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

[austinpop]

SOtM is installing their new superclock, the sCLK-EX, into it. I'm not yet sure what to expect except that May Park has suggested the improvement will be "huge." I will be their first guinea pig...we'll see. I'll report what I hear once I get it back. I'm sure Sonore has similar upgrades up their sleeves also. Competition is good and I love where this is all going!

 

Hi Roy,

Is this a one off, or do they plan to offer this as an optional add-on to the SMS-200?

 

I would be very interested to hear your observations of this change in terms of SQ?

 

What's the price, do you know?

[austinpop]

A question for those who made a network bridge under W2012R2.

 

I have a discussion with Audiophil, who states that in W2012R2 a network bridge is not possible in minimal server mode or core mode, only in GUI. I agree with Core-mode, but not with minimal server mode : I happily listen at the moment to minimal server mode optimized with AO's script and the network bridge I made in GUI mode is intact. Or are we talking about different types of network-bridge?

 

Who else got the network-bridge setup working in W2012R2 minimal server mode?

 

Hi Peter,

 

Did you make the bridge before applying AO?

 

Or is it still possible after a machine has been AO'd?

[austinpop]

I only make the bridge in GUI (didn't try to do it AO'd GUI) and then go to minimal server mode and use AO's script there. I think you cannot make a bridge while you are in minimal server mode.

 

You can connect the renderer directly to your PC in a separate sub-net as I described earlier and that can be done even in AO'd core mode. Potentially also in W2016-core. I will try that later. First I will exercise in W2012-core the necessary command line commands to configure my ethernet-ports.

 

Yeah sorry, I haven't tried AO yet, just gathering info before taking the plunge!

[austinpop]

Even ripping CDs with dBPoweramp while in minimal server mode sounds noticeably improved compared to the same CD rip while in GUI mode.

Whoa there! I know we're in less-understood territory, and I respect your insights, but this one crosses into the twilight zone for me!

 

Are you saying that files ripped running dBPoweramp in minimal server mode are forever imbued with an SQ improvement vs files ripped in the regular way?

 

How can that be?

[austinpop]

I suspect the benefits we hear are due to the elimination of unnecessary and redundant hardware of the switch, and it's power supply. A switch is just another computer with multiple NICs a processor, RAM and software. It makes sense that reducing the number of devices touching the music signal leads to improved SQ.

 

This does not mean there may not be specialized switch's optimized for music that will sound great.

 

Yes, I always feel a Pang of envy when I hear of @romaz 's fancy switches.

[austinpop]

It sounds nutty, but I think (as a few people have figured out), adding a wifi adapter to your network bridge could likely make this work for you.

 

I know of one CA member who was stuck at the exact same place that you are. Adding a tiny, cheap wifi dongle (this one: wifi dongle) made everything work (meaning, HQP was finally able to see his microRendu in NAA mode with a network bridge and direct connection). Even if this doesn't sound so appealing, it's probably worth the $8 expense just to try it.

 

I don't know this for sure, but I think (possibly) that some of the network bridges people are creating are, for some reason, not fully able to handle all the multicast routing that is required by HQPlayer. Whether this is due to their network adapters, their OS, software they have installed, I have no idea...

 

By the way, when I experimented a little (and tried a direct connection without creating a network bridge by 1) putting NIC 2 on a different subnet than my regular network and 2) running Open DHCP Server software to have it provide an IP address to the microRendu that was connected to NIC 2), I got stuck at the same place as you (I could ping the microRendu, could open the microRendu's web interface, could even play music to the microRendu in RoonReady mode, but could not get HQPlayer to see the microRendu in NAA mode).

 

Yup - that includes me.

 

Lots of issues with 2 x Ethernet bridge. Added a wifi adapter - technically, recreated by selecting 3 adapters (2 Ethernet, 1 wifi) and creating a bridge.

 

Flawless from there.

[austinpop]

In adding the wifi adapter are you enabling it entirely such that it is an active network device with a separate path to a wireless router?

 

Wifi adapter is enabled, but not connected.

 

So the only active path to the router is via one Ethernet interface only.

 

W10 Enterprise.

[austinpop]

It's interesting how these little improvements can add up to something quite significant and very meaningful.

 

Some have suggested there's no reason that two music players that have bit-perfect output should sound any different and these filters definitely counter that theory. These AO sound signature/filters are not in the music path at all. According to AudioPhil, they merely affect CPU/memory handling and yet from 1A to 4D, you go from a more precise sound to a warmer and lusher sound and with this direct connection, in my system, there is no mistaking the difference.

 

The common strategy that AudioPhil had always suggested was to use setting 4D for the Audio PC (renderer) as it resulted in a smoother and "tube-like" presentation and to consider something like setting 1B for the control PC as this resulted in a more precise and detailed presentation. I believe these are his personal preferences for a dual PC setup. Well, we have no option to impact the sound signature of either the mR or sMS-200 but what I have found in my own comparisons is that the mR has greater body and maybe a touch more organic bloom to it while the sMS-200 has greater detail clarity and precision. Which is better is up to the user but as for my personal preferences, the sMS-200 has a slight edge since it just sounds more resolving. With my mR, I first went with 1A on my Mac Mini and detail clarity with the the mR went up considerably which I found very much to my liking. In fact, it felt as if my mR had become transformed into the sMS-200 as far as detail clarity but 1A resulted in a fairly flat and almost mechanical presentation. 1B improved soundstage depth and fluidity but 1C was my preferred sweet spot in the end. Still plenty of detail clarity but much more depth. These filters alone are worth the price of admission of AO for me in this setup.

