Jump to content
IGNORED

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


Message added by The Computer Audiophile

Important and useful information about this thread

Posting guidelines

History and index of useful posts

Most important: please realize this thread is about bleeding edge experimentation and discovery. No one has The Answer™. If you are not into tweaking, just know that you can have a musically satisfying system without doing any of the nutty things we do here.

Recommended Posts

Larry, while I agree that your fiber to USB box is doing little, I think those of us with the NAA architecture (fifo buffer containing actual dac drivers and USB card) is doing more, and I often wonder if these SATA, etc issues are more important to the NAA (uRendu and SOtM notwithstanding) than the server.  And it is quite important to me, for example, as I seriously contemplate some of your recommendations on my server (replace Platinum ps with Titanium, replace SSD with 6TB hdd, thus also "replacing" NAS, install hdd via non-standard M.2 adapter, etc),  Your perspective and results could very well be looked at as quite different than mine, as you are really going direct, per se, and the latencies and driver issues are server-based for you.   I dunno.  This whole thread was predicated on using an NAA or streamer and measuring direct ethernet connections from the server, but now this has evolved (and not a bad thing at all) to overall architectural sq design.  And for that, the architectures of our signal paths are often very different.

 

Oh well, It's not a lot of money to try it, and it is easily undone.  :)

Link to comment
15 minutes ago, lmitche said:

GreenLeo,

 

I have spent hours chasing down random disk writes in Windows 10, finding the source, and researching a way to kill the processes associated with this random activity.  There are logging processes all over Windows.  In the Windows 10 optimization thread you will find various scripts that many of us have contributed to prevent this random disk writing behavior.

 

Left on it's own, WIndows makes promiscuous use of the hard disk and will write many megabytes of data per second despite zero user activity and a disabled swap disk.  Just run the resource monitor available of the performance tab of the task manager.  Once in the resource monitor watch the disk activity tab.

 

Bring a barf bag.

 

Despite my efforts, I still have 10s of thousand of bytes of logs flowing every second.  Very annoying.  

Hi Imitche,

 

Thank you for your reply.  It seems that I need to buy a good SATA cable then.

 

At the moment I'm using Windows Server 2016 Core mode that runs literally a handful of processes as shown using the command tasklist /FI "STATUS eq RUNNING".  However, I can't find a tool that can show the disk activity in Core Mode.  Any recommendation?

Link to comment
1 hour ago, ted_b said:

Larry, while I agree that your fiber to USB box is doing little, I think those of us with the NAA architecture (fifo buffer containing actual dac drivers and USB card) is doing more, and I often wonder if these SATA, etc issues are more important to the NAA (uRendu and SOtM notwithstanding) than the server.  And it is quite important to me, for example, as I seriously contemplate some of your recommendations on my server (replace Platinum ps with Titanium, replace SSD with 6TB hdd, thus also "replacing" NAS, install hdd via non-standard M.2 adapter, etc),  Your perspective and results could very well be looked at as quite different than mine, as you are really going direct, per se, and the latencies and driver issues are server-based for you.   I dunno.  This whole thread was predicated on using an NAA or streamer and measuring direct ethernet connections from the server, but now this has evolved (and not a bad thing at all) to overall architectural sq design.  And for that, the architectures of our signal paths are often very different.

 

Oh well, It's not a lot of money to try it, and it is easily undone.  :)

Ted, I'm beginning to think it may be time for a road trip.  

Pareto Audio aka nuckleheadaudio

Link to comment
28 minutes ago, greenleo said:

Hi Imitche,

 

Thank you for your reply.  It seems that I need to buy a good SATA cable then.

 

At the moment I'm using Windows Server 2016 Core mode that runs literally a handful of processes as shown using the command tasklist /FI "STATUS eq RUNNING".  However, I can't find a tool that can show the disk activity in Core Mode.  Any recommendation?

