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

[ElviaCaprice]

1 minute ago, amir57bs said:

 

Usb cable in my idea is very very important .

overally if you like emotional sound (true tone plus good micro dynamic) i think combination of gordon rankin codes plus purist cable is very good.

i do not like ultra smooth and clean sound of many systems , i prefer emotional sound .

 

If I wanted that kind of sound I'd do it with tubes in a SET amp/preamp.  Not a cable.  Best to use a USB adapter or shortest cable possible.

[amir57bs]

1 hour ago, ElviaCaprice said:

 

If I wanted that kind of sound I'd do it with tubes in a SET amp/preamp.  Not a cable.  Best to use a USB adapter or shortest cable possible.

 

90% of tubes amplifier/pre are not emotional. they sound bad .

most clean systems cut micro dynamics and sound boring. lowering noise is not easy without altering micro dynamics

 

[semajm85]

Besides the inclusion of the superclock....are there any other improvements to the sms200,ultra?

[ElviaCaprice]

nevermind

[austinpop]

@romaz

 

Roy, could you also ask if there is jitter introduced by the ISO-Cat6 isolator, and even perhaps the dCBL-Cat7 filters?

[pam1975]

slightly off-topic, can anyone recommend a good ethernet PCI express card?

Thanks!

[ElviaCaprice]

2 hours ago, amir57bs said:

 

90% of tubes amplifier/pre are not emotional. they sound bad .

most clean systems cut micro dynamics and sound boring. lowering noise is not easy without altering micro dynamics

 

I couldn't disagree more.  All we want is the digital streaming data, zeroes and ones to enter the DAC unencumbered.  You don't shape digital data, yes, you can shape the electrical noise/jitter.  All we want is the upmost detail, separation and biggest soundstage from the DAC, if you want to color your sound, this is the time to do it in analog.

electrical noise/jitter will effect the DAC's analog stage, if you consider this noise micro dynamics then I want no part of it.

[Johnseye]

34 minutes ago, pam1975 said:

slightly off-topic, can anyone recommend a good ethernet PCI express card?

Thanks!

 

Intel E1G42ET

[pam1975]

13 minutes ago, Johnseye said:

 

Intel E1G42ET

thanks!

[Johnseye]

 

5 hours ago, amir57bs said:

 

Usb cable in my idea is very very important .

overally if you like emotional sound (true tone plus good micro dynamic) i think combination of gordon rankin codes plus purist cable is very good.

i do not like ultra smooth and clean sound of many systems , i prefer emotional sound .

i think ethernet renderer may give you ultra clean ultra smooth ultra low noise sound but i prefer intimate sound of pcie usb card with AS gordon rankin codes.

the best way to get emotional sound is to keep the way minimal.

adding reclockers or ethernet devices like microrendu add complexity of streaming data transfer and it will degrade intimacy of sound.

 

5 hours ago, ElviaCaprice said:

 

If I wanted that kind of sound I'd do it with tubes in a SET amp/preamp.  Not a cable.  Best to use a USB adapter or shortest cable possible.

 

This should clear things up.  Shamelessly copied...

 

Regardless of how good your clock source, the cable you use to connect clocks and digital audio signals can affect the clocking accuracy quite dramatically. Word clock signals are conveyed as a simple square wave running at the sample rate, and are always conveyed using a matched‑impedance interface designed specifically to convey high‑frequency signals. In fact, it's the same basic interface as is used for conveying video signals, with 75Ω terminations at each end and cables with a characteristic impedance of 75Ω. As a result, the signal that falls out at the far end of the clock cable is generally pretty similar to what went in, and although the inherent cable capacitance will tend to round off the vertical edges slightly, every clock transition will be consistent, which is vitally important.

It's also critically important to terminate word clock signals correctly. The word clock input terminal (normally a BNC connector) on some devices is deliberately unterminated (high impedance), while some are permanently terminated and others are switchable. If the line is unterminated, the clock signal will be reflected back down the cable, which could prevent some devices from detecting and locking onto the word clock information. The usual way to apply a termination in these cases (as shown by the diagram, left) is to fit a BNC T‑piece to the word clock input terminal, attach the clock cable to one side and a 75Ω termination plug to the other. Double terminations (where an extra termination is added inadvertently to one already on the line) will result in a low-level signal with the same potential problem.