 

My sMS-200 is still out but upon its return, I will be surprised if I prefer the same signature and filter settings.

 

I have now proven to myself through blind testing that Process Lasso adds a small but still meaningful improvement, especially when Roon is run in Bitsum Highest Performance mode. I am not yet convinced Fidelizer Pro is adding anything further.

 

Interesting comments!

 

I've been evaluating both the mR and the sMS-200 over on this thread: http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/auralic-aries-mini-vs-sonore-microrendu-vs-soul-music-sms-200-listening-impressions-31499/index2.html

 

I just installed AO on my bridged machine last night, although I am on W10 Enterprise. I have to say, I'm not finding a night and day difference with AO, but it's early days, so extended listening is needed.

 

At least initially, I found the 1-4 and A-D filter effects to be very subtle. I currently have the sMS-200 in the system, and settled on 3C as my preferred setting.

 

I've just gone back to no FMCs on the direct link, and will see how that sounds. More to come.

[austinpop]

BTW - not to digress too far on AO - but does anyone know what the typical process count should be after AO is applied?

 

In my case, on W10 Ent, with no apps running, I have about 27-29 each of background and Windows processes.

[austinpop]

Which thread Larry?

 

Not at my desk, so slave to the limited search on the iPhone app!

 

 

Sent from my iPhone using Computer Audiophile

[austinpop]

It's interesting how these little improvements can add up to something quite significant and very meaningful.

 

Some have suggested there's no reason that two music players that have bit-perfect output should sound any different and these filters definitely counter that theory. These AO sound signature/filters are not in the music path at all. According to AudioPhil, they merely affect CPU/memory handling and yet from 1A to 4D, you go from a more precise sound to a warmer and lusher sound and with this direct connection, in my system, there is no mistaking the difference.

 

The common strategy that AudioPhil had always suggested was to use setting 4D for the Audio PC (renderer) as it resulted in a smoother and "tube-like" presentation and to consider something like setting 1B for the control PC as this resulted in a more precise and detailed presentation. I believe these are his personal preferences for a dual PC setup. Well, we have no option to impact the sound signature of either the mR or sMS-200 but what I have found in my own comparisons is that the mR has greater body and maybe a touch more organic bloom to it while the sMS-200 has greater detail clarity and precision. Which is better is up to the user but as for my personal preferences, the sMS-200 has a slight edge since it just sounds more resolving. With my mR, I first went with 1A on my Mac Mini and detail clarity with the the mR went up considerably which I found very much to my liking. In fact, it felt as if my mR had become transformed into the sMS-200 as far as detail clarity but 1A resulted in a fairly flat and almost mechanical presentation. 1B improved soundstage depth and fluidity but 1C was my preferred sweet spot in the end. Still plenty of detail clarity but much more depth. These filters alone are worth the price of admission of AO for me in this setup.

 

My sMS-200 is still out but upon its return, I will be surprised if I prefer the same signature and filter settings.

 

I have now proven to myself through blind testing that Process Lasso adds a small but still meaningful improvement, especially when Roon is run in Bitsum Highest Performance mode. I am not yet convinced Fidelizer Pro is adding anything further.

 

Once again, thanks for your insights into AO. I've just tried it and written up my experience here - http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/auralic-aries-mini-vs-sonore-microrendu-vs-soul-music-sms-200-listening-impressions-31499/index2.html#post634586 . I applied AO and some additional W10 tweaks.

 

Sure enough, I did find that the AO filter settings can tune the sound signature of the 3 endpoints in my comparison. In my case, my preferences were:

• sMS-200 - 3C
  
• mR - 2C
  
• Aries Mini - 2B
  

 

Another important finding for me was that, finally, with AO and W10 tweaks, I am hearing better SQ without FMCs in the direct connection to the bridge! This was one of the differences I had in my findings with you guys in the past. Now I hear it clearly.

 

I haven't yet tried Process Lasso, but I did manually set MinimServer (or technically javaw.exe) to Realtime priority. Is there anything else PL does other than enabling persistent setting of priorities?

[austinpop]

Thanks for your post. It seems the evidence is mounting that these FMCs in the direct path can cause more harm than good. These FMCs have switching voltage regulators. I suspect they also have inferior clocks and so the good they might do with regards to blocking RF noise could be offset by these other problems.

 

Yes, I agree.

 

What I found was that the direct path sounds better - overall - but it's not cut and dried. The FMCs do seem to grow the soundstage, but I also noticed more glare and harshness.

 

So I agree - there are likely competing forces at play with FMCs in the path, and depending on the rest of the chain, the net result could either tilt in favor of FMCs or without.

 

I really wonder what something like an Etalon isolator would do instead. Anyone want to send one along for testing?

[austinpop]

Yeah it's rare here in the US, but has gotten mentioned as being superior to the more commonly mentioned EmoSystems EN-70HD.

 

 

Sent from my iPhone using Computer Audiophile

[austinpop]

For those who have a nas Synology with a double nic you can run a bridge with a root command. I have tested on a* Synology ds214 + and it works with Sotm sms 200 and minimserver or asset upnp .

 

Open ssh option in nas configuration

*With putty ,connecting to your nas with ip of your nas .**

 

Sudo su

( password is the same as you synology.)

 

insmod /lib/modules/stp.ko

insmod /lib/modules/bridge.ko

brctl addbr br0

brctl stp br0 off

ifconfig br0 (ip of your nas) netmask 255.255.255.0 up

brctl addif br0 eth0

brctl addif br0 eth1

ifconfig eth0 0.0.0.0 promisc up

ifconfig eth1 0.0.0.0 promisc up

route add default gw (ip of you router) dev br0

exit

 

Kudos!