Greenleo, Sorry, but I don't know Windows Server 2016.  The resource monitoring tool has to be there somewhere.  Here is a site that may help:

 

http://serverfault.com/questions/810948/windows-server-core-network-and-disk-resource-monitoring

Pareto Audio aka nuckleheadaudio

Link to comment
5 hours ago, lmitche said:

Hmmm, I can't find any 2gb DDR4 memory availability,  Guessing you ordered ddr3?

 

 

 

That's right, I ordered DDR3.  As I was forced to go to an older 1150 socket motherboard to be able to run in IDE mode, DDR3 was the only option.

Link to comment
5 hours ago, greenleo said:

Hi Roy,

 

Any reasons that an OS drive is always busy in Windows?  Given that an audio PC just runs a few processes, after a short time, all the files to be run or be read would already been cached in the RAM, would it?  Also, by disabling the page file, there is no need in reading OS disk again.  The OS disk should not be busy.

 

Not sure if Windows needs to write something secretly into OS disk.  If not, then the OS disk activity may be reduced to to zero and the influence of the OS disk may be reduced to minimal.

 

This is a question that has puzzled me for a long time.  

 


 

Would you mind to tell us were to get the Intel X25E SSD?

  

3

 

Larry has addressed this very well already and so I won't go into it again.  

 

I got my Intel X25E from an IT supply store on EBay.  Here is a good review of this SSD:

 

http://www.tomshardware.com/reviews/intel-x25-e-ssd,2158-7.html

 

As a server grade device, this SSD has an MTBF of 2 million hours but what is more impressive is that as an SLC drive, it has a write endurance of 2 petabytes (2,000 terabytes) of random writes.  This drive should virtually last forever.

Link to comment
5 hours ago, lmitche said:

"I think the general consensus is that speed and power = noise and slower and less powerful is generally less noisy." - Roy

 

"You think these things matter so much even when we're talking about a music server connected to an endpoint like sMS-200, rather than a music player directly feeding a DAC?" - Rajiv

 

Aren't these two questions related?  First off we are all using two box designs with one box doing, heavy (hqplayer), or heavier lifting (SMBD server), the second converting from ethernet to usb, or fiber to usb each powered by an LPS-1.  My fiber to USB box uses very little power 5V DC@ 0.35A maximum for the R1USB30 itself. So, I think we are really talking about the same music sever - endpoint, high power - low power architecture,

 

With galvanic isolation in place, through either fiber or ethernet, I don't think that the power usage of an isolated server machine makes any difference, but that the cleanliness of the power does.

 

To me the mystery is why dicking around with something like the sata connection has such a huge impact.  I don't get it and can only guess we are exposing the impact of timings differences between various connection points by varying this parameter.  What do you think?

 

7

 

One of the goals of my server build is to see what would happen if I replaced as many of the clocks that are in the direct signal path within the server as possible.  As I will be replacing 4 bad clocks with 4 excellent clocks that will be very cleanly powered, like many of you, I have wondered whether such devices like the sMS-200 Ultra will still make a difference?  In a way, this experiment has already been done and the question has already been answered.  SOtM has developed a new version of their sMS-1000SQ music server that now incorporates their latest super clock and in their testing, the improvement was significant.  Just for fun, May told me that they decided to connect this server to their new tX-USB Ultra (their USB regen device that also incorporates their new superclock) and while they thought this might perhaps result in a small incremental improvement, May told me they were quite surprised to hear a "huge" improvement.

 

Today, I look at reclocking differently than I did even a couple of months ago.  Replacing a clock isn't just a "reclocking" event, it is a signal regeneration event and with each regeneration, there is an opportunity to create something new -- a fresh start.  With respect to a music stream, this means fixing (or at least partially fixing) a signal that was previously damaged.  

 

You have stated it very well.  At its core, both the mR and sMS-200 do nothing more than convert an incoming Ethernet signal into a USB signal.  This is similar to what a USB-to-SPDIF converter does.  Likewise, your Adnaco converts a USB signal from copper to fiber and then back to copper again.  In each instance, a signal is converted but with any conversion that takes place, a reclocking and, therefore, a regeneration of the signal has to take place.

