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

[LTG2010]

21 minutes ago, mozes said:

Has anyone ordered a PSU from Sean Jacobs of Custom HiFi?

I'm waiting for mine, he's made all the modules, waiting for the case and transformers, this and following week, then put it all together.

[rickca]

6 minutes ago, LTG2010 said:

I'm waiting for mine, he's made all the modules, waiting for the case and transformers, this and following week, then put it all together.

Can you tell us more about your specifications?

[Kritpoon]

1 hour ago, LTG2010 said:

I'm waiting for mine, he's made all the modules, waiting for the case and transformers, this and following week, then put it all together.

 Can you share some details specifications of your PS order?  Is it multi-rails? Order lead time?

[mozes]

1 hour ago, LTG2010 said:

I'm waiting for mine, he's made all the modules, waiting for the case and transformers, this and following week, then put it all together.

Great pls let is know your impressions when you receive it. 

[LTG2010]

55 minutes ago, Kritpoon said:

Can you share some details specifications of your PS order?  Is it multi-rails? Order lead time?

@rickca & kritpoon,

Sure, its a custom design for my server, with two toroids and 9 rails (started of with 5 but you know how it is)

4 x atx rails ( 1x5v 5amp, 1x3.3v 1.5amp, 2x12v 1.5amp)

2 x 5v 1.5 amp (for os ssd & storage hdd)

3 x 9v 1.5amp for SOTM trio ( txUSBexp, sCLK ex & tx USBHubin)

Lead time is based on transformer delivery time. Sean offered 3 types of toroid, off the shelf as used in his DC3 PS, 2 week delivery next grade up, and 6 week delivery based on a specially chosen iron core and using a hand operated machine with very tight tolerences, 6 weeks basically because there is a waiting list/ queue. (same type as used in Zenith SE)

This week or the next is the 6th week so the toroids will be with Sean soon and he's made all the modules. The case was a special order to fit both toroids and big heatsinks for the regulators. There will be one umbilical for the cabling which I will feed into the server case with a gland and connect it up internally. If opting for atx rails you'll need some technical knowledge as Sean can't provide connectors for these rails due to an exclusivety agreement with Innuos, but its not dificult too look up and I'll post what ive done.

Cost is  £2530 I paid a deposit, will pay some more as soon as the transformers and case arrive.

48 minutes ago, mozes said:

Great pls let is know your impressions when you receive it. 

 Will do, as soon as I get it all going.

[One and a half]

11 hours ago, flkin said:

Ok, so here are the results of our SOtM Stable and Antipodes comparisons. Firstly I have to thank Khun Aek @Kritpoon for kindly coming over with his equipment and leaving it behind for a week. The world is truly small and to have two top tier front ends systems within a stone throw of each other in Bangkok is a nice surprise for us both. 

 

 

There are two schools of though about sources. One is to have an optimised computer with low noise, properly powered and even re-clocked perhaps feeding a DAC directly. The other is to use a simpler server of music like a Macbook or Mini and pass the signal through sophisticated rendering and cleaning systems. Both approaches have avid supporters and here we are looking at an instance of both approaches with interesting results.

 

As we have a lot of equipment and many various connections required to get each setup to work properly, I will first list (did it before but to summarise, here it is again) out the different equipment we had on that day of the test:

 

Source:
Antipodes running Roon Server
MacBook Pro running Roon Server 

 

The MacBook Pro is unoptimised (actually running as a Plex movie server for my kids as well) but bridged between an Orbi network WiFi secondary router (w/Teddy Pardo 12/2) that acts as the only data input point in the room and the SOtM clock modified D-Link DSG-105 switch.

 

 

Antipodes system:
- Antipodes DX Gen 3
- UltraRendu (LPS-1)
- Mutec Ref.10
- Mutec MC3+usb (mod for LPS-1 DC) w/Habst Digital BNC 5N Cryo Silver, 70ohm cable to the Ref 10
- BNC Coax cable Digiflex Gold II 

 

 

SOtM Stable:
- 2 SOtM dCBL-CAT7 ethernet cables and an iSO-CAT6 ethernet cleaner box
- The Trifecta consisting of the 3 items: D-Link DSG105 w/clock mod (LPS-1), sMS-200 w/clock mod (sPS-500), tx-USBultra (sPS-500) providing the clock connections
- IsoRegen (LPS-1)
Clock:
- Cybershaft OP14 clock w/ original Cybershaft 50ohm cable

 

 

Rest of my system:
Vinnie Rossi DAC 2.0 --> Vinnie Rossi DHT preamp (w/Takatsuki TA-300B tubes) --> YBA Passion 1000 mono blocks --> Wilson Watt Puppy 6 with the WP7 woofers 

 

Here are the tests we did and had 3 pairs of ears on them. We mostly agreed on the differences as they were not subtle. No blind tests were conducted. In all cases, the MC3+ was clocked by the Mutec Ref.10 via the Habst BNC cable.

 

          1. Cybershaft OP14 vs. Mutec