 

If you've read this thread, you know I beat my head on this for several days - look at http://www.computeraudiophile.com/f10-music-servers/novel-way-massively-improve-sq-sms-200-and-microrendu-31110/index6.html#post619478

 

I think my script was very similar, except in mine I brought up eth0 and eth1 and added them to the bridge before bringing up the bridge. But hey, whatever works!

 

In my case, I moved on and decided to run the bridge on Windows. Which really opened Pandora's box, because I then discovered that running MinimServer on the W10 bridge box sounded better than running Minim on the NAS. Go figure.

 

Anyway, great job cracking this - I'm sure it'll benefit many people.

[austinpop]

What they are creating for me is a "one off" but they plan to offer this super clock as an optional add-on for future sMS-200 customers. I'm guessing there will be an option for current owners to send in their units for an upgrade but it will require a new larger chassis and a slight reworking of the internals. The price with the new clock + sMS-200 has not been established. I believe they are having some problems sourcing components (the proper chassis) and so their timeline has been delayed but April is what I am hearing.

 

Just how good this clock is, I'm not sure. Chan King Girand Michel, owner and maker of Pachanko cables based in France, who I have used to make me UP-OCC grade SATA cables (highly recommended, btw) uses this new clock now and he said it makes TCXO and OCXO like "child's toys" which is a pretty bold statement. SOtM is telling me this is the best clock they have ever made and considering how good the clock in the sMS-200 is, I felt it was worth the risk to try.

 

My situation is unique and will be different from what SOtM will be offering but there's no reason you couldn't approach SOtM with your own custom requests which is what I have done.

 

In my correspondence with May Park, she stated that she has a prototype sMS-1000SQ with the new clock at the USB output and that the improvement was significant. She decided to add their new tX-USBUltra (expected release date in April) to the chain which is essentially a USB relocker using this same super clock. This USB-regenerator with superclock sits between their music server and their DAC just like a USB Regen would and since both the server and USB-regenerator use the exact same clock, she wasn't expecting much improvement (if any) but was quite surprised by just how much more detailed and dynamic the SQ had become. In effect, she is hearing the same thing that I am with my reclocking switch in my direct connection path.

 

So in my situation, since their super clock can reclock up to 4 components (for $1150), I used 2 clocks for the sMS-200, 1 clock for a LAN switch and 1 clock for their USB-to-SPDIF converter. While my Chord DAVE sounds best via USB compared to SDPIF, I will see what triple reclocking will offer. I have no reason to doubt that this clock will improve the sMS-200 and the switch using this direct connection but it remains to be seen if going SPDIF and triple reclocking will yield any further improvement in my situation. There is the possibility it could be a mixed bag and somehow negatively impact the wonderful balance I have achieved with this direction connection. I am in uncharted territory for sure.

Hi @romaz

 

Any update on this? Did you receive your sCLK-EX, and if so, how does it sound?

 

I did contact May, and she indicated they were working on an sMS-200Ultra, which is the integrated unit, and the timeline is about 2 months, which is consistent with April.

 

I am wondering whether to wait for this to buy the sMS-200. In my recent evaluation, I paired it with a chain of Intona and RUR on USB, and that gave a significant SQ bump over the direct USB connection to the DAC (Codex). This Ultra config would make sense if it eliminates the need for external reclockers like the RUR.

 

But galvanic isolation is still a question. I suspect your DAVE has it's own galvanic isolation as to render something like the Intona unnecessary. But with the reclocker built into the streamer, I suspect adding an Intona, while adding GI, will negate the benefit of the sCLK.

 

Interestingly, I found even the mR benefits from the Intona/RUR combo in the USB chain, even though it has a Regen built in.

 

This may be off tangent to the direct Ethernet discussion. Sorry.

[austinpop]

Yes, busy weekend. My SOtM bundle arrived on Friday and I had some initial problems with powering them but they have been sorted out.

 

As before, what spurred this whole venture was my unexpected observations with my Paul Pang switch with TCXO clock. I bought this relocking switch 2nd hand on a whim because I was curious to know what reclocking an ethernet stream might sound like just prior to the stream entering either my music server (Mac Mini) or NAA (sMS-200 or mR). With my NAA in the standard configuration (connected directly to a router), this reclocking switch made a small difference but barely worth the $100 I spent for it even when powered by an LPS-1. With either NAA directly connected to my Mac Mini and with this reclocking switch in the "direct" path, however, I was quite surprised by the significant improvement in terms of soundstage and clarity.

 

Obviously, while TCXO clocks are better then the standard clocks that come in all-purpose network switches, Paul Pang sells an even better OCXO version with 10x better stability than his TCXO switch and so my attention was drawn there. About this time, I had taken delivery of an "UP-OCC grade" SATA cable for my music SSD that was hand made for me by Chan King Girand Michel, owner of Pachanko Cables based in France and he told me how his new SOtM sCLK-EX "super clock" made his Paul Pang V4 USB card with OCXO clock like a "child's toy."

 

Based on conversations with May Park of SOtM, it appears they were quite surprised also by how much better their new super clock, the sCLK-EX, was sounding compared to their previous best clock. Unfortunately, because of the larger size of this clock board, they were unable to retrofit it into the sMS-200's existing case. It turns out, however, that there was enough space in their USB-to-SPDIF converter, the tX-USB HD, to accommodate this clock and so this is where things got interesting.

