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

[romaz]

22 minutes ago, ElviaCaprice said:

 

You lost me here, Roy.  You have yet to replace your mobo clocks but you saw a big improvement in SQ with the SOTM products and new clocking.  I thought you were in the mindset that working back from the DAC upstream that the best clock is first next to the DAC and so on. 

I don't know that we can say that better clocking in the mobo is really the panacea that we are hoping for, without the ability to provide clean power throughout the mobo.  Which in my opinion comes first and foremost over any new clocking.

All we know, so far from observation, is that placing a bad clock after a better clock, downstream, seems to effect SQ negatively.  Same can be said for power supplies.

 

It could just be that once we can power the server/mobo with quality power throughout the stream that clocking becomes insignificant thru the remaining downstream as long as we maintain clean power.  Idea being that there is no need to fix the audio stream which was effected by bad power in the beginning stages of the stream.

 

 

Yes, working back from the DAC is the most cost-effective way of improving SQ.  The best clock should be placed just before the DAC.  What AmusedToD is suggesting is to replace the clock in his router which is just about as far upstream from the DAC as you can get.  If you do this and you don't do anything with the bad clocks in between the router and his tX-USBultra, I don't believe you will notice much (if any) improvement at all.  Having placed my reclocking switch before my server, I am hard pressed to hear a worthwhile improvement.

 

Yes, I believe good power to the clock is as important as good power to the component itself and so that is the limitation of changing the motherboard clock(s) since there are many subclocks in the motherboard that will not be cleanly powered.  Those I have communicated with who have replaced their motherboard system clock felt it resulted in a definite but not a mindblowing improvement and so I am prepared for that.  This is why I think changing the clock in your output card (whether it be a USB card or a NIC) potentially results in even greater improvement.  I still believe that with all of this, there will be benefit from another lightweight endpoint like an sMS-200Ultra or tX-USBUltra but what I am hoping is that I won't need 3-4 endpoints.  I am looking for simplicity.  

 

Regarding the importance of clean power relative to the importance of clocks, just know that with an SR7 or LPS-1 powering server and each endpoint, these reclocking devices are still making a significant difference.  Because the only difference between a stock SOtM product and an Ultra is the clock, there is no mistaking what a good clock can achieve irrespective of good power.

[Johnseye]

36 minutes ago, ElviaCaprice said:

 

Yes, but you will fail to address still the faults of the regulators on the mobo which create poor power regardless of how clean the original power supply is.  At least so far.

 

I repeat, which you missed.

It could just be that once we can power the server/mobo with quality power throughout the stream that clocking becomes insignificant (to a certain degree) thru the remaining downstream as long as we maintain clean power.  Idea being that there is no need to fix the audio stream which was effected by bad power in the beginning stages of the stream.

 

I will also add as long as those remaining clocks are sufficient enough not to create new jitter/noise.  If I am correct then the streaming chain becomes much more simple and short, server to DAC.

 

Could be.  I don't think we'll know for sure until we replace every clock on every device we control, then pull them out one by one until we hear the difference.  Only then will we know if it benefited one person's sound quality.  Then enough people need to repeat it until a consensus is formed.  Then repeat the process for the regulators or any other variable.

[ElviaCaprice]

I would also like to add that we don't know yet if a $200 clock is any better than a $10 clock.  It just may be that the $10 clock when implemented correctly with clean power is just as good as the $200 one.  I guess we will know soon enough with the ISO Regen.

[romaz]

42 minutes ago, AmusedToD said:

Regaring bridging two LAN cards in my PC, please note I am using a late 2014 iMac, so I am not really sure this can be done due to the construction of the computer. If you have a suggestion how to overcome this, I would highly appreciate that.

 

 

If you have Thunderbolt, I would suggest you get a Thunderbolt LAN card.  They're inexpensive.

 

43 minutes ago, AmusedToD said:

But what about my WIFI (PC) -WIFI (streamer) proposal for establishing connection between my PC and sms200? I know some users have reported SQ gains by using a wifi dongle instead of LAN (and May also suggested to me this can be a legitimate "upgrade" path).

1

 

I have not tried this and so I can't comment.  If you decide to bridge your LAN ports, let us know what you think how it compares.

[ElviaCaprice]

5 minutes ago, romaz said:

Regarding the importance of clean power relative to the importance of clocks, just know that with an SR7 or LPS-1 powering server and each endpoint, these reclocking devices are still making a significant difference.  Because the only difference between a stock SOtM product and an Ultra is the clock, there is no mistaking what a good clock can achieve irrespective of good power.

