (It's been a while since I've tried to catch up on this thread.)

 

+1. If you hadn't already figured it out, I too preferred ISO-R before tX.

 

It's just like grammar - IR before tX except after sMS. OK - it's nothing like grammar.

 

1 minute ago, limniscate said:

If I'm remembering correctly, I like the ISO Regen before the tX, but @austinpop liked it better after.

 

Nope, same as you. 

[austinpop]

4 hours ago, Bamber said:

While testing the modded W4S DAC2V2SE DAC I am now able to set the Jitter Eliminator setting to level 1 and also disable it.  I couldn't lower this setting below the default level 3 before the mod without frequent dropouts.

 

Very interesting! I guess this is the DAC's way of confirming your input has very low jitter.

 

Curious what this does, but I'm not expert on DAC internals. 

[austinpop]

2 hours ago, Bamber said:

This setting controls the digital tracking for the ESS 9038PRO.   The lower the number the narrower the tracking window becomes and SQ can improve.

 

Previously if I lowered this setting to 2 while playing DSD128 or DSD256 tracks it resulted in dropouts.

 

So much to explore! Take your time @Bamber!

 

BTW - Does the DAC2v2 SE do native DSD or does it require DoP?

[austinpop]

Just a quick teaser on my next area of exploration.

 

I've had my eye on the Cybershaft reference clock, ever since @hols wrote about it. I also got @zephyr24069's help and advice. He put me in touch with Kenji at Cybershaft, from whom I learned that their entire production run was sold out, and there was no ETA on a new product with the quality of the OP-14.

 

Well, my luck has turned. I was able to locate an OP-14, and it has been shipped to me. My intent is to use this as a reference clock for my trifecta, feeding the single 50 ohm master clock input of my tX-USBultra.

 

Just to set the context, the OP-14 is a curated unit, where the OCXO has been measured to have a 1Hz phase noise of -114.0 to -114.9 dBc/Hz. The Ref 10, by contrast, is even better, with a claimed phase noise of ≤  -116 dBc/Hz at 1Hz. But at a third of the cost used, the OP-14 seems like a promising option.

 

Once I have my findings, I will report back. 

[austinpop]

32 minutes ago, atxkyle said:

Has anyone else using the microRendu experienced frequent network conflicts (manifesting in loss of playback from Roon until you reboot the mR) when using this bridged-connection scheme?  And figured out how to resolve it?  

 

I suspect it has to do with the microRendu using a dynamic connection - static IP not supported in their OS. I contacted Sonore - the response was that they don't/won't support this connection scheme.

 

I've only ever had borrowed mR's for a few weeks at a time, but I've never seen any network issues with bridging, using DHCP with the mR.

 

I don't think the problem and potential solution lie in static IP addressing.

 

If Jesus can't help you, then your last resort is a hail Mary. 

[austinpop]

1 minute ago, rickca said:

This is your best line in quite some time.  Bravo!

 

I'm on a roll these days.

 

Good to see you back Rick! Glad Irma took a detour to the west.

[austinpop]

4 minutes ago, Bamber said:

Well done sourcing the OP-14, perHabst another listening session with @limniscate can be scheduled?

 

Yes indeed! Waiting for his Paul Hynes SR7 to come in as well, and then we'll have a lot to dig into.

[austinpop]

Hey folks, 

 

By the (absolutely zero) authority vested in me by virtue of my OP status  I am going to propose something, with all due respect.

 

Can we conclude this discussion about timelines and delays on PH PSUs? I feel like both sides have aired their views quite coherently.

 

At this point, everyone can make their own choices. Beyond this, I feel any further discussion needs to be either via email/PM, or in another thread.

[austinpop]

Patrick,

 

I asked Paul about this a couple weeks ago. His best guess was end-Sept/early Oct.

 

Here is a bit more detail from his email. I hope he doesn't mind my posting it here!

