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

[austinpop]

A couple clarifications.

 

7 minutes ago, austinpop said:

HDPlex 100W LPS, supplying 2 x 12V rails

Eventually, this will be replaced by the Paul Hynes SR7MR2DRXL that Eric has on order. That will be another Ultra-polooza for sure.

 

7 minutes ago, austinpop said:

Here are some burning questions we want to answer:

? Hmm, hope that wasn't a Freudian slip? ?

 

[austinpop]

1 hour ago, ElviaCaprice said:

Which only begs the question.  If reclocking is so important then why not reclock several times in the same component?  I suspect that its not so much about the actual clock itself (to a certain degree of quality, that being your DAC), but the quality of the power that clock receives.  Too bad you guys don't have an ISO Regen to throw into the mix.

 

Right now, we are going on purely empirical findings, motivated by @romaz's research. John Swenson has indicated this whole area is one he is actively researching, but is not ready to discuss his theories at this time. Over on the Mutec side, there are similar reports of people cascading MC-3+ USB reclockers to good effect.

 

We are at the dawn of understanding this stuff. Once harnessed, I am sure manufacturers will find clever ways to maximize performance - even for example, like you speculate, multiple times in a component.

 

Until then, I approach this with a spirit of fun, adventure, and the joy of learning.

 

It's a great time to be ...

[austinpop]

32 minutes ago, mozes said:

@austinpopJust to let you know, I will be doing these tests next week when my VR will have arrived:

1- tX-USBultra vs. Iso-Regen

2- Nimitra>IR>tX>Brooklyn vs. Nimitra>tX>IR>Brooklyn

I think this way we will almost have all the answers that we want so far.

 

Awesome, @mozes! Really looking forward to your findings.

 

I am also expecting @Bamber's kind loaner IR/LPS1 when it ships - hopefully in mid-July.

[austinpop]

24 minutes ago, BigAlMc said:

 

... loads of fun too ...

 

That is the operative phrase!

[austinpop]

15 minutes ago, rickca said:

Eric, Rajiv and Moussa I'm certainly looking forward to the results of your experiments.

 

I'm just wondering ... why does it make sense that multiple reclockings with the same quality clock is better than one? 

 

Wouldn't one top quality reclocking be better than multiple reclockings with something less good?

 

I'm not looking for a scientific explanation at this point, just something plausible.  Even without this, if it sounds better, it sounds better!  So carry on.

 

Some day soon, this will be better understood by the best minds in computer audio. Until then, we only have the empirical data to go on.

[austinpop]

Woot!

 

My DIY serial DC cable - while messy looking - works as designed. We are a GO for 2 x LPS-1s to power Eric's 12V Ultra gear.

 

[austinpop]

Progress update on index... 25%

 

I have now gone through about a quarter of this thread (25 pages) and populated the index on the first page with selected postings, organized by topic. I will continue progress as time permits. Have a look and see what you think so far.

 

Call for volunteers

 

I have 2 topics for which I'd like help in assembling useful posts:

• Computer mods - mobos, NICs, drives, etc
• Adnaco fiber mods

If you want to help, here's what I need. Just assemble useful links to posts in this thread, one link per line, and send to me. I'll add it to the index. No formatting or any other work necessary.

[austinpop]

27 minutes ago, greenleo said:

I suggest to compare the following in particular:

Modded switch + 200Ultra to DAC

                             vs

200Ultra+txUSBUltra to DAC

 

Actually, I've done this before, both myself, with @limniscate, and even @romaz has reported on this. 

 

In out setups, the effect of the two (switch vs. tX) were both positive but different. Both added more timbre, articulation, and resolution. However, the tX added greater solidity and improved bass, while the switch enlarged the soundstage, and created more air around instruments. So I would be hard pressed to pick one over the other.

[austinpop]

1 hour ago, TopQuark said:

@limniscateand @austinpop I wish you have Paul Hynes in the chain up to the DAC.  Where are you guys located at?  I have a spare in Silly Valley area if you are nearby.