 

If the regen/reclocking device has a better clock than the preceding server, then there is the potential for the new signal to have better fidelity than the preceding signal.  If the regen/relocking device is a lower impedance device than the server that preceded it, then there is the potential for the new signal to sound less confined and more dynamic.  If that regen/reclocking device is more cleanly powered than the server that preceded it (or in the case of the Adnaco or upcoming Iso Regen where you have galvanic isolation), then there is the potential for the newly generated signal to be cleaner as well.  

 

These lightweight add-on devices, however, can't act in isolation and must be fed by a server that will be responsible for the heavy lifting.  At the same time, I still strongly believe that these add-on devices can't completely fix a damaged stream and that the best way to address a problem is to prevent or minimize a problem to begin with and so while I have yet to fully objectively validate my beliefs, it is my belief at this time that you need both a well implemented server as well as a well implemented lightweight endpoint in the same way that you would want Michael Phelps swimming your 100m and Usain Bolt running your 100m.  No one component (or one person) can do all things well.

 

With your SATA findings, my guess is that it comes down to noise and latency.  In your previous setup, where your Adnaco USB and your hard drive were both going through the PCH, since you use your hard drive as both an OS and data drive, what would have to happen was that with every OS call and with every stream of a music file, a signal would originate from your hard drive that would travel through your PCH before eventually reaching the CPU.  Once processed by the CPU, it would then have to travel back through the PCH in order to get to your Adnaco.  Not only does it have to navigate through a potentially more congested pathway multiple times but from what I can gather, all PCHs run off of a 1-1.5V rail which suggests this is probably also a very noisy pathway.  Now that you have both your SATA ports and your Adnaco on a more direct, less congested and less noisy pathway, it's quite possible that it was never really SATA or USB that was the issue but how these buses are positioned within your motherboard.

 

As the topic of designing an audiophile-grade motherboard was brought up recently, I started thinking to myself what such a motherboard would entail.  Surely, it's not just about removing noisy switching regulators and unnecessary ICs.  Well, I came across an interesting paper that highlights some of the challenges and it is indeed incredibly complex and further lends credence to my belief that you will always want some lightweight endpoint to follow your server.  Here is that paper.

 

http://www.ti.com/lit/an/szza009/szza009.pdf

Link to comment
1 hour ago, romaz said:

 

One of the goals of my server build is to see what would happen if I replaced as many of the clocks that are in the direct signal path within the server as possible.  As I will be replacing 4 bad clocks with 4 excellent clocks that will be very cleanly powered, like many of you, I have wondered whether such devices like the sMS-200 Ultra will still make a difference?  In a way, this experiment has already been done and the question has already been answered.  SOtM has developed a new version of their sMS-1000SQ music server that now incorporates their latest super clock and in their testing, the improvement was significant.  Just for fun, May told me that they decided to connect this server to their new tX-USB Ultra (their USB regen device that also incorporates their new superclock) and while they thought this might perhaps result in a small incremental improvement, May told me they were quite surprised to hear a "huge" improvement.

 

Today, I look at reclocking differently than I did even a couple of months ago.  Replacing a clock isn't just a "reclocking" event, it is a signal regeneration event and with each regeneration, there is an opportunity to create something new -- a fresh start.  With respect to a music stream, this means fixing (or at least partially fixing) a signal that was previously damaged.  

 

You have stated it very well.  At its core, both the mR and sMS-200 do nothing more than convert an incoming Ethernet signal into a USB signal.  This is similar to what a USB-to-SPDIF converter does.  Likewise, your Adnaco converts a USB signal from copper to fiber and then back to copper again.  

 

Not that this changes your overall points.  But the Adnaco, I believe, only converts once to USB from Fiber.  It receives the stream from the PCIe bus, converts direct to Fiber.

 

I fail to see your point about the sMS-1000SQ server that incorporates the new super clock, how did they incorporate it?  Did they test whether the sMS-200 Ultra made a difference with the new server?  With both the 200 and the tx-usb Ultra?  Server plus tx-usb Ultra, no 200?