 

For those of you that know the Chord DAVE, you know that USB is its best input. It is the only input that is synchronously tied to DAVE's clock whereas the other inputs have to go through DPLL first. Furthermore, DAVE's USB input is a "floating" USB design and it's galvanic isolation is excellent. As previously stated, I heard no improvement whatsoever with an Intona Industrial. Furthermore, the RF isolation techniques that others were raving about (ie FMCs, isolation transformers) hardly resulted in any improvement (if any at all). If there was any leakage current that was being generated by my Paul Hynes SR7, it was resulting in no sonic penalty that I could detect as my SR7 sounded noticeably better than my highly regarded LPS-1.

 

Just to be sure, I had an Aurender W20 in house a while back and directly compared the W20's USB output against its SPDIF and AES/EBU outputs to DAVE. Despite using a set of highly regarded High Fidelity Cables SPDIF ($4,500) and AES ($7k) cables that I had on loan and despite Aurender's claim that their AES output sounds best on the W20, USB still won out. So the prospect of buying this USB-to-SPDIF converter from SOtM (even with its super clock) didn't sound like money well spent.

 

What made it interesting was the sCLK-EX has the option of 4 independent outputs and SOtM's USB-to-SPDIF converter only required one of these outputs and so I approached SOtM with the idea of using the other 3 outputs for other components. As it turns out the sMS-200 has 2 clocks which left me with one available clock and so I elected to use this for my cheap $20 Trend Net 5 port switch I had lying around. This switch was convenient because (1) I already owned it and (2) I could power it with my 5V LPS-1.

 

Here are a few photos of the triple combo:

 

[ATTACH=CONFIG]33396[/ATTACH]

 

[ATTACH=CONFIG]33397[/ATTACH]

 

Here is the tX-USB HD opened up and you will see the clock board as well as the USB to SPDIF converter board in the same chassis. Out of the clock board, you can see 4 black clock cables (30cm each) leaving it for their desired destinations (1 for the USB to SPDIF converter, 2 for the sMS-200 and 1 for the switch):

 

[ATTACH=CONFIG]33398[/ATTACH]

 

Here is a photo of my switch opened up. Along with the clock input, you can see a large capacitor that SOtM added for me. This switch happened to have 2 switching regulators. They were able to replace the larger one with a low noise linear regulator. They had no substitute for the smaller switching regulator and so that one was left in place but "cleaned up." The cost of the capacitor and linear regulator upgrade was $70. Unfortunately, this switch has a power saving feature built in (it turns out all modern 5 port unmanaged switches do) and so I had to accept that this switch has compromises.

 

[ATTACH=CONFIG]33399[/ATTACH]

 

Here is a photo of this layout along with my 2 Mac Mini's that I am using for testing. Both Mac Mini's incorporate Uptone's MMK allowing either to be powered by my Paul Hynes SR7. My SR7 is also powering my tX-USB HD. For the time being, my LPS-1 is powering my sMS-200 and switch but I have another Paul Hynes SR7 on order.

 

[ATTACH=CONFIG]33401[/ATTACH]

 

Before I move on to SQ, during a previous post, I had postulated that for reclocking to result in an improvement in SQ, the component that receives this nicely reclocked signal needs to have clock(s) that are at least as good or better. Whether this is actually true or not, I can't be sure but it makes sense to me and it is how I reasoned that my PPA switch would sound noticeably better when used before either of my NAAs compared to before my Mac Mini and it's "not so great" stock clocks. In fact, I surmise that this could also represent one important reason why the mR and sMS-200 were previously much more incapable of revealing the qualities of a well-tuned upstream source because with the standard way of directly connecting your mR/sMS-200 to a router, the wonderful signal that your finely tuned music server generates ends up getting molested by all the bad clocks in your network path (routers, switches, FMCs, etc). The network path is not such a benign path after all.

 

So how does triple clocking sound using clocks of identical quality? I'll start with the upgraded sMS-200 (soon to be called the sMS-200 Ultra). With the stock sMS-200 vs the mR and with each powered by a switching 9V iFi PSU, I actually prefer the mR. The sMS-200 sounds more detailed but the presentation is quite thin and anemic whereas the mR has nicer body. When powered by the LPS-1, this thinness improves considerably and while both the mR and sMS-200 benefit greatly from a superior low-impedance PSU like the LPS-1, the sMS-200 scales better to my ears -- it is more resolute. Powered by the Paul Hynes SR7, the gap grows further in favor of the sMS-200 although this gap isn't necessarily enormous and when the 2 were A/B'd amongst a group of friends (our local audiophile society), some of us favored the greater detail resolution of the sMS-200 and others favored the slightly richer, full-bodied sound of the mR although within our group, the ratio is 4:1 in favor of the sMS-200.

 

With the upgraded clock, with only 2 days of listening thus far and with this clock probably requiring further burn-in, there is no longer any debate about which sounds better. The level of clarity, the layering of the finest details, the precision of timing, the accuracy of timbre and the soundstage are now at a much higher level. I have never heard the timbre of the piano sound this real in my system before...ever. There is still that characteristic thinness that I equate as SOtM's house sound when compared against the mR but I now equate this fullness in the mR's presentations as a coalescing of detail. Where you have a hundred violins sounding like 20 violins with the mR, with the sMS-200, you get a better sense that a hundred violins are actually playing and so with this new clock, there is not only better timbre but also better timbre variation. Where sMS-200 and mR were previously close, with this new clock, the sMS-200 has moved to a whole 'nother zip code.