 

 

That would be great. But we're not local, sadly. We're in Austin, TX. You'd have to ship it to us.

[austinpop]

8 hours ago, Confused said:

@austinpop Good luck with all this! I will enjoy reading about you exploits and findings.  I have to admit that my initial thinking was perhaps similar to @greanleo when reading your kit list.  Now that you have the DX-USB Ultra in the chain, I would be interested in the results with and without the tX-USBultra.  I would see this along the lines of you finding the absolute optimum set-up with all the options available, then establishing the delta reduction, if any, if the tX-USBultra is removed.  I think the logic here is simplifying the set up, and indeed seeing the influence of the removal of one of the more expensive components in the chain.  Just curious!  In fairness, I can see that you consider that the tX-USBultra has already proven it's worth, and the driver here is to add more clocked devices to the chain, not remove them, so essentially this goes back to how I started this post, I will follow all this with much interest, and good luck!

 

This would be a good time for me to step back and explain my thinking here. I can see it is causing some confusion.

 

First of all, the kit I listed is the aggregate of my and Eric's gear. I would never dream of buying all this stuff myself! Our respective chains are:

• Rajiv: modded Zyxel switch > modded sMS-200 > tX-USBultra
• Eric: modded Dlink switch > sMS-200ultra > dX-USBultra (and a Singxer SU-1)

 

So, the theme of the exercise this week is  - now that we have all this gear at our collective disposal, let's take the opportunity to answer some what-if questions!

 

We have already established on my chain that going from 2 reclockings (modded sMS-200) to 3 (add tX-USBultra) to 4 (add modded switch) each added sonic value, and staggering value, at that. Yes, this adds cost, and I'll come back to that. But the question that remains is - what lies beyond 4? 

 

Roy is going down the path of looking upstream, at reclocking his music server, and even his broadband router and modem. What our collection of gear allows us to do is to add a 5th and 6th reclocking, but still downstream of the music server. The purpose of this experiment is simple exploration. Whether or not the findings warrant additional spending is a completely orthogonal consideration.

 

Which brings us to cost. Let's consider a baseline that is quite typical for folks here at CA: a microRendu (+LPS-1) and an ISO-Regen (+LPS-1). Adding up the cost, this works out to about 1690 USD. Contrast this with what I consider the best value SOtM trifecta here - https://sotm-usa.com/collections/ultra-series-mods/products/sms-200-mod-tx-usbultra?variant=40555632908 - and add in a couple of LPS-1s to power. That comes out to 2540 USD.  This is a difference of 850 USD.

 

(Clarification: I know I have 3 LPS-1s in my chain, but I am finding the 3rd isn't strictly necessary, and I may repurpose it elsewhere.)

 

How much of an improvement (or how different) is the Ultra trifecta over the mR/IR baseline? I'll be able to study this when I receive my loaner IR in a few weeks, along with borrowing Eric's mR if he still has it. There are other experiments of interest with the IR, but from a cost perspective, this is the crux (pun intended) of the matter.

 

For me personally, it would take a significant SQ jump to convince me to spend more on this digital front end of my chain. While the outcome of this study will provide interesting insights, it may not correlate with where the next $500 or $1000 I spend on my audio chain will go. Should I spend more here, or say upgrading my DAC or amp? These are personal choices that we all have to make.

[austinpop]

May I humbly ask we not go down the path of "what part of my system should I upgrade next?" 

 

While this thread is hardly a model for staying on-topic,  a question like the above is so open-ended, it would just veer into chaos.

[austinpop]

21 minutes ago, greenleo said:

Rajiv,

Just to clarify:

You mean for equal SQ, LPS-1 for the modded switch is not necessary but necessary for the 200Ultra & the txUltra, right?

 

If so, this is a great news and money can be spent elsewhere.