(JRiver) Jetway barebones NUC (mod 3 sCLK-EX, Cybershaft OP 14)  (PH SR7) => mini pcie adapter to PCIe 1X => tXUSBexp PCIe card (mod sCLK-EX) (PH SR7) => (USPCB) Chord DAVE => Omega Super 8XRS/REL t5i  (All powered thru Topaz Isolation Transformer)

Link to comment
21 hours ago, romaz said:

 

Sorry, I believe I didn't read your post properly and so I failed to answer your question.  With the standard way of connecting an mR or sMS-200 to a router first, I don't think these things matter as much.  With the direct connection, while transparency to the recording has increased, so has transparency to the qualities (and deficiencies) of the music server itself.  As long as I continue to hear improvements, I will probably continue to push along.

 

Think of it this way. Say you had a direct connection:

  • Music Server > USB > DAC

Let's say you improve the signal integrity by some factor S (using better clocks), and reduce noise (HW/SW optimization, zero leakage loops PSUs, etc) by a factor N.

 

Now consider the streaming case:

  • Music Server > Switch > sMS-200 > dX-USB HD > DAC
  • ______________< ---------- sCLK-EX -------------->

where your 3 intervening devices (switch, sMS-200, and dX-USB HD) have the best clocks (sCLK-EX) and the best PSUs.

 

The central question then becomes - do those improvement factors of S and N that you make on the music server still have the same impact? Or are they attenuated by a factor a (0 < a < 1) due to the 3 other intervening devices that are reducing noise and improving signal integrity?

 

Qualitatively, it sounds like you expect that a > 0, else this exercise would be pointless. I think I expect that a < 1, and eager to find out vicariously through your experiments if it is closer to 0.8 or 0.2. :D

 

 

Link to comment
4 hours ago, ElviaCaprice said:

 

Not that this changes your overall points.  But the Adnaco, I believe, only converts once to USB from Fiber.  It receives the stream from the PCIe bus, converts direct to Fiber.

 
1

 

Within Adnaco's PCIe card, there is most definitely a conversion that has to take place.  All PCIe adapter cards I have seen, whether it be PCIe to SATA or PCIe to USB have clocks which makes sense since the PCIe bus, SATA and USB all operate at different speeds.  As I look at Adnaco's PCIe card, there is indeed a clock on its PCB.  That same clock is also on their receiver module's PCB that converts fiber back to copper which is to be expected.  This is important to know since each clock will cost money if the goal is to replace them.  

 

4 hours ago, ElviaCaprice said:

I fail to see your point about the sMS-1000SQ server that incorporates the new super clock, how did they incorporate it?  Did they test whether the sMS-200 Ultra made a difference with the new server?  With both the 200 and the tx-usb Ultra?  Server plus tx-usb Ultra, no 200?

 
2

 

Some have asked the question about whether devices like an sMS-200 Ultra or tX-USB Ultra, which contain SOtM's new super clock would be necessary at all if all the bad clocks within the server itself were already replaced with their super clock.  

 

With SOtM's sMS-1000SQ, they don't generally expect people to use this server with an sMS-200 since this server incorporates their highly acclaimed tX-USBexp card which is an audiophile-grade PCIe USB card that has a very good clock of its own and so with this server, the intention is to connect straight to the DAC.  With their Ultra version of this server, they replaced the motherboard's system clock and the clock to their tX-USBexp card with their new super clock.  As such, they weren't expecting any further significant improvement when they attached the tX-USB Ultra and yet, they claim that adding the tX-USB Ultra resulted in a huge further improvement.

Link to comment
14 minutes ago, lmitche said:

As such, they weren't expecting any further significant improvement when they attached the tX-USB Ultra and yet, they claim that adding the tX-USB Ultra resulted in a huge further improvement.

 

The manufacturers don't know anymore then we do.

 

 

This is how I view digital.  It is a frontier which much that is still not well understood.