 

Moving on to the switch, as previously stated, I looked at this as a freebie since I had a free clock output. Given that this switch was a true "cheapie" with certain undesirable attributes (power saving feature), I prepared myself that the improvement could be only a bit better than my Paul Pang switch. I was wrong. The improvement is huge. This switch nearly doubles the improvement I am hearing with the improved clock in the sMS-200. I am just flabbergasted!

 

Moving on to the USB-to-SPDIF adapter, as before, because I know my DAVE sounds better via USB compared to SPDIF and because previous experiments with other USB-to-SPDIF adapters including an Audiphilleo 1SE custom configured to be powered by an LPS-1 failed to sound better than straight USB, I was prepared for this to not result in any SQ improvement and possibly sound worse. Once again, I am quite surprised that there is further improvement. While the improvement is not as large as what I am hearing with the sMS-200 or the switch, it is a notable and welcome improvement, nonetheless. How best to describe this triple reclocking combo? Just utterly mesmerizing with regards to realism.

 

Of course, this means it is now game on for my music server. I don't know if it will be diminishing returns from here on but I have already begun a conversation with SOtM about sending them a specific Gigabyte motherboard with dual LAN ports. This mobo has 5 clocks including from what I can gather independent 25MHz clocks for each LAN port, a 24MHz system clock and a 25MHz PCIE clock and I am aiming to replace 4 of these clocks. SOtM has promised to send me their phase noise and stability measurements for their new superclock but regardless, my ears have told me all that I need to know -- this is one incredible clock that could possibly bring about a new revolution to music servers.

Wow, excellent findings.

 

Would you say the sCLK removes the value of USB reclockers between the sMS-200 and the DAC?

 

I'd like to make sure I understand the new chain. Is this the best case config described above? I've shown clock inputs in the component chain, and the PSU below each component.

 

Component ---- Mac Mini (bridged) > Trendnet switch (sCLK-1) > sMS-200 (sCLK-2,3) > tX-USB-HD (sCLK-4) > DAVE

PSU ---------------------(SR7) ------------------- (LPS-1) --------------------- (LPS-1) ---------------- (SR7)

 

Given this, I will definitely wait for the sMS-200Ultra before I buy!

[austinpop]

So Romaz given the great bridging tweak you have given us I just now have to know more about this switch. What is it, where did you get etc etc? Do you believe that the doubling of sq impact requires the super clock SMS -200 downstream of it to yield such results? I find this so fascinating and I smell a new project coming up for me.

 

Indeed - I agree this seems like a new project in the air!

 

@romaz - it seems like what you've constructed with your modified Trendnet switch, precise sCLK, and LPS-1, is a Regen/RUR for Ethernet! Would you agree? Until now, the focus on the Ethernet side has been isolation - either passive (EMO, Etalon) or active (FMCs), but not so much on the clocking.

 

The question now becomes this. With your Ethernet reclocking switch in place, do you still need bridging on your music server? I.e. could you revert to this configuration, without any loss of quality:

 

Music server -----------> Trendnet with precise clock -----------> sMS-200Ultra

Router _______________________|

 

Or do you still find this to have better SQ:

 

Router <------> Music server (bridged) -----------> Trendnet with precise clock -----------> sMS-200Ultra

 

Thanks again for blazing a new trail here!

[austinpop]

I wish that this was true and if it were, then replacing the clock in the sMS-200 should have sufficed but in my system, with each clock upgrade, the improvements continued.

 

In my simplistic way of looking at this, there is no clock that is perfect as even the finest atomic clocks will have some level of phase noise (clock jitter) and instability over time and so the best that a clock can do is to not cause harm to a signal but in reality, all clocks will degrade a signal.

 

Where a really good clock seems to make a difference is in recovering some of the damage made by a bad clock that preceded it. Can it completely undo damage by a bad clock (or a series of bad clocks)? My guess is no but that with repeated reclockings, it would appear that each reclocking can further refine or restore the signal and there are many posts here on CA of how people have reported improvements by placing several USB Regens in series. I used to own a TotalDac d1-monobloc DAC and this DAC includes a very good reclocker and this DAC seemed to benefit as well from having several of these reclockers in series. As such, TotalDac's best "twelve" DAC actually has 2 reclockers.

 

What I am thinking is that it would be best to have as few bad clocks as possible in your chain making it less important to have so many great clocks at the end of the chain to rescue the signal timing. Lastly, I am guessing that if you have an entry level DAC with a mediocre clock, all of these efforts may not make as much of a difference because it should be the very last clock that matters the most. When you look at your digital front end, I believe strongly that the DAC is the most important piece but your DAC can only be as good as the quality of the signal it is fed.

 

@romaz - I agree with you about the fact that it is not sufficient to simply add reclocking (and for that matter, galvanincally isolation, and high-quality LPS) to just the component upstream of the DAC. This is where I started, since it seems to be a logical premise. But like you, empirically, I have found that improvements continue as you move further upstream. I'll be honest - I wish it were not true!

 

Second, the idea of high-quality clocks on streamers is not new. The Auralic Aries Femto also claims to use high-precision FemtoClock in their streamer. It would be interesting to see how the sMS-200 Ultra and the Aries Femto stack up. Do you know what the metric of goodness is for external clocks like the sCLK-EX? You made an important point that the benefit of reclocking is ultimately gated by the quality of the DAC's clock itself. My DAC, the Ayre Codex, has scaled very well so far with the addition of reclockers and isolators, but I have seen no published specs regarding the accuracy of its internal clocks. So I suspect, it may be hard to find these metrics for DACs.