 

I found that I could share an LPS-1 set at 7V (using a Y-cable) to power both my Zyxel modded switch, and the modded sMS-200. There is a slight increase in SQ by powering each with their own LPS-1, but not worth the $400 extra in my opinion.

 

Note: The above is true for a modded sMS-200, not for the sMS-200ultra, since I don't have the latter. If it's like the tX-USBultra, I suspect it draws close to the capacity of the LPS-1, as the LPS-1 powering my tX-USBultra gets pretty hot by itself.

[austinpop]

1 hour ago, greenleo said:

 Is it because the mobo of the clock is in the txUltra?  Maybe the users of 200Ultra may share their experiences.

 

Yes I think the sclk-ex board draws significant current.

[austinpop]

3 hours ago, Bamber said:

I haven't sent either to SOtM to use the extra clocks.  I think I need to invest in higher end power supplies because the SOtM Ultra combo doesn't sound any better than the mRendu/Recovery combo so far.

 

That is surprising and disappointing, I'm sure.

 

Sorry - I'm on a mobile device, so can't see your sig, where you may have this listed.

 

What are the PSUs you are using currently for:

- mR

- RUR (Recovery)

- sMS-200ultra

- tx-USBultra?

[austinpop]

23 minutes ago, Bamber said:

My 12v power supplies are all low end Teddy Pardo 12/2, McRU 12V and a Zero Zone 12V, at least compared to Romaz's PHs.

 

To complete the picture, what power supplies are you using with the mR and Recovery?

 

Regarding PSUs - in our tests this week, we should be able to validate if 2xLPS-1s in series, set to 7v and 5v respectively, can adequately power the 12v Ultra components. I don't see why not.

 

While I'm sure the SR7DRXL is in a class by itself, the 2xLPS-1 and the VR Mini's that @mozes uses are certainly fantastic alternatives, available within a week or two rather than several months.

 

Also - I know SOtM has mentioned (i.e. Lee to Roy) that the 12V versions sound better, but it would be nice to validate that. But you would really need 2 sMS or tX Ultra components side by side, where the voltage was the only difference, and a variable voltage power supply that could supply both ranges at equal quality.

 

33 minutes ago, Bamber said:

The SotM combo is more detailed but not natural sounding at this stage.

 

This certainly sounds like an artifact of the PSU.

[austinpop]

10 hours ago, limniscate said:

It has begun!

 

It's a mess!

 

It ended up being an 8 hour session, and we still have a lot of pending experiments. To be continued...

 

We have some initial findings, but too tired to write them up right now. Plus, we want to validate our findings in the light of a new day.

 

Stay tuned. 

[austinpop]

9 hours ago, hols said:

I would like to share my experience with the SOtM ts-USB Ultra with 10M reference clock input which has arrived a week ago.

 

...Cybershaft OCXO 10M reference clock...

 

Very cool stuff!

 

Do you know if the Cybershaft is available in the US?

[austinpop]

6 hours ago, Cornan said:

 

May I suggest that you try powering the switch with battery and the charger disconnected from AC mains. IME especially network switches sounds best with battery power. I strongly believe switches kick a lot of noises back into the AC mains. 

 

I'm working on writing up my results, but to this point - we did intend to try a battery, since we didn't have enough LPS-1's to go around. But sadly the battery turned out to be not sufficiently charged (operator error). So this will have to wait for another day.

 

But seriously - I have never found a battery to sound better than an LPS-1.

[austinpop]

2-day Ultra-vaganza Report

 

I see @limniscatehas already leaked some important results to the press. Typical lawyer move.  I, OTOH am a scientist™, so what follows is the summary report from our listening sessions. We conducted at least 14 distinct experiments, so I'll be reporting those here.

 

A couple of overarching comments. I am going to confine most of our findings to better, equal, or worse SQ, without trying to describe each improvement. Where possible, I'll add some terse adjectives to describe, but no more. Also, while Eric and I agreed on many things, we didn't always hear the same effects and the same magnitudes. So what I describe here are my impressions alone. Eric can jump in and add his views as he chooses.