 

"To answer your questions regarding the SR4,

1)    I revised the printed circuit board layout of the SR4 to reduce the chassis machining requirements, which caused a delay with the production boards. I am reluctant to take pre orders until I have verified the electrical and sonic performance of these revised boards and it will be approx 2 weeks before they arrive. Once I am happy with the performance I will begin producing them and taking orders. I have all the other materials in stock to produce the first batch so I expect this to be towards the end of September/early October if all goes according to plan.
2)    The overall specification will be the same as the SR7 rails with the exception of available power delivery and the regulator output within current ratings will be below 3 milliohms.
3)    All my power supplies are galvanically isolated from AC ground and the supply rails are completely floating with respect to chassis earth as this eliminates interaction when using multiple power supply rails in a sound system. The mains transformers are configured as balanced transformers with balanced Schottky rectification to provide a high degree of common mode isolation from mains even before the voltage regulator supply rejection is factored in. You should not have any problems with leakage current artefacts when using these power supplies."

[austinpop]

19 minutes ago, rickca said:

Can someone please explain what that means?

 

Rick, it refers to output impedance. See this post for more details:

https://www.computeraudiophile.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-the-sms-200-and-microrendu/?do=findComment&comment=676168

 

It is a reference to the spec for the SR7, which was:

Output impedance < 3 milliohms DC to 100 KHz

 

Low output impedance is a key measure of a PSU's quality, and 3 milliohm over that frequency range is rather impressive.

[austinpop]

25 minutes ago, auricgoldfinger said:

I would like to share my impressions from adding an isolating transformer with a floating secondary along with a DC voltage regulator/power supply to my system.  Thanks go to @Cornan for posting the concept and recommending the Gophert power supply (I am using a CPS-1610 16V 10A model) and @austinpop for recommending the BK Precision 1604A isolating transformer.

My objective was to improve the power supply to my Synology DS213+ NAS, where all of my music files reside.  Frankly, my expectations were low, but since the financial outlay was relatively small, I decided to give it a try.

I received the isolating transformer first.  There was a perceptible improvement, but I was not especially excited about the result.  However, after installing the Gophert, I heard a very dramatic improvement in sound quality.  Music became much more detailed and dynamic, images sharpened, and the soundstage deepened.  I heard the magic to which @Cornan has previously alluded.  (I am still only using the stock cable for the isolating transformer.  @Cornan says to expect further improvement from a star-quad ac power cable.)

 

 

I have now ordered a Tripp Lite IS500HG (with 4 outlets) and another Gophert for my router.  Before setting it up with the router, I plan to experiment using it to power my microRendu, ISO REGEN, and tX-USBultra.  The mR and IR are currently powered by separate LPS-1’s, and the tX-USBultra is powered by an sPS-500 with Pangea ac cable.  I am anticipating some interesting results.

 

Very cool!

 

Please keep us posted. I'm glad you Went-pher-it! 

[austinpop]

BTW @auricgoldfinger - what is the path from your NAS to your mR?

 

What components and cables?

[austinpop]

@romaz posted his latest update over on the Ref 10 thread:

 

 

Given how that thread went off the rails, I hope you come back and post here, Roy!

[austinpop]

On 9/16/2017 at 8:14 PM, romaz said:

Like many of you, I value simplicity whenever I can get it which is one of the reasons I went away from Ethernet endpoints like the sMS-200ultra.  As a minimum, I know I could be very happy using a single sCLK-EX board (and its 4 clocks) and forgo reclocking the modem/router/switch, Iso-Regen and tX-USBultra.  This means that with a single box server with integrated sCLK-EX board, a 4-rail SR7, and a REF10, I would feel I was in end-game territory but I will say that adding the reclocked modem/router/switch, ISO-Regen and tX-USBultra definitely improves things further.  

 

Hi Roy,

 

Fascinating report, as always. I did have a few questions - especially regarding reclocking the modem/router/switch. 