 

That is true, but it still remains to be seen if the $10 clock is just as effective as the $200 clock.  What was replaced was obviously not sufficient enough.

[romaz]

34 minutes ago, ted_b said:

Roy, thanks for all this.  I am contemplating trying the S3B (as Larry is recommending it too).  My comparison will be an NAA setup (I have two; Windows NAA and microRendu as NAA).  If S3B brings more clarity and body, then great, I jettison the NAA idea.  But Jussi's NAA approach offers some benefits (like simple dac driver environment) that I;m not sure the S3B will also ameliorate.  What say you?

 

 

Ted, I have yet to compare Adnaco + tX-USBUltra vs my NAA string from a SQ standpoint.  The advantage of straight USB for me is compatibility and versatility.  Both mR and sMS-200 have different modes and the latest mR update now adds Spotify Connect which is great but with straight USB, I can use just about any software player that will run on Windows and not have to worry.  With my Chord DAVE specifically, it also frees me up to use Chord's ASIO driver which has broader capabilities including native DSD playback (instead of DoP).  While DoP sounds just as good, I am finding more stuttering with it whereas native DSD playback results in no stuttering at all even with my native DSD256 files.

 

For me, the only advantage of NAA was the noise buffer that Ethernet provides but as we now know, despite having multiple cleanly powered, reclocking isolation devices in series, there is still enough noise in the Ethernet pathway that Ethernet cables can make a big difference.  With the Adnaco fiber solution, I don't see how it can be beat. 

[romaz]

9 minutes ago, ElviaCaprice said:

 

That is true, but it still remains to be seen if the $10 clock is just as effective as the $200 clock.  What was replaced was obviously not sufficient enough.

 

 

Not sure what you mean by "sufficient enough."  I think the Ultra clock has exceeded my expectations.  With other clock upgrades I have experienced in the past including dCS' $14k master clock option for the Vivaldi, I don't recall hearing improvements of this magnitude.  Having owned or tried various reclocking devices in my system including a USB Regen, W4S Recovery, various iFI reclocking devices, PS Audio LANRover, etc., I have yet to hear this level of improvement.  The problem with any device, no matter how good, is that SQ will never be quite as good as the real thing and so there will always be room for improvement and so we each have to define what "sufficient enough" means.

 

For me, the beauty of SOtM's sCLK-EX is that you have at your disposal 4 independent clocks that can be powered with a single PSU and so even if you place these clocks in dirty environments like a music server, the clocks themselves remain cleanly powered.  If you buy any OCXO, these clocks generally require 1A of current each.  To have to buy 4 OCXOs requires 4 LPS-1s or 4 rails from an SR7 which makes the whole concept inexpensive and inconvenient.  We don't yet have phase noise or stability measurements on SOtM's latest clock and it's quite possible there are better clocks out there but is the very best super duper clock going to be better than the collective benefit you would get from 4 of SOtM's clocks?  To me, it's unlikely.  

[AmusedToD]

24 minutes ago, romaz said:

 

If you have Thunderbolt, I would suggest you get a Thunderbolt LAN card.  They're inexpensive.

 

 

I have not tried this and so I can't comment.  If you decide to bridge your LAN ports, let us know what you think how it compares.

 

Ok, I think I didn't get you right the first time. I do have a LAN input on my IMac, and at the moment my iMac is directly connected in bridged mode to my sms200 via an entry level Audioquest Pearl CAT7 cable, bypassing the router. And yes, I confirm the SQ gain is not insignificant compared to the "standard" connection which includes a router and/or switch. But since my PC is some 4m away from my sms200, I really can't afford a 4m high end Ethernet cable, therefore I decided to try the wifi link between my PC and streamer.

 

[limniscate]

@romaz

 

Do you think I should just skip out on the iSO-CAT6 and get three dCBL-CAT7 cables instead?

[romaz]

2 minutes ago, limniscate said:

@romaz

 

Do you think I should just skip out on the iSO-CAT6 and get three dCBL-CAT7 cables instead?

 

 

If you are operating with that type of budget, also get an iSO-CAT6 and see what you think.  Even though my experience with it has not been overwhelming, it contrasts with the experience of others who have found much more significant improvement with it.  Just make sure you can return it if you're not happy with it and I believe that option will be there.

[ElviaCaprice]

I agree, Roy.  At the moment the sclk EX clocking board is the only game in town for multiple clock replacements.  I am keeping the Adnaco option open and couldn't agree more with you assessment on software flexibility and format versatility. 