 

With that out of the way, on to the experiments.

1. Compare mR/LPS-1 to sMS-Ultra powered by HDPlex 12V
   • No contest. The SQ improvement was yuge. 
   • Bigger, more dynamic, more resolution, better articulation.
      
2. Replace HDPlex 12V with 2xLPS-1s with serial cable to get 12V
   • Another large jump in SQ
   • blacker, more natural, bigger image
   • For the bulk of experiments below, only LPS-1s were used. Where necessary, we used a Y-cable or a serial cable.
      
3. Add in D-Link switch modded for sCLK-EX (referred to as D-switch in rest of report), powered by LPS-1/5V
   • Another huge jump in SQ.
   • We continue to be amazed at the benefit of an sCLK-EX modded switch!
   • Bigger, more natural, smoother, less fatiguing
      
4. Compare Rajiv's bifecta (?) of modded sMS-200 +modded Zyxel switch (referred to later as Z-switch) with Eric's combo of sMS-200ultra+D-switch
   • More or less a wash, with a slight preference for Rajiv's setup
   • Suspect it is the Z-switch sounding better than D-switch.
   • ASIDE: we were looking for a way to compare the efficacy of the SMB connectors vs. hardwired, but couldn't find a way to just isolate that. Eric's clock cables from Digikey were about the same length as Rajiv's hardwired cables (0.2 to 0.3m range)
   • Similarly, while we had no way to isolate just this variable, we saw no evidence that equipping the Ultra component as a 12V device was any better SQ than the standard 6.5-9V. When SOtM first proposed this option, their rationale was not SQ, but convenience for folks who already had 12V PSUs.
   • For me, the ability to power the Ultra components with LPS-1s makes them far more cost attractive. No need for super-expensive VR Mini or SR7's. Although - the SR4's that Paul just announced up-thread sounds rather intriguing!
      
5. Compare Z-switch with D-switch, both with sMS-200ultra
   • Z-switch by a nose
   • More boogie factor, PRAT, bit more dynamic
   • Remember, every piece is more than just a clock. The entire component design matters too. It looks like the Zyxel sounds a tad better.
      
6. Does using the same sCLK-EX board sound better/worse than separate boards?
   • Compare sMS-200 mod > tX-USBultra with sMS-200ultra > tX-USBultra
   • We could not hear any significant difference.
      
7. dX-USB HD Ulta > AES out vs. tX-USBultra > USB out
   • Baseline: Z-switch > sMS-200 mod > tX-USBultra
   • Comparison: Z-switch > sMS-200 mod > dX-USB HD Ulta
   • Different strengths;
     • dX: better articulation, microdynamics
     • tX: deeper bass, more weight and slam
   • If picking one, would pick the dX for the Yggy
      
8. Add 5th reclock - on USB
   • Baseline: Z-switch > sMS-200 mod > tX-USBultra
   • Comparison: Z-switch > sMS-200 mod > tX-USBultra > dX-USB HD Ultra
   • Definite SQ bump, but smaller in magnitude 
   • Heard combination of tX and dX strengths from experiment 7 results
   • Would not pay $1k for tX to get this small an improvement.
   • But maybe an IR? Intriguing thought - can't wait for my loaner from @bamber
      
9. Add 5th reclock - on Ethernet
   • Baseline: D-switch > sMS-200ultra > tX-USBultra
   • Comparison: Z-switch > D-switch > sMS-200ultra > tX-USBultra
   • Very small SQ bump
   • Definite case of diminishing returns
      
10. Add 6th reclock
    • Baseline: Z-switch > sMS-200 mod > tX-USBultra > dX-USB HD Ultra
    • Comparison: D-switch > Z-switch > sMS-200 mod > tX-USBultra > dX-USB HD Ultra
    • BUT: ran out of LPS-1s at this point, so had to power the D-switch (and the sMS-200ultra for the sCLK-EX) via 5V and 12V taps from the HDPlex
    • Not good. Sounded worse than baseline. Smearing, loss of detail and microdynamics
    • Conjecture: The HDPlex just ain't good enough for this.
    • Will retry this when the mighty SR7MR2DRXL arrives
       