 

Now, regarding the idea of minimizing bad clocks. One way would be to shorten the chain. What about an experiment attaching a USB data store to the sMS-200 (Ultra) directly? My initial thought was a USB flash drive, although there too, I'm sure the "clock" on the controller is probably pretty crappy. Not sure there's any such thing as an externally clocked USB drive in the SOtM catalog, is there? Of course, even if there were, this would hardly be the Roon-like music playback experience. It would have to be controlled by MPD, but could be a proof of concept of end-to-end high-quality clocking.

 

The reason I ask about flash drives is that I just did this experiment in my recent comparion of the sMS-200, mR, and Aries Mini, as someone requested it. Have a look here - http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/auralic-aries-mini-vs-sonore-microrendu-vs-soul-music-sms-200-listening-impressions-31499/index3.html#post637708 . Of course, in my case, I did not have the sCLK-EX improvements in place on the sMS-200.

[austinpop]

romaz said:

Here are some excerpts I've collected from Paul's e-mails to me over the past months.

 

Just double checking - none of the SR3/5/7 are ultra-capacitor designs, correct?

 

I have no doubt that these are superb power supplies, as they should be, given the component quality, design, and attention to detail that is evident.

 

I do want to stress that there may be situations where the LPS-1 is the better supply to use, your experience to the contrary notwithstanding. It all depends on whether the LPS-1, at the location it's deployed, breaks up a ground leakage loop. In that situation, its effect can be massive. So beware - the LPS-1 is either just a very good LPS, or a truly transformative component. Which will it be for you? Unclear. You'll have to try it and see.

 

I posted about this on another thread, so rather than repeat myself, have a read:

 

 

 

[austinpop]

No, none of Paul's supplies use ultracaps. They are connected to ground.

 

You are absolutely correct, we each have different listening environments and so you will need to figure out what part of my listening experience will be applicable to your own. Having said that, I have had other DACs come through my system recently (brought in by members of our local audiophile society) for comparison against my DAVE that don't have USB galvanic isolation and regardless of the DAC used, there has been unanimous consensus of which is better.

Actually my DAC (Ayre Codex) too seemed to be fairly immune to the charms of the LPS-1 directly upstream of it.

 

Here is my optimal chain. Notice the LPS-1's optimal location at the FMC upstream of the Aries Mini.

 

• Aries Mini: W10 bridge > FMC (Teradak) > FMC (LPS-1) > Aries Mini (Auralic LPS) > Intona > Vbus2 > RUR (el cheapo LPS) > Ayre Codex DAC
  

I remember while my LPS-1 was still on order, Alex was convinced that in my chain, the LPS-1's ideal location was powering the RUR, and I was completely sure he was right. After all, logic dictates that the combination of GI on the USB provided by Intona, and the GI from AC provided by the LPS-1 would create the perfect storm of goodness for me.

 

Instead, I found the LPS-1 quite underwhelming at that location. In fact, I found the el cheapo to sound better! It wasn't until someone (this was on the Overall Isolation thread) suggested that I try it upstream at the FMC did I have that WOW experience with it. See - http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/overall-isolation-network-universal-serial-bus-industry-standard-cables-connectors-and-communications-protocols-between-computers-and-electronic-devices-and-power-29916/index23.html#post611383

[austinpop]

@romaz -

 

I've been thinking about these new experiments of yours. We as a community are fortunate to have someone with your curiosity, financial means, highly resolving audio equipment, and most of all, your ears going down this path. Thanks for doing this.

 

It's also important to place this in the context of the taxonomy that @JohnSwenson and others have postulated about system optimization. For example, here is just one of John's posts: http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/overall-isolation-network-universal-serial-bus-industry-standard-cables-connectors-and-communications-protocols-between-computers-and-electronic-devices-and-power-29916/index12.html#post597250

 

He talks about timing, signal integrity, and leakage loops.

 

So far, while leakage loops have been explored in both the Ethernet and USB domain, the focus of timing and signal integrity has been primarily on the USB side.

 

Your experiments with applying a high-precision clock on the network switch is the idea that timing and signal integrity matter very much on the Ethernet side too. To be honest, this one seems hard to reconcile with the fact that network propagation is inherently buffered and error-correcting, so should be insensitive to these things.

 

But as always, the empirical results precede the science that explains them.

 

I for one will be watching your clocking experiments with great interest. In your case, you are starting with a DAC with outstanding clocks. So far, you've cleaned up the clocks on 3 of the upstream components from the DAC, by my count: the Trendnet switch, the sMS-200, and the dX-USB-HD, and you noticed an incremental improvement with each step.

 

I admire your intent to clean up "all noisy clocks!" The key finding will be - how far upstream does this cease to matter?

 

<Tongue firmly in cheek> - Maybe you'll find yourself shipping precision clocks to your ISP to deploy at their headend directly upstream of your house. Just kidding!

[austinpop]

Knowing how a "femto clock" will benefit a component is not easy to gauge on paper. It seems everyone and their uncle is promoting a femto clock these days. I used to own the Aries with the femto clock and LPS upgrade and to be honest, compared against the base Aries without the clock, the difference was never night and day. I used to own a W4S DAC 2 and having owned it for a few months, I decided to send it back to W4S for their femto clock upgrade. While there was an improvement, it wasn't anything to write home about. During my evaluation of dCS's $13k master clock, I was somehow expecting this incredible improvement given the expense of that clock but found myself disappointed once again.