1. Where do the two sCLK-EX reside in your system now? One is in the tX-USBultra. But the other? In your server enclosure?
2. Is your ISP ingress close to your audio setup? My issue is that my ISP ingress is a 100 feet away in the opposite end of the house on a different floor. So modem/router clock mods would be tricky.
3. What would you recommend for a Roon server that didn't require a from-the-ground-up custom build, that could still be modded to have the system clock, USB and ethernet clocks delivered from an sCLK-EX board? And be powered by a 12V LPSU?

Appreciate your thoughts.

 

[austinpop]

1 hour ago, ElviaCaprice said:

 

Seeing that Roy (Romaz) has gone with a USB only stream and found it superior to any added Ethernet leg.  It's probably a good time to move on from this thread. 

 

58 minutes ago, mozes said:

+1 

I think it's a good idea if we move on to a new thread that focuses on optimizing USB audio. It is still relevant to those who use ethernet renderers as they still need usb to connect to a dac.

 

Guys, I personally don't see the need for a new thread, but if you would like to create one, you are always welcome.

 

42 minutes ago, romaz said:

Hi, sorry for the delay in posting.  It was always my intention to post my ultimate findings here and that will still happen but it will take a few days.  Despite how big this thread is, Rajiv has done an excellent job indexing it and curating it and so much thanks to him!

 

5 minutes ago, phusis said:

 

I find the Ethernet approach to have (long) reached its "saturation point" in regards to the growing slew of cables, clocks, PSU's and what not, almost all of them ever upgradable over short intervals. Even with the sonic refinements here made (in which this thread has been truly inspiring with all the contributions from its posters and the willingness in particularly from SOtM to participate on the basis of this with product development and individual wishes) the complexity and relative expense of many components can seem daunting and over the top (yet still cheaper than the all-out music servers like Aurender and others). Of course it's up to each one to assess when to say "enough is enough," and still be able to achieve great SQ through this solution with a simpler, less ambitious setup. Initially my interest with the Ethernet approach came with the advent of small network players (mR and sMS-200) and their simplified, dedicated audio-only computers and lesser power need, but it soon branched into a more sophisticated development with many areas to attend care to; in itself a fascinating process to follow (at least viewed from a distance). From my chair I'll stay with the current state of my Ethernet-based setup for a longer while (save for a few tweaks within it), simply for the reason being that it now sounds rather natural and uninhibited - while still being relatively simple component-wise - and that it has an easy, reliable and no-fuss use.

 

Thanks for all your thoughtful comments.

 

This space (of computer audio streaming optimization) is a big one, and can accommodate subscribers of both approaches - USB direct, or Ethernet endpoints. Roy's explorations have shown that both are rich troves of opportunity for optimization. I, for one, am not looking for "the winner" here. Most of us do not have the luxury of making wholesale changes on a dime. As @phusis so eloquently puts it:

 

10 minutes ago, phusis said:

Any development from this point in the future ahead would necessitate the outlook for more simplicity and an obvious advance in SQ - a combination of which may not be easily attained.

 

And finally, I strongly urge that we stay away from pitting manufacturers against each other on this thread. Those arguments can and often do easily derail threads here on CA. As OP-designate, I am going to be trying my best to dissuade any such ratholes we may be in danger of falling into.

[austinpop]

46 minutes ago, mozes said:

I am also fine with this thread, but I guess that our contributions lately extend beyond microrendu and SMS-200. Just a thought, maybe a small edit to the thread title to be more encompassing.

 

How about we change the title to "A novel way to massively improve the SQ of computer audio streaming"? Are people OK with that?

[austinpop]

2 hours ago, Lebouwsky said:

Will the old link still work?

 

Gulp!  Never thought of that.

 

But - I just checked the links in the index, and it looks like all is good.

[austinpop]

Listening Impression with the Cybershaft OP-14 Reference Clock - Part 1

 

As I mentioned a few days ago, I was very fortunate to recently acquire a Cybershaft OCXO Premium OP-14 reference clock. The current OCXO Premium line is sold out. A few weeks after my initial inquiry, Kenji @ Cybershaft informed me he had produced a unit that passed the OP-14 classification (more on this below), and would I be interested in buying it? 