 

 

S/PDIF, AES3 and ADAT interfaces convey the audio data in a 'channel' coding format that also embeds word and bit clocks (and some other useful information). Basically, the signal switches the frequency of a square‑wave signal between 1.5 and 3MHz (for a 44.1kHz sampled signal). So we're talking pretty high‑frequency stuff here — well into video frequency territory, rather than normal audio. Trying to get this kind of signal down a mic cable is difficult because of the relatively high capacitance, which is why specific very low-capacitance 'digital audio cable' is manufactured and recommended for this application. (Other cable types intended for high‑speed data transfer can also work well, particularly Cat5 Ethernet cable).

 

The result of sending data down a cable intended for digital data transfer.shows the result of sending an AES3

 

The first signal through 2m of Canford Audio's DFT digital audio cable. What this graph is showing is a series of sections of digital audio data, all overlaid over one another. The top, horizontal blue line represents the high data value (binary 1), and the lower blue line represents the low data value (binary 0). Individual data bits obviously exist only at one of these two levels, and the divisions between one data bit and the next are represented by the vertical blue lines.

As you can see, with this digital audio cable the source signal is being received still looking very square and clean, and each individual 'bit cell' (also referred to as the 'eye' pattern) is clearly identifiable and very 'open.' The small red rectangles within each bit cell represents the minimum eye pattern area that must be kept clear for the receiver to be able to interpret the data reliably — in effect, this is the 'decision zone', where the receiver has to decide whether the data is high or low for that specific data bit.

 

Similar results when using standard mic cable of the same length.

 

The second graph shows exactly the same thing, but this time with the signal passing through 2m of standard professional star‑quad microphone cable. The rising and falling edges between each bit cell have obviously been slowed down dramatically, and follow a ragged curve rather than the crisp up or down switching seen in the DFT cable. This is due entirely to the capacitance of the cable storing, and then releasing over time, the electrical energy from each change of signal voltage. While the damage is very obvious, it's still not disastrous and the data can still be received reliably — but you can certainly see how the eye pattern is starting to 'close up'. Now imagine what would happen if you tried to send a signal down many tens of metres of mic cable! In practice, it would become unrecoverable after about 15 to 25m.

The critical aspect of this waveform damage is that the rising and falling edges of each bit cell are used to define the bit‑clock transitions, and the distortions caused by cable capacitance can therefore introduce clock jitter. The diagram above shows how the data is conveyed in the S/PDIF and AES3 interface formats. The audio data ones are conveyed by switching the signal between high and low (or vice versa) mid‑way through the bit period (ie. sending the 3MHz square wave), while audio data zeros are conveyed by remaining at a constant level for the entire bit period (sending the 1.5MHz square wave). This arrangement is actually very similar to the way standard linear timecode works — although that system operates by switching between 1kHz and 2kHz signals, giving rise to the characteristic warble. This kind of channel code arrangement (often called a Bi‑Phase Mark Code) has the advantage that the recovered data is not affected by the overall signal polarity — all that matters is whether the signal transitions midway through a bit cell period or not, rather than whether it is a high or low voltage at any particularly time. However, as the diagram shows, the amount of clock jitter is directly affected by the audio data content to some degree, because of the different cable‑capacitance charging and discharging times that occur.

 

With S/PDIF and AES formats, clock jitter is directly affected by the audio data content to some degree, because of the different cable‑capacitance charging and discharging times that occur.

 

Ideally, the receiving circuitry will be able to remove this interface jitter, but not all devices manage this equally well (as shown in the investigations into A‑D converter clock‑recovery above), and if the embedded clocks are to be used as a clock reference (as is common practice in D‑A converters, for example), then this interface jitter can become part of the overall system's clock jitter, potentially resulting in reduced D‑A (or A‑D) performance.

[amir57bs]

13 minutes ago, Johnseye said:

 

With S/PDIF and AES formats, clock jitter is directly affected by the audio data content to some degree, because of the different cable‑capacitance charging and discharging times that occur.

 

Ideally, the receiving circuitry will be able to remove this interface jitter, but not all devices manage this equally well (as shown in the investigations into A‑D converter clock‑recovery above), and if the embedded clocks are to be used as a clock reference (as is common practice in D‑A converters, for example), then this interface jitter can become part of the overall system's clock jitter, potentially resulting in reduced D‑A (or A‑D) performance.

 

Thank you

i could not focus on reading all the text, it is tooooooooo long  :-))))

[ElviaCaprice]

Any cable can pick up RF noise or increase impedance over distance, especially USB or DC, not sure what's going on with ethernet.  That is why you want as short as possible of a cable.  You don't need expensive cables to do this.  Just properly built with proper impedance.  If you are noticing a difference with cables, more than likely you have a more serious problem somewhere else.  That's been my experience. 