 

There are obviously many things to consider when looking at a clock upgrade. I am not a clock expert but as I see it, the two parameters most look at are phase noise (which is a measure of a clock's jitter) and stability (both short term and long term) with good short-term stability being much more important in audio. Here is a very helpful comment from John Swenson regarding the pros and cons of an atomic clock (which many consider as the reference standard when it comes to clocking) vs a good OCXO. Note also that he states that it isn't just the clock's characteristics that you have to consider but what your DAC must be capable of to benefit from this clock:

 

"A rubidium standard offers NO advantage over the very good OCXO. The rubidium standard has two systems, a rubidium oscillator, which has high jitter but very good long term stability and an OCXO with very low jitter but not as good long term stability. Some complex circuitry in side there reads both and every so often slowly tweaks the OCXO to match the long term averaged frequency of the rubidium oscillator.

 

Because the OCXO is an adjustable oscillator it actually has slightly higher jitter than the fixed OCXO. Audio could care less about absolute frequency accuracy over tens of years time frame so the rubidium version has higher jitter and costs more, not particularly a good combination for audio use. (unless it is all about bragging rights, but that is something else!)

 

The phase noise specs for those cybershaft OCXOs are actually very good for the price. The big issue with any such external frequency reference is how it gets into the DAC and what happens to it in there.

 

First off, many frequency standards are sine wave output, a lot of DACs that have external inputs want a square wave not a sine wave. Make SURE the reference and the DAC will work together before spending any money.

 

Almost no DAC or audio device uses 10MHz directly. In order to use it the frequency has to be converted. What that conversion does to the phase noise of the input can vary wildly. The absolute best systems out there are at about on par with the phase noise from the premium. So with a premium you would be getting about twice the jitter inside the DAC. With the limited the internal jitter is going to be several times higher than the reference, hence there is not going to be much of an actual difference in the DAC, for a much bigger cost.

 

The above assumes your DAC has a state of the art frequency conversion circuit, these are pretty rare and expensive but COULD exist in one or two audio devices. The problem is that any device with such a conversion circuit probably already has a REALLY good local oscillator, so using one of these OCXOs going through the conversion is not necessarily going to give you lower jitter in the DAC. It may, but it may not.

 

John S."

 

 

I have come to appreciate that devising a good clocking scheme for a DAC is not an easy thing to do, that there are many challenges to to consider and overcome and that throwing a super duper femto clock with excellent characteristics into the mix is hardly ever enough to fix a much more complex problem. This is what Rob Watts, designer of my Chord DAVE had to say about femto clocks:

 

"The issue of clocks is actually very complex, way more of a problem then in simply installing femto clocks. People always want a simple answer to problems even if the problem is multi-dimensional and complex. I will give you a some examples of the complexities of this issue.

 

Some years back a femto clock became available, and I was very excited about using it as it had a third of the cycle to cycle jitter of the crystal oscillators we were using. So I plugged it in, and listened to it. Unexpectedly, it sounded brighter and harder - completely the opposite of all the times I have listened to lower jitter. When you lower jitter levels in the master clock, it sounds smoother and warmer and more natural.

 

So I did some careful measurements, and I could see some problems.

 

The noise floor was OK, the same as before, and all the usual measurements were the same. But you could see more fringing on the fundamental, and this was quite apparent. Now when you do a FFT of say a 1 kHz sine wave, in an ideal world you would see the tone at 1 kHz and each frequency bucket away the output would be the systems noise floor. That is, you get a sharp single line representing the tone. But with a real FFT, you get smearing of the tone, and this is due to the windowing function employed by the FFT and jitter problems within the ADC, so instead of a single line you get a number of lines with the edges tailing of into the noise. This is known as side lobes or fringing. Now one normally calibrates the FFT and the instrument so you know what the ideal should be. Now with a DAC that has low frequency jitter, you get more fringing. Now I have spent many years on jitter and eliminating the effects of it on sound quality, and I know that fringing is highly audible, as I have done many listening tests on it. What is curious, is that it sounds exactly like noise floor modulation - so reduce fringing is the same as reducing noise floor modulation - they both subjectively sound smoother and darker with less edge and hardness.

 

So a clock that had lower cycle to cycle jitter actually had much worse low frequency jitter, and it was the low frequency jitter that was causing the problem and this had serious sound quality consequences. So a simple headline statement of low jitter is meaningless. But actually the problem is very much more complex than this.

 

What is poorly understood is that DAC architectures can tolerate vastly different levels of master clock jitter, and this is way more important than the headline oscillator jitter number. I will give you a few examples:

 

1. DAC structure makes a big difference. I had a silicon chip design I was working on some years back. When you determine the jitter sensitivity you can specify this - so I get a number of incoming jitter, and a number for the OP THD and noise that is needed. So initially we were working with 4pS jitter, and 120dB THD and noise. No problem, the architecture met this requirement as you can create models to run simulations to show what the jitter will do - and you can run the model so only jitter is changed, nothing else. But then the requirements got changed to 15 pS jitter. Again, no problem, I simply redesigned the DAC and then achieved these numbers. So its easy to change the sensitivity by a factor of 4 just by design of the DAC itself - something that audio designers using chips can't do.

 

2. DAC type has a profound effect on performance. The most sensitive is regular DSD or PDM, where jitter is modulation dependent, and you get pattern noise from the noise shaper degrading the output noise, plus distortion from jitter. R2R DAC's are very sensitive as they create noise floor modulation from jitter proportionate to the rate of change of signal (plus other problems due to the slow speed of switching elements). I was very concerned about these issues, and its one reason I invented pulse array, as the benefit of pulse array is that the error from jitter is only a fixed noise (using random jitter source with no low frequency problems). Now a fixed noise is subjectively unimportant - it does not interfere with the brains ability to decode music. Its when errors are signal dependent that the problems of perception start, and with pulse array I only get a fixed noise - and I know this for a fact due to simulation and measurements.