I take it your now on hold until your new mobo is built.  Or will you go ahead, now, and send the Adnaco off for updates?

[Confused]

This may be the stupid post of the day, but there is a thought in my head and I'm just going to ask the question.  There is much talk of utilising the four clocks in the sCLK-EX.  I get this.  My question is could you utilise one clock tap to two or more devices?  If this works, it would obviously put the two devices fed into perfect sync, but is it possible.  Maybe it's undesirable for some reason if it would work?

[rickca]

2 hours ago, limniscate said:

This new Rowland PSU looks like an Uptone LPS-1 type solution for DAC and Pre.

The Jeff Rowland PSU works only with the Aeris DAC and Corus preamp.  It is $8200 list and since I have an Aeris my dealer is dying for me to try it ... he is sure it will never come out of my system.  Apparently, the PSU does not do bank switching like the LPS-1.

[limniscate]

Just now, rickca said:

The Jeff Rowland PSU works only with the Aeris DAC and Corus preamp.  It is $8200 list and since I have an Aeris my dealer is dying for me to try it ... he is sure it will never come out of my system.  Apparently, the PSU does not do bank switching like the LPS-1.

But it does completely isolate the DAC and pre from AC, right?

[rickca]

5 minutes ago, limniscate said:

But it does completely isolate the DAC and pre from AC, right?

Absolutely, yes.  In that respect it is like an LPS-1.  A really, really expensive LPS-1!  The stock Aeris comes with a very sophisticated outboard SMPS.  The PSU is designed to replace that SMPS.

 

The stock Aeris is a wonderful DAC.  I'm happy to discuss this further, but it's way OT in this thread.

 

[ted_b]

1 hour ago, romaz said:

I can use just about any software player that will run on Windows and not have to worry.  With my Chord DAVE specifically, it also frees me up to use Chord's ASIO driver which has broader capabilities including native DSD playback (instead of DoP).  While DoP sounds just as good, I am finding more stuttering with it whereas native DSD playback results in no stuttering at all even with my native DSD256 files.

I have a Windows NAA so those ASIO "freedoms" are with my NAA too (the only real downside of Linux NAA).  So, although I have one less thing to worry about, I may still try the S3B.

[guerph]

@AmusedToD

It's been some time since I've tried the USB dongle on the sMS-200, but I vaguely recall it sounding a bit "thinner." I didn't stick with it long enough to really say for sure. I tried it before I upgraded to a wireless mesh system. I've seen conflicting info from SOtM on which method produces better results. I again seem to recall hearing @MayfromSOtM saying some months back that Ethernet is "preferred." Based on how much money SOtM (and others) make on selling Ethernet cables, I highly doubt you'll get them to admit the $15 dongle we bought off Amazon yields equivalent (let alone better) results. I will definitely give the dongle another shot once my tx-USBultra arrives.

[lmitche]

5 hours ago, romaz said:

 

Ted, I have yet to compare Adnaco + tX-USBUltra vs my NAA string from a SQ standpoint.  The advantage of straight USB for me is compatibility and versatility.  Both mR and sMS-200 have different modes and the latest mR update now adds Spotify Connect which is great but with straight USB, I can use just about any software player that will run on Windows and not have to worry.  With my Chord DAVE specifically, it also frees me up to use Chord's ASIO driver which has broader capabilities including native DSD playback (instead of DoP).  While DoP sounds just as good, I am finding more stuttering with it whereas native DSD playback results in no stuttering at all even with my native DSD256 files.

 

For me, the only advantage of NAA was the noise buffer that Ethernet provides but as we now know, despite having multiple cleanly powered, reclocking isolation devices in series, there is still enough noise in the Ethernet pathway that Ethernet cables can make a big difference.  With the Adnaco fiber solution, I don't see how it can be beat. 

Hi Roy, I'm delighted to hear you are enjoying the Adnaco solution.  I may have missed it, how are you powering the USB box and tx-USBUltra?  I'm very much looking forward to hearing the results of your Adnaco reclocking effort.