Anyways, this is all needless to argue over for this thread.  More importantly is what we are hearing and how components/clocking/power supplies are effecting that outcome.  I'll leave the cable mythology alone from here on out.

[hasticus]

12 hours ago, romaz said:

That means whether they use 6V or 12V, the number of watts consumed is the same.

this is possible I think only if there is DC-DC converter on-board, which is not very good. please correct me if I'm wrong.

[ElviaCaprice]

24 minutes ago, hasticus said:

this is possible I think only if there is DC-DC converter on-board, which is not very good. please correct me if I'm wrong.

Nope, your right.  I agree, one less regulator is always better.  But there is going to be other regulators for lower voltage needs, unavoidable. 

[Johnseye]

9 hours ago, romaz said:

sCLK-EX essentially uses an internal master clock by which the 4 clocks derive their timing and so all 4 clocks are synchronized.  He believes strongly that this clock can be improved upon by a good external 10MHz master clock such as one from Esoteric or dCS.  SOtM plans to build their own external master clock in the near future.  As previously mentioned, an external clock means more wires and the need for another good PSU.

 

Looking forward to hearing more on this topic.  I've read some different information on external master clocks from no improvement to degradation.

 

9 hours ago, romaz said:

 

In Lee's opinion, the sMS-200Ultra will sound best at 12V.  I would presume this applies to the tX also.  This has nothing to do with these devices being "understarved" at lower voltages as has been suggested.  These devices consume the same power regardless of voltage.  That means whether they use 6V or 12V, the number of watts consumed is the same.  The benefit of higher voltage is less current draw.  As stated many times on this thread, the less current a component draws, generally it means less noise in the ground plane.  If a component consumes 12 watts, at 12V, that equates to 1A of draw, whereas at 6V, that equates to 2A of draw.  The downside of higher voltage can be greater heat generation and so a manufacturer has to balance these traits.

 

It was a question, not suggestion.  I honestly don't know.  My question was: "Do we end up starving some devices of power, limiting their performance, in efforts to use certain devices or to keep power input levels low?"

 

As you point out, decreasing the voltage increases the current.  If higher current means less draw, and less noise then why are people suggesting undervolting these devices?  This is another question.  I honestly don't know and am looking for an answer.  In the long run undervolting could damage the device at the expense of saving power.

 

Others just pointed out there must be a converter on board.  Can you confirm with Lee?

 

 

9 hours ago, romaz said:

 

An sMS-200ultra will sound slightly better than a standard sMS-200 that has been modified to become an sMS-200ultra purely because longer clock cables have to be used.  No other reason and the difference shouldn't be great.  No word yet on the rebate to upgrade from standard to Ultra but it is essentially a trade in is my understanding.

 

SOtM doesn't yet know the sonic impact of the SMB connectors but they feel this is a better way to go.  This also means that you can now get beyond a 40cm clock cable and extend up to 1M but the longer this length, the greater the deterioration in SQ.

 

Definitely more convenient and allows for flexibility with other devices without having to ship it all back for modification.  A mobo or switch with multiple SMB interfaces could allow for different clocks to be used in the future as this technology, or our knowledge of it, advances.  But as with all interfaces vs. directly soldered on board, or even soldered wires, there is a high probability of issues; noise, distortion, etc.

 

It sounds like they just used the tx chassis with the sMS-200 board and made soldered wire connections to the sCLK-EX board.  Is this your understanding as well?  Do you know why they haven't just used the same clocks from the sCLK-EX directly on the sMS-200 board?  I can guess, but hoping you heard something direct.

 

9 hours ago, romaz said:

 

If 2 components have identical clock requirements (ie frequency and current draw) such as with 2 NIC cards or 2 switches, then one clock from the sCLK-EX can be used to clock both components with no detriment to SQ.  This means potentially, a single sCLK-EX can clock more than 4 items.

 

Do you or SOtM know the limit?

 

9 hours ago, romaz said:

 

Lee feels strongly that optical is a bad idea.  Lots of jitter.  I thought this applied mainly to things like an Intona and not as much with optical but according to Lee, he has been looking at this for the past 5 years and can't get around the fact that even optical results in considerable jitter leading to a more "closed" sound which is why he has intentionally avoided this type of GI and so his advice is to stick with passive isolation.  I have to agree that when I tested the Adnaco, as I previously posted, it did nothing to open up the soundstage, and while replacing 2 clocks in the transceiver could make up for the high jitter produced, the optical connection will potentially undo all the good done by any previous clocking efforts meaning that my efforts to reclock my motherboard, NIC card, router, etc, will be seriously impacted.  As such, I have decided to use my Adnaco as a hub for music storage where it made a remarkable difference.  I will still plan to change the clock in the PCIe card to address jitter caused to the music stream as it passes through optical but I will no longer plan to change the two clocks in the outboard USB transceiver since the impact of those two clocks will essentially be compromised as the signal goes through the optical.  This means I will also be abandoning my fiber NIC and FMCs since this would negate the benefits of changing the 2 clocks in my router.  Huge change of plans.  Ugghh.