Link to comment
1 hour ago, romaz said:

 

Within Adnaco's PCIe card, there is most definitely a conversion that has to take place.  All PCIe adapter cards I have seen, whether it be PCIe to SATA or PCIe to USB have clocks which makes sense since the PCIe bus, SATA and USB all operate at different speeds.  As I look at Adnaco's PCIe card, there is indeed a clock on its PCB.  That same clock is also on their receiver module's PCB that converts fiber back to copper which is to be expected.  This is important to know since each clock will cost money if the goal is to replace them.  

 

 

I agree, replacing those clocks would make for an interesting component, the Adnaco.  As is, I don't even consider it a fair comparison.

1 hour ago, romaz said:

 

Some have asked the question about whether devices like an sMS-200 Ultra or tX-USB Ultra, which contain SOtM's new super clock would be necessary at all if all the bad clocks within the server itself were already replaced with their super clock.  

 

With SOtM's sMS-1000SQ, they don't generally expect people to use this server with an sMS-200 since this server incorporates their highly acclaimed tX-USBexp card which is an audiophile-grade PCIe USB card that has a very good clock of its own and so with this server, the intention is to connect straight to the DAC.  With their Ultra version of this server, they replaced the motherboard's system clock and the clock to their tX-USBexp card with their new super clock.  As such, they weren't expecting any further significant improvement when they attached the tX-USB Ultra and yet, they claim that adding the tX-USB Ultra resulted in a huge further improvement.

 

My big question is:  Which has better sound quality and how far apart are they. 

 

I am at a crossroads here after acquiring the 2Qute.  How to achieve the best SQ, pros and cons.

 

Go the Ethernet route:  sms-200 Ultra

Pros:  Probably? gives the best SQ regardless of the amount of tweaking on the direct USB route.

Cons:  Limited to audio only.  That new sms-200 ultra will probably require a more expensive linear power supply, LPS-1 not sufficient.  

 

Go the direct USB route:  Adnaco w/new clocks

Pros: Can do audio and video sound output.  How much difference in sound quality from Ethernet route?  One LPS-1 will be sufficient for the Adnaco, one LPS-1 for new Sotm sCLK-x, use extra clocks for motherboard system server and an add on tx-usb hub to follow the Adnaco if so desired?

Cons:  Is untested, not a cheap component either, Adnaco, new clocks may not work. 

 

When I don't know which path to take.  I do nothing and keep researching for answers.  Meanwhile enjoy the audio I have.

(JRiver) Jetway barebones NUC (mod 3 sCLK-EX, Cybershaft OP 14)  (PH SR7) => mini pcie adapter to PCIe 1X => tXUSBexp PCIe card (mod sCLK-EX) (PH SR7) => (USPCB) Chord DAVE => Omega Super 8XRS/REL t5i  (All powered thru Topaz Isolation Transformer)

Link to comment
2 minutes ago, ElviaCaprice said:

 

I agree, replacing those clocks would make for an interesting component, the Adnaco.  As is, I don't even consider it a fair comparison.

 

My big question is:  Which has better sound quality and how far apart are they. 

 

I am at a crossroads here after acquiring the 2Qute.  How to achieve the best SQ, pros and cons.

 

Go the Ethernet route:  sms-200 Ultra

Pros:  Probably? gives the best SQ regardless of the amount of tweaking on the direct USB route.

Cons:  Limited to audio only.  That new sms-200 ultra will probably require a more expensive linear power supply, LPS-1 not sufficient. 

 

Go the direct USB route:  Adnaco w/new clocks

Pros: Can do audio and video sound output.  How much difference in sound quality from Ethernet route?  One LPS-1 will be sufficient for the Adnaco, one LPS-1 for new Sotm sCLK-x, use extra clocks for motherboard system server and an add on tx-usb hub to follow the Adnaco if so desired?

Cons:  Is untested, not a cheap component either, Adnaco, reclocking may not work. 

 

 

I think the big X factor is the soon-to-be-released Iso Regen by Uptone.  I suspect this could be a better option than the Adnaco and should pair well with their LPS-1.