[seeteeyou]

On 5/12/2017 at 4:16 PM, romaz said:

With the support from a group of friends, we have purchased a few more motherboards to test including a SuperMicro board that I mentioned in a previous post.  This SuperMicro motherboard is a SoC (System on a Chip) board that has no external PCH and so all of its ports including SATA, USB, LAN, PCIe, etc. have a much more direct and low latency path to the CPU.  Moreover, this motherboard has the option of being powered either via an ATX PSU or a 12V DC supply and this ability proved to be invaluable.  While there are probably better ATX PSUs around, we powered this motherboard initially with an EVGA SuperNova G2 850w ATX PSU that we had on hand and without question, it sounded very good and very dynamic.  For $140, this is an excellent ATX PSU and looking at reviews, its ripple values at low load (10-20%) are excellent with all 4 rails (4.6-7.6mV) and handily outperforms the HDPlex DC-ATX converter I bought (at least on paper).  As usual, there are no reported measurements on output impedance as this parameter seems to hold no value for reviewers or most manufacturers.  Anyway, we then compared this to the SR7 and even with all the noisy switching regulators within this motherboard that get in the way, the improvement we heard in terms of improved dynamics and this tonally more vivid and more holographic sound was just stunning.  It was similar to what I heard when I went from my HDPlex to the SR7 with my Mac Mini.  Having now heard this, there was no way I could continue on with my ASRock uATX board and the HDPlex DC-ATX converter.

 

How do we interpret the differences at low load?

 

EVGA SuperNOVA G2 850 W

https://www.techpowerup.com/reviews/EVGA/SuperNOVA_G2_850/10.html

 

EVGA SuperNOVA 650 P2

http://www.kitguru.net/components/power-supplies/zardon/evga-supernova-650-p2-review/6/

 

Cooler Master Masterwatt Maker 1200W MIJ

http://www.kitguru.net/components/power-supplies/zardon/cooler-master-masterwatt-maker-1200w-mij-psu-review/5/

http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story3&reid=508

https://www.newegg.com/Product/Product.aspx?Item=N82E16817171104

Quote

Most power supplies today convert the 5V and 3.3V from the 12V rail. This can affect the power delivery of the 12V rail under extremely high loads and as such, the MasterWatt Maker 1200 MIJ has separate power conversion for the 5V and 3.3V rails.

 

If MasterWatt Maker 1200 MIJ were actually as good as a grand gets, could we even make our own DC cables with StarQuad wiring to power stuff like sMS-200 (12V) / sCLK-EX (5V) or wireless routers (12V) by any chance?

 

Or does that quote imply that each and every ATX PSU (including Masterwatt Maker 1200W MIJ) might be only good for 12V because both 5V and 3.3V rails must go through their own voltage regulators at some point?

[amir57bs]

On 2/22/2017 at 1:04 AM, romaz said:

Here are soPaul does not have a presence here on CA and for those who wish to correspond with him, you can do so via e-mail at [email protected].

I have send an email but no answer!

is this address OK?

[sbilotta]

18 minutes ago, amir57bs said:

I have send an email but no answer!

is this address OK?

Try [email protected]

[greenleo]

13 hours ago, rickca said:

The Jeff Rowland PSU works only with the Aeris DAC and Corus preamp.  It is $8200 list and since I have an Aeris my dealer is dying for me to try it ... he is sure it will never come out of my system.  Apparently, the PSU does not do bank switching like the LPS-1.

I wish I had 8.2k, LOL.

 

I am eager to heard your impression☺

[Johnseye]

5 hours ago, amir57bs said:

I have send an email but no answer!

is this address OK?

 

I've been trying to contact him as well, have received no reply in several days, and I already paid him.  I'm hoping he's either in Munich or just out of communication temporarily.  

[romaz]

16 hours ago, lmitche said:

Hi Roy, I'm delighted to hear you are enjoying the Adnaco solution.  I may have missed it, how are you powering the USB box and tx-USBUltra?  I'm very much looking forward to hearing the results of your Adnaco reclocking effort.

 

Hi Larry.  Yes, the Adnaco is as good as you've been saying.  I am presently powering it with an LPS-1.  I am powering the tX-USBUltra with my other LPS-1.  I have a double regulated SR7 that is now officially on its way to me from Scotland (tracking number in hand) and so this will be used for any and all endpoints.

[Albrecht]

21 hours ago, ElviaCaprice said:

I would also like to add that we don't know yet if a $200 clock is any better than a $10 clock.  It just may be that the $10 clock when implemented correctly with clean power is just as good as the $200 one.  I guess we will know soon enough with the ISO Regen.

Hi,

Stupid question perhaps, but are the clocks in the SoTM products, namely the "ultra" isolation device and the "ultra" DDC $200 clocks? My prob here is the $1000 price point for these boxes whose intention is to mitigate the poor performance of USB audio. The very good F-1 DDC previously blew away most all of the more expensive DDCs, and not only because of the clock. Real world, I know of at least several people who are using DDCs instead of a poor implementation of the USB input on their DACs.