 

I asked both May and Lee which would sound better:  SOtM USB card with sCLK-EX + tX-USBultra vs sMS-200 modded to Ultra status + tX-USBultra?  Neither are sure as they haven't done this testing.  My guess is an sMS-200ultra + tX-USBultra would have the greater potential with the X-factors being the "direct connection" path, the switch that you can throw in for very little money plus that fact that you have the option of using a pair of dCBL-CAT7s.

 

 

I suspect Lee has used a jitter meter to determine this.  I feel for you Roy as you've gone through a lot of effort and expense only to talk with Lee, turning things upside down.  This is an excellent example of where we make our presumptions through personal listening exercises.  As many here have hypothesized, running the packets through fiber will filter the upstream power and noise.  However there are different quality emitters, and a difference between LED and laser.  There are also different types of detectors which will introduce distortion and noise, all effecting the S/N ratio.  With fiber you're converting the packets to light and back to electrical again, and it seems Lee has a good understanding of the price paid for that conversion in jitter.

[lmitche]

17 hours ago, romaz said:

 

 

Lee feels strongly that optical is a bad idea.  Lots of jitter.  I thought this applied mainly to things like an Intona and not as much with optical but according to Lee, he has been looking at this for the past 5 years and can't get around the fact that even optical results in considerable jitter leading to a more "closed" sound which is why he has intentionally avoided this type of GI and so his advice is to stick with passive isolation.  I have to agree that when I tested the Adnaco, as I previously posted, it did nothing to open up the soundstage, and while replacing 2 clocks in the transceiver could make up for the high jitter produced, the optical connection will potentially undo all the good done by any previous clocking efforts meaning that my efforts to reclock my motherboard, NIC card, router, etc, will be seriously impacted.  As such, I have decided to use my Adnaco as a hub for music storage where it made a remarkable difference.  I will still plan to change the clock in the PCIe card to address jitter caused to the music stream as it passes through optical but I will no longer plan to change the two clocks in the outboard USB transceiver since the impact of those two clocks will essentially be compromised as the signal goes through the optical.  This means I will also be abandoning my fiber NIC and FMCs since this would negate the benefits of changing the 2 clocks in my router.  Huge change of plans.  Ugghh.

Hi Roy,

 

I have had discussions with several engineers, including one with John Swenson at RMAF 2015, raising concerns about optical fiber and audio sound quality. Like Mr. Lee, these concerns were relayed in a matter of fact way.

 

Unfortunately I hear better SQ with the Adnaco, then without, using my i7-6700k based hqplayer upsampling machine at DSD512. The Adnaco works at the PCIE bus level and at 10gb fiber speeds, so is a very different animal then the FMCs used in our early, failed, fiber FMC experiments.

 

As you say, the Adnaco delivers the most solid audio image I've ever heard from any source.  I am now enjoying listening to the Adnaco with the ISO REGEN in various configurations. But without hearing it for myself, and without any disrespect for these esteemed audio engineers, I cannot accept the conclusion that optical fiber is bad for audio until I've hear something better for myself.  I look forward to hearing the SOTM solution and comparing it to various other options including the Adnaco?ISO REGEN/SOTM devices.

 

I should also say that the level of openness, size of the soundstage, extension . . . blah, blah, blahI am hearing on a daily basis is so incredible that it is beginning to feel very foolish to pursue better SQ.  The system is high resolution in every single way I can contemplate.

 

Enjoy Munich,

 

Larry

[greenleo]

4 hours ago, lmitche said:

Hi Roy,

 

I have had discussions with several engineers, including one with John Swenson at RMAF 2015, raising concerns about optical fiber and audio sound quality. Like Mr. Lee, these concerns were relayed in a matter of fact way.

 

Unfortunately I hear better SQ with the Adnaco, then without, using my i7-6700k based hqplayer upsampling machine at DSD512. The Adnaco works at the PCIE bus level and at 10gb fiber speeds, so is a very different animal then the FMCs used in our early, failed, fiber FMC experiments.