11. Compare dX with Singxer SU-1
    • Baseline: Z-switch > sMS-200 mod > tX-USBultra > dX-USB HD Ultra (all LPS-1)
    • Comparison: Z-switch > sMS-200 mod > tX-USBultra > SU-1 (unmodded, built-in PSU)
    • Different strengths;
      • dX: more dynamic, larger
      • SU-1: more detail, slightly more natural and relaxed
    • Once again, a component is more that just the clock. Clearly, DDCs are not created equal, and the SU-1 shines.
    • I was really impressed with the SU-1. I am very curious to redo this comparison once Eric mods it and adds an LPS-1, as he plans. It would not surprise me to see it best the dX.
    • OTOH - what happens when the dX is fed a 10MHz master reference clock?! 
    • But purely bang for buck, the SU-1 is incredible.
       
12. Ethernet experiments - Impact of iSO-Cat6 isolator
    • Baseline: Z-switch > sMS-200 mod > tX-USBultra > dX-USB HD Ultra (all LPS-1) - same for all Ethernet experiments
      • server to switch: generic Cat6 cable
      • switch to sMS: BJC Cat 6a
    • Comparison: changed server to switch: generic Cat6 cable > iSO-Cat6 > BJC Cat 6a
    • Small SQ bump. 
    • I wouldn't pay $300 for this small an SQ change
       
13. Ethernet experiments - Impact of dCBL-Cat7
    • Baseline:
      • server to switch: generic Cat6 cable > iSO-Cat6 > BJC Cat 6a
      • switch to sMS: BJC Cat 6a
    • Comparison: changed switch to sMS: dCBL-Cat7
    • Noticeable SQ bump. Wow, this cable does sound better
    • I wouldn't pay $500 for this modest SQ bump
       
14. Ethernet experiments - Impact of full loom ($2600 worth of cables)
    • Baseline:
      • server to switch: generic Cat6 cable > iSO-Cat6 > BJC Cat 6a
      • switch to sMS: dCBL-Cat7
    • Comparison: 
      • server to switch: dCBL-Cat7 > iSO-Cat6 > dCBL-Cat7
      • switch to sMS: dCBL-Cat7 > iSO-Cat6 > dCBL-Cat7
    • Significant SQ bump. Wow, this full loom does sound good!
    • But still not the magnitude of adding the modded switch to the sMS-200ultra
    • $2600 for this improvement?  Lord have mercy!
    • No thanks!

Concluding Remarks

• The sMS-200ultra plus a modded switch is the lowest cost winner here. I think this is the true uniqueness of the sCLK-EX approach, as for whatever reason, previous standalone Ethernet reclocking switches, as reported by Ray, have just not yielded this much SQ improvement.
• For me, my trifecta (Zyxel modded switch, modded sMS-200, tX-USBultra) is still the sweet spot of bang for buck. I really lucked into this combo
• There is a definite case of diminishing returns by adding a 5th reclocking, both on the Ethernet and the USB side. Given this, I'm curious to see how impactful @romaz's sCLK-EX mods of his mobo, NICs, router, modem etc pan out.
• We need more data, so this is a conjecture only. I would hazard to say that the biggest bang for buck is by upgrading the system/USB/Ethernet clocks on the streamer, and reclocking once on Ethernet and USB. Subsequent reclockings become harder to justify in terms of price-performance.
• Once we get the Paul Hynes SR7, we can delve into the 6th reclocking to validate my conjecture of diminishing returns.
• The SU-1 really impressed me. I am intensely curious to see how much it improves with an LPS-1. This makes a lot of sense for the Yggy. I personally have no use for this device in my setup as
  • my Ayre Codex only has USB and optical inputs, and sounds way better with USB
  • about 25% of my library is DSD, so until I get a DAC that can accept DSD over AES, I'll be sticking with USB
• The fact that the SU-1 at the end of the chain still managed to sound so good could mean one of two things:
  • Roy's conjecture that clock chains should be in ascending quality up to the DAC has some unknown nuance to it, or
  • the clock implementation of the SU-1, using the Crystek 975 clock chip is close in quality to the sCLK-EX - a clock about which we have no technical info
• The above is bolstered by @mozes's findings with the IR in conjunction with the tX, which similarly held its own, and also employs the Crystek 975 as its clock chip.
• All switches modded for sCLK-EX and upgraded regulators/capacitors still sound different. To my ears the Zyxel sounded a tad better than the DLink, but my choice of it for modding was dumb luck!
• The dCBL-Cat7 cables are definitely da bomb, as Roy indicated. But here again, we are well into the realm of diminishing returns. And the cost is hard to justify - for me. 