 

Yes I would! 

 

Of course, my initial hope was to find a used unit - my budget is limited, after all. But this unit is rare to find used, and probably commands a resale price close to retail anyway. So I ordered it, along with their under-$20 0.5m 50 ohm clock cable.

 

Before I get into the listening impressions, a word about this device. Here is the web page: http://www.cybershaft.jp/a-products/premiumop.html. To view, allow Google to translate it for you. Some general comments:

• As I understand it, Cybershaft classifies units with OP-11 to OP-14 designations based on a process of curation of OCXO's by actual testing and calibration. The OP-14 has the lowest phase noise characteristics (Offset 1 Hz) of -114.0 to -114.9 dBc / Hz.  
• My unit came with its own calibration certificate showing it measured at -114.8 dBc / Hz at 1Hz, and -132.4 dBc / Hz at 10Hz.
• In comparison, the Mutec Ref 10's phase noise ratings are:
  • 1 Hz:               ≤  -116 dBc/Hz
  • 10 Hz:             ≤  -142 dBc/Hz
• The Mutec Ref 10 has much more impressive numbers. In addition, it has 8 galvanically isolated clock outputs, to the OP-14's single output. So these units address different needs, and the feature differences go a long way toward explaining the price difference. NOTE - I don't really have a deep knowledge of how these numbers relate to SQ, so please don't ask me! I am curious, of course.
• Still, for my trifecta, requiring a single reference clock input, the OP-14 seemed a perfect fit. 
• A note about availability. Kenji @ Cybeshaft told me they are very soon going to roll out a new product - the OCXO Platinum - which will have a phase noise rating of -109dBc/Hz. He would not say whether they would offer curated lower phase noise offerings like the OPxx. Regardless, if you are interested in these, do contact Kenji Hasegawa and let him know your interest.

 

But how does it sound? That comes next. I'll try to post that tonight so as not to keep y'all on tenterhooks for too long!

[austinpop]

Listening Impression with the Cybershaft OP-14 Reference Clock - Part 2

 

Since my clock chain is hard to visualize from what's in my signature, I've updated my system topology picture to give you a clearer idea of how things wire together. Hopefully, it helps understand the flow better. See the diagram below.

 

 

And here is a photo of the latest "spaghetti chain." I would characterize this more as a pasta medley - a hearty mix of linguine, pappardelle, and penne.

 

 

I've been doing a lot of listening since my Cybershaft arrived. However, I waited (im)patiently until about 50 hours of burn-in, and then again at 100 hours, to do any comparisons, or draw any conclusions.

 

Roy tells me that on his Ref 10, it is easy to turn each clock output on/off, which makes A/B comparisons of with/without reference clock really easy. On the OP-14, there is no switch. Also, it's important not to power cycle the unit, as it can take at least an hour for the OCXO to stabilize. The tX-USBultra continues to function if the reference clock is hot plugged in or out. My technique was to disconnect and reconnect on the OP-14 output, as it was easy to reach from my listening chair.

 

The first time I did this, I went from ref clock connected to disconnected. It took my brain a few seconds to register, but then it hit me. WOW - what a difference! It was like deflating a balloon. Some of the magic died. And yet - this was what, just a few days ago, I thought was the best my system had ever sounded. Connect the OP-14 back, and the goosebumps were back!