 

As you say, the Adnaco delivers the most solid audio image I've ever heard from any source.  I am now enjoying listening to the Adnaco with the ISO REGEN in various configurations. But without hearing it for myself, and without any disrespect for these esteemed audio engineers, I cannot accept the conclusion that optical fiber is bad for audio until I've hear something better for myself.  I look forward to hearing the SOTM solution and comparing it to various other options including the Adnaco?ISO REGEN/SOTM devices.

 

I should also say that the level of openness, size of the soundstage, extension . . . blah, blah, blahI am hearing on a daily basis is so incredible that it is beginning to feel very foolish to pursue better SQ.  The system is high resolution in every single way I can contemplate.

 

Enjoy Munich,

 

Larry

Larry,

 

Any comment of the SQ of Adnaco + IsoRegen vs Adnaco + txUSBUltra if you've tried both.  I think this is a hot issue that many CA members and all members watching this topic are most interested! 

[greenleo]

18 hours ago, amir57bs said:

 

Usb cable in my idea is very very important .

overally if you like emotional sound (true tone plus good micro dynamic) i think combination of gordon rankin codes plus purist cable is very good.

i do not like ultra smooth and clean sound of many systems , i prefer emotional sound .

i think ethernet renderer may give you ultra clean ultra smooth ultra low noise sound but i prefer intimate sound of pcie usb card with AS gordon rankin codes.

the best way to get emotional sound is to keep the way minimal.

adding reclockers or ethernet devices like microrendu add complexity of streaming data transfer and it will degrade intimacy of sound.

I want the true sound, if the record is bad, then just sound bad would be fine to me.  However, I respect we have different favors, each to to their own.

[romaz]

On 5/18/2017 at 8:07 AM, Forehaven said:

Can you post a link to the switch you're talking about Romaz?

 

Thanks

 

Any switch will work but make sure you use one that can be powered by a good PSU like the LPS-1 and so avoid a 9V switch if you plan to use an LPS-1.  The PSU will once again make a big difference.

 

This is what I used just because I had it lying around:

 

https://www.trendnet.com/products/proddetail?prod=510_TEG-S50G

 

I would suggest you consider something like the Zyxel that Rajiv used as I believe it is a better switch.  

[romaz]

On 5/18/2017 at 8:43 AM, austinpop said:

One question in my mind is whether there is an ultimate "best of both worlds" approach that looks something like this:

 

 

 

That would be pretty incredible.

[romaz]

On 5/18/2017 at 9:29 AM, limniscate said:

@romaz

 

So if I'm using Schiit Yggy, do you think it's better for me to use the dX-USB Ultra or the tX-USB Ultra?  I know most people think that the AES input is the best, but I haven't really noticed a difference between the AES and the USB.  Also, would it be better to use a tX-USB Ultra and dX-USB Ultra with the Yggy?

 

That has been my impression with the Yggy, that it sounds best via AES but if you can't tell the difference, then it may not matter whether you go with the tX vs the dX.  The dX is versatile since it has optical SPDIF, coax SPDIF and AES outputs.

[romaz]

On 5/18/2017 at 10:48 AM, tboooe said:

Question to those who have had their switch modded by SOTM.  Will they mod any switch?  I am thinking about having them upgrade my AC Wireless bridge I am using that includes a 4 port switch.

 

https://www.amazon.com/D-Link-Wireless-Gigabit-Extender-DAP-1650/dp/B00JFOP688?th=1

 

I don't see why they couldn't mod a wireless bridge, they just have to be able to open it up and access the system board.  It would be interesting to see how upgrading the clock on that device improves it.

[pam1975]

@austinpop which version of the Zyxel router are you using?

 

[romaz]

On 5/19/2017 at 1:19 PM, AmusedToD said:

 

The sms200ultra gives you the same benefit of reclocking external devices (2 devices max), as I posted above. Confirmed by SOTM. But you get away with only one expensive linear PSU, and you get a single box solution.

 

Both the sms200Ultra and the txUSBUltra share the same chassis and the same clock, the difference is that the sms200Ultra has a built in streamer (the sms200) and that's why it's $1200, whereas the txUSBUltra doesn't have the steamer part so it's $210 cheaper. Everything else is identical.

 

It was great to meet you in Munich yesterday!

 

Part of the reason for the higher cost of the sMS200ultra is that it requires 2 clock inputs from the sCLK-EX.

[romaz]

On 5/19/2017 at 0:23 PM, Johnseye said:

Based on that information I won't be pursuing the splitter option.  Maybe another LPS-1 or a third rail.

 

 

That would be ideal, John.  You could also use a splitter cable and see what you think.  You may or may not care if you notice a small degradation.