It's great to see @hols's and @mozes's comments today (coincidentally) regarding the ISO-Regen and the Cybershaft reference clock. While I'm not sure when I'll get a chance to audition the latter, I should be receiving @Bamber's ISO-Regen/LPS-1 in a matter of a few weeks.

 

I'll post my teaser of IR experiments I plan to try a few days before results - as is my habit  - to whet your appetite. We will also be revisiting some of these experiments with the arrival of the Singxer SU-1 mods, LPS-1 to power it, and the SR7MR2DRXL.

[austinpop]

13 minutes ago, guillaume31 said:

Thanks for the detailed report @austinpop !!

 

Do you think the LPS-1 has enough oomph to power the 6.5V-9V sMS-200Ultra  ?

 

I cannot say authoritatively, as I have never directly used one, but I strongly expect so.

 

Works great with the tX-USBultra.

[austinpop]

3 minutes ago, rickca said:

 

I'm pretty sure this is just a typo, but the clock in the IR is the Crystek 575.  The ShenzhenAudio website description of the SU-1 says it uses Crystek CCHD-575/957 Series clock, so the SU-1 uses either the 575 or the 957.

 

I don't think there's any such thing as Crystek 975.

 

Yes - typo.

[austinpop]

8 hours ago, AmusedToD said:

 

I am convinced after your thorough comparisons. Thanks you so much for your time and effort!

 

I will definitely be sending my sms200Ultra for a switch mod. Could you please write down the exact model number of the Zyxel switch you were using in your test?

 

Welcome.

 

This is the switch. https://www.zyxel.com/us/en/products_services/8-Port-Desktop-Gigabit-Ethernet-Switch-GS-108B-v3/

 

Note it takes a 5V DC input, although for mine, I asked May to ensure the upgraded regulators would be happy at 7V too. 

 

I have no idea why the Newegg unit just referenced a couple posts up is showing 12V input. 

[austinpop]

7 hours ago, guillaume31 said:

Since we are in the "network bridge" post , how about the sMS-200Ultra in bridge mode versus the sMS-200Ultra connected to the modded switch ?

 

What I like in the modded switch is that in my set-up I could also connect my HTPC to it and get some benefits but on the other hand I've been trying to reduce the number of boxes and just got rid of my optical network

 

As mentioned, the purpose of the switch in this direct connection is purely to reclock. Have a look at my explanation of topologies - linked in index in the first post of this thread.

[austinpop]

5 hours ago, BigAlMc said:

Question. In either your lawyer or scientist capacities  do you feel that the switch/SMS combo in B upstream of the SU-1 still add sufficient SQ improvement to justify adding these? My SU-1 outputs I2S to my Directstream DAC so I'm tempted by these improvements but worried they'd get diminished as it's not USB going into the DAC.

 

I guess the other aspect I need to fully figure out is whether it's feasible (and viable in terms of finding someone to do it) to also replace the clock on the SU-1 with the sCLK-EX clock.