 

How to describe the improvement with the OP-14? I'll talk about it in the context of these pieces of music I happened to use while comparing:

• Mahler Symphony 3: Sehr langsam "O Mensch"
   
•  Sibelius Symphony 6: 1st movement
   
• Buena Vista Social Club: Pueblo Nuevo
   
• Mahler Symphony 5: Scherzo
   
•  Gorecki Symphony 3: 1st movement
   
• Beethoven Piano Concerto No.5 "Emperor": 1st movement

 

The improvement comes in these major areas:

1. Rock solid image
   • The sonic image seemed bigger in all 3 dimensions, but also had a sense of solidity and stability to it.
   • This was especially noticeable in the big orchestral Mahler and SIbelius pieces
2. Dynamic
   • It seemed like the music had a greater dynamic range.
   • There was a perceived reduction in the noise floor.
   • In the Mahler 3rd, the movement begins with an extremely low level harmony in the strings. I'd never heard this so clearly and so dramatically before.
3. Dimensionality of individual instruments
   • In Pueblo Nuevo, the piano sounded more real than before
   • Ditto in the Beethoven piano concerto
4. Incredible resolution and detail
   • What I love about the Mahler pieces in my list is how small groupings of instruments play with and against each other to really bring out the emotion. In sonic terms, it allows individual instruments to shine. With the OP-14, these instruments - like the french horn, the oboe, the strings had incredible texture and dimension.
5. Improved bass
   • The Gorecki starts with a minimalist melody in the double basses and cellos. With the OP-14, I really felt the growl in the pit of your stomach.
6. Smoother tonality
   • Overall, there was a further reduction in harshness, brightness, and fatigue. There was a richness, a smoothness, and an incredible sense of ease.

 

Just when I thought my trifecta couldn't sound any better, along comes another improvement that raises the bar further. I am just gobsmacked by the sheer magnitude of this improvement. And all this without any change to my DAC. Frankly, after the last round of improvements, I was pretty convinced that to achieve significant gains past this point, I'd have to upgrade my DAC first. But the Codex keeps scaling up to reach new heights. Pretty darn impressive.

 

And finally, a word about cost benefit. The OP-14 isn't cheap, so is the incremental benefit worth the cost? For me, the answer is yes, because of two reasons:

1. I find it difficult to go back to the prior configuration without feeling a sense of profound loss.
2. The SQ jump with the OP-14 is much greater than what I heard when I replaced the Codex with the DAVE in an earlier listening session. That price delta was HUGE, much higher than the price of the OP-14.

Clearly, in my system, the investments in the front end of the digital chain seem to pay bigger dividends in SQ than just in the DAC. Eventually, I'm sure everything matters, but at this point, I feel like the investment in the OP-14 is giving me excellent value.

 

As always - every system is different, so YMMV!

[austinpop]

3 hours ago, zephyr24069 said:

Great post with a lot of detail!  Your listed points parallel with what I've heard out of an OP14 level curated unit from Cybershaft with my Esoteric stack and are also what I'm hearing with my new Cybershaft OPxx curated unit that was custom-build with several features that may make it into other units in the future. The improvements are extremely noticeable and when I remove the clock from the circuit, my experience is the same, i.e. that I miss it dearly and the transformation low-phase noise OCXO 10 mHz clocking brings to my system.  These clocks effectively upgrade my transport-DAC combo "in place" and allow everything to achieve its full potential. As you say,...YMMV; this has been a very interesting trip to say the least!

 

Thanks, and that is good to hear you reinforce my impressions. I am also intrigued to hear of the OPxx unit you're using. You've clearly known this company a lot longer than I.

 

I didn't get any clarity from Kenji about whether the new Platinum line, which is not yet out, will also have a curated OPxx variant, like the Premium line did. It sounds like that is still going to be the approach? Or am I reading too much into this?

 

2 hours ago, mozes said:

Thanks for posting your impressions @austinpop it is really amazing that there is no limit to SQ improvements. I share your findings that improvements in the digital front end pay generous dividends without costing rediciulous $$$.

I am very intrigued by the upcoming SOTM master clock, I trust what these guys are doing. Most importantly it will be more affordable than the Ref 10.

 

Hi Moussa,

 

Yes this really sets the stage for the entrance of the SOtM clock. Will it have multiple outputs? What are the specs, in comparison to the Cybershaft and the Mutec? And most importantly, how does it compare sonically!