 

The other option I am very curious about is to go Toslink, especially with your 2Qute.  Here is an inexpensive PCIe audio card I have been looking at with interest that has a Toslink output:

 

http://www.sybausa.com/index.php?route=product/product&path=64_77_80&product_id=122

 

https://www.amazon.com/Syba-SD-PEX63081-Surround-CM8828-Chipset/dp/B00FWFNY92

 

It has a clock that can be replaced.  I also see 2 switching regulators that can be replaced with much better linear ones and numerous capacitors that can probably be upgraded.  This could be used to make a very good single box setup.

 

Link to comment
13 hours ago, lmitche said:

Once in the resource monitor watch the disk activity tab.

 

Bring a barf bag.

 

LOL!  Every time I look at Windows I feel that way. :S

 

32 minutes ago, lmitche said:

The manufacturers don't know anymore then we do.

 

9_9

Especially since we don't have the time to dig into the fringe weeds the way you guys do!  But we watch and come up with new product ideas to hopefully address/cure some of the ills.

Or maybe take it all in an entirely new direction...   (Can't wait until John's move is done and his lab is back up.)

Link to comment
1 hour ago, romaz said:

 

I think the big X factor is the soon-to-be-released Iso Regen by Uptone.  I suspect this could be a better option than the Adnaco and should pair well with their LPS-1.

 

The other option I am very curious about is to go Toslink, especially with your 2Qute.  

 

Thanks, but toslink is not an option, need DSD.

 

I've bought most of Uptones products in the past couple years, great products, but I doubt the Iso Regen will be in the cards for the 2Qute stream.  Don't need the USB galvanic isolation with the 2Qute.  Think in my case the Sms-200 Ultra would be the best route.  Use my Ifi Nano iDSD LE for my Video via direct USB.  There the Iso Regen would be a fit.

 

Just waiting to see what the specs are on the sms-200 Ultra and further reports on other components.   Your exploration has helped a lot Roy, along with others feedback.

 

Just don't feel comfortable with the Adnaco route, Larry.  Not enough information/comparisons or feedback.  Not a cheap device and resale is poor, if it doesn't work out.  If I find one cheap on resale, I would give it a shot.

(JRiver) Jetway barebones NUC (mod 3 sCLK-EX, Cybershaft OP 14)  (PH SR7) => mini pcie adapter to PCIe 1X => tXUSBexp PCIe card (mod sCLK-EX) (PH SR7) => (USPCB) Chord DAVE => Omega Super 8XRS/REL t5i  (All powered thru Topaz Isolation Transformer)

Link to comment
52 minutes ago, hopkins said:

This should make for an interesting product: Raspberrypi hat with isolation, reclocking, and high quality regulators:

 

http://www.diyaudio.com/forums/digital-line-level/305583-digione-rca-bnc-output-rpi.html

 

Spdif out only. 

 
 

 

I'm always glad to see new things by resourceful individuals or companies come to market and challenge the status quo and even raise the bar, especially at such a low price point.

 

This statement, in particular, caught my attention:  "Basically what we think, its that DigiOne will be the best digital transporter on the market at ANY price (yes including those 5kEUR/USD transporters)."  

 

That is a pretty bold statement that will be very hard to live up to and could end up causing people not to take them seriously.  Unfortunately, some of the very best servers I have heard sell for well over $5k and so if a $5k server is as good as this individual has heard, I think it may be a stretch to suggest that this will be the best transport at any price.  Furthermore, 50-100uV of noise and 1-3ps of jitter is very good but not close to the best there is.  Still, for 120 Euros, that's a pretty good deal for anyone looking for an inexpensive SPDIF endpoint.

Link to comment
5 hours ago, romaz said:

 

The other option I am very curious about is to go Toslink, especially with your 2Qute.  Here is an inexpensive PCIe audio card I have been looking at with interest that has a Toslink output:

 

But Toslink is limited to 192kHz, is it?

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...