 

3. The DAC degrades clock jitter. What is not appreciated is that master clock jitter is only the start of the problem. When a clock goes through logic elements, (buffers level shifters, clock trees gates and flip-flops plus problem of induced noise) every stage adds more jitter. As a rough rule of thumb a logic element adds 1 pS of more jitter. So a clock input of 1pS will degrade through the device to be effectively 4 pS once it has gone through these elements (this was the number from a device I worked on some years ago). So its the actual jitter on the DAC active elements that is important not the clock starting jitter.

 

The benefit I have with Pulse Array is that the jitter has no sound quality degrading consequences - unlike all other architectures - as it creates no distortion or noise floor modulation. Because the clock is very close to the active elements (only one logic level away), the jitter degradation is minimal and there are no skirting issues at all. This has been confirmed with simulation and measurement - its a fixed noise, and by eliminating the clock jitter (I have a special way of doing this) noise only improves by a negligible 0.5 dB (127 dB to 127.5 dB).

 

This is true of all pulse array DAC's even the simpler 4e ones. In short the jitter problem was solved many years ago, but I don't bleat on about it as its not an issue and because it's way too complex a subject to easily discuss.

 

Pulse Array is a constant switching scheme - that is it always switches at exactly the same rate irrespective of the data, unlike DSD, R2R, or current source DAC's. This means that errors due to switching activity and jitter are not signal dependent, and so is innately immune from jitter creating distortion and noise floor modulation and any other signal related errors. The only other DAC that is constant switching activity is switched capacitor topology, but this has gain proportionate to absolute clock frequency - so it still has clock problems.

 

I plan to publish more detailed analysis of this, but from memory all of my DAC's have a negligible 0.5dB degradation due to master clock jitter, so its a non issue.

 

And yes you are correct, the absolute frequency is quite unimportant, so forget oven clocks, atomic clocks etc. Also the clock must be physically close to the active elements,with dedicated stripline PCB routing with proper termination. Running the clock externally is a crazy thing to do, as you are simply adding more jitter and noise and an extra PLL in the system."

 

So, to answer your question, you can go around asking about the phase noise and stability measurements of various clocks and they will serve as a useful starting point of reference for comparison but will they guarantee that they will sound good? Based on the above comments by John and Rob, I think you know the answer to this already as there is so much more to consider. What I will say is that where all my other "clock upgrade" experiences from the Aries to the dCS Master clock have been underwhelming, this quad clock board by SOtM has been anything but and it has expanded my horizons of just what a good clock can accomplish when implemented properly.

 

Thanks. That is very educational. So in your opinion, there are no metrics that can tell you how good or bad a clock upgrade will sound?

[austinpop]

Agreed that so much of a listening experience is system dependent but of course, personal preference also comes into play.

 

Should you decide to try SOtM's new clock for your sMS-200, you may wish to also try their upcoming tX-USB Ultra after your Intona as SOtM's clock will likely surpass the performance of the Crystek clock in the Recovery. My time with the Recovery even when powered by the LPS-1 never lead to this level of step change.

 

What I was hoping is that the Ultra would enable direct attach to the DAC without need of an Intona or tx-USBUltra.

[austinpop]

Neither the sMS-200 Ultra nor the tX-USB Ultra will provide galvanic isolation. If this is something you need, you will probably still require the Intona. Should you use the Intona, you will get the GI you seek but the Intona will degrade your SI in other ways which is why you would then add the tX-USB Ultra after it. This is the type of complexity I find myself running away from when possible but sometimes the benefits are hard to ignore.

 

Well, this is where things get complicated, isn't it? Yes, in my current chain, I do have an Intona and RUR between streamer and DAC. And yes, the Intona does improve SQ. But the Intona would qualify as a bad clock, compared to the precision of the new SOtM sCLK-EX. So the chain would then become:

 

sMS-200Ultra ---------------> Intona --------------> tX-USBUltra -----------> DAC

(good clock) -----------------(bad clock) -------------(good clock)

 

The questions then becomes:

1. does the benefit of the Intona (galvanic isolation) outweigh the harm (bad clock)?
   
2. If yes, then is the good clock on the sMS-200 Ultra wasted?
   

 

I guess the only way to answer this will have to be to compare:

1. sMS-200 Ultra ---> DAC
   
2. sMS-200Ultra ---> Intona ---> tX-USBUltra ---> DAC
   
3. sMS-200 ---> Intona ---> tX-USBUltra ---> DAC
   

 

I'd have to score some review samples to try that!

[austinpop]

BTW - I had an email back from May on some questions and she gave me two nuggets of info:

 

 

1. the sMS-200 Ultra's SQ is improved significantly in their own listening tests by adding the tX-USB Ultra in series. Obviously, we'll need to validate that in the community, but @romaz 's findings with the dX-USB HD (with sCLK-EX clocking) certainly support this.
   
2. She doesn't have the exact specs finalized, but she thinks the sMS-200 Ultra will require at least 1.5A (this is not an official number, so don't hold me to this) - which means we may have to look at other PSes than the LPS-1 to power it. It will be interesting to see what their upcoming sPS-500 PS really is, and whether it is ultracapacitor based. She could not share any details yet.