 

5 hours ago, AmusedToD said:

I already asked SOTM about modding the SU-1 and they said no.

 

1 hour ago, Bamber said:

Leter from Singxer responded they won't perform any mods.

 

1 hour ago, ElviaCaprice said:

I don't understand why anyone wants to replace the Crystek 575 in their SU-1.  This clock functions solely for the I2S output and from these testing results is as good or better than the Sclk EX, although they didn't even use the I2S output.

 

 

It is not surprising that manufacturers aren't enthused about modding their gear. SOtM is an exception in that they at least consider mods of non-audio gear, but I think we could all appreciate that manufacturers are busy implementing their roadmap of future products and have neither the time nor inclination to consider mods. What would be proper is to communicate customer expectations, feature requests, and try to raise the  priority of things like clock quality.

 

This whole exercise has led me to a new appreciation of the nuances of and complexities of clocking. While the experience with the sCLK-EX Ultra gear has been a revelation, it is also very encouraging to hear - second hand - about how good the ISO-R is, and first hand, the Singxer SU-1. I think this points to the fact that the eventual performance of a component is what I would term the "effective clock quality." What I mean by this is that sure, a killer clock chip or board can propel a system to better SQ, but the ultimate result is an aggregate of the entire system design. 

 

Nobody knows for sure the "raw" quality differences between the Crystek 575 clock and the sCLK-EX board. But that is academic. What matters is that in aggregate, how do components built using one or the other compare? I am not at all surprised by the quality of the ISO-R, as I am a huge fan of Alex and John. They have a gift for producing amazing gear at very attractive price points. Singxer is a new one for me, but I am similarly impressed.

 

I am really looking forward to trying the ISO-R. My selfish interest in my system is to:

• do a straight comparison by swapping it with the tX-USBultra
• determine SQ benefit from adding it in the chain upstream of the tX like @mozes and @hols

From a group perspective, I suspect there will be more than a passing interest in this config:

• modded switch > sMS-200ultra > ISO-R > modded SU-1 > DAC.

Fun times!

 

5 hours ago, BigAlMc said:

My wallet starts sweating nervously at this stage because if I could get the BigAlMc 'trifecta' of switch, SMS-200 & SU-1 all sharing the sCLK-EX clock then the case for adding a Ref 10 seems a lot stronger! (Rajiv - I hope I haven't infringed upon your 'trifecta' copyright here. Perhaps you can ask Eric as the lawyer ).

 

The possibility to further bump SQ with a high quality, but reasonably priced reference clock like the Cybershaft, or the forthcoming SOtM ref clock, is the next area of intense curiosity for me.

 

This is the one area where the dX with reference clock input could have an edge. But what if the next-gen Singxer also included a reference clock input? So use your voices as customers. Ask SOtM for an improved dX that better exploits the sCLK-EX (to my point about a DDC being about more than just the clock), and to support an I2S output. Ask Singxer to give you a next gen DDC that obviates the need for mods, and enables a ref clock input.

 

[austinpop]

1 minute ago, BigAlMc said:

I've discussed some of this stuff on the PS Audio forums as they have a great model where you can interact directly with the designers of the product (like Ted Smith for the Directstream DAC) and the CEO (Paul McGowan).

 

I've even shamelessly asked them when they'll produce an audiophile switch  and then enjoyed it when other PSA customers jump onto that and +1 it....

 

I'm a big fan of PS Audio, and consider the P5 AC regenrator in my system to be one of most significant upgrades I've made.

 

I do think we have to cut manufacturers a break. As consumers, we have the luxury to try all kinds of wacky mods and empirical improvements, but it's quite another thing for manufacturers to sell products - that they have to stand behind, and support - that bake in these ideas without a lot more testing and solid engineering.

 

On the flip side, I do find it disappointing when vendors (not naming names, but some even here on CA) start pooh-pooh'ing findings or bad mouthing competitors.

 

So it's two sides of the coin. The good manufacturers ones can navigate this stuff way better than the others.