[austinpop]

3 minutes ago, ElviaCaprice said:

My biggest concern is the power supply ability of the master clock?  Hopefully the SOtM solution is adjustable to external power.

 

I hear you. In an ideal world, I'd like to see more manufacturers enable an external PSU, so the consumer can decide.

 

FWIW, Cybershaft does seem to have paid some attention to their internal LPS - quoting from their translated site:

• For the power supply section, use a special order large R core transformer with a minimum leakage flux. In addition, SiC Schottky barrier diode, large capacity low ESR capacitor, OSCON is adopted and ultra low noise linear power supply is realized. The occurrence of spurious caused by power supply noise is reduced as much as possible by phase noise verification.

[austinpop]

3 hours ago, auricgoldfinger said:

Keep an eye out for an official announcement from SOtM regarding their sCLK-OCX10 reference clock.  The specs should be released in the next several hours.

 

I've seen the specs. It looks very promising. I hope May releases the details soon.

[austinpop]

3 hours ago, auricgoldfinger said:

 

Here they are:  http://www.sotm-audio.com/sotmwp/english/portfolio-item/sclk-ocx10/

 

Seems like there are 3 options.

 

And here is the press release:

http://www.sotm-audio.com/sotmwp/english/press-release-rmaf-2017/

 

It looks like you can configure your choice of quality/performance upon purchase to Standard, Advanced, or Reference. And - it will have 4 outputs!

 

On first glance, the Reference phase noise values match or exceed the Ref 10, but as always, how it sounds is what matters.

 

I asked May about pricing, but she asked us to wait a bit longer as they were not ready to reveal that yet.

 

I'm really looking to the demo in RMAF.

[austinpop]

41 minutes ago, rickca said:

Any idea what these performance levels really mean? 

 

At a simplistic level: 

• lower number, good
• higher number, bad!  

But more seriously, let's overlay the numbers with the Cybershaft and the Ref 10.

 

 Typical Phase noise(dBC/Hz)

                                                                                           My

                                       sCLK-OCX10                       Cybershaft    Mutec
     Frequency  Standard  Advanced  Reference       OP-14         Ref 10 (published)
         10Hz  :       -125         -135        -140                   -132.4              ≤ -142
       100Hz  :       -150         -160        -160                   -141.9              ≤ --155
         1KHz  :       -160        -163        -165                                             ≤ --160
       10KHz  :       -165        -165        -165
      100KHz  :      -165        -165        -165
      Noise floor:                                                                                        ≤ -166

 

So, on paper these numbers looks very good. Let's hope the SQ of the unit matches up. Based on their track record with the sCLK-EX, I feel rather optimistic that it will.

 

Oh, and also - I'm very glad to see it has a 12V DC input, so you can use your favorite PSU with it.

[austinpop]

5 minutes ago, rickca said:

Sorry, what I meant to ask was what are the dBc/Hz for standard/advanced/reference.  Please update your post to say dBc/Hz rather than dB/Hz.  So you provided what I wanted ... where did you get that information?

 

Sorry about the dB instead of dBc. It was in the text from the SOtM product page, and I just cut and pasted. See: http://www.sotm-audio.com/sotmwp/english/portfolio-item/sclk-ocx10/ . Scroll down to Specifications.

 

The Cybershaft numbers are from the calibration sheet that came with my unit.

 

The Ref 10 numbers are from here: https://www.mutec-net.com/product_ref_10.php#data

 

5 minutes ago, rickca said:

The other thing I asked is whether SOtM has ever provided a similar phase noise graph for the sCLK-EX.  I'd love to see that.

 

They haven't, as far as I know. I would, too!

 

5 minutes ago, rickca said:

What exactly is different about a Reference vs Advanced vs Standard unit.  Different hardware?

 

I am guessing it's the same system hardware, but different OCXOs for each. But this is just a guess. Or - they may just curate, like Cybershaft.

 

Let's not speculate (he said, after doing just that).  We'll know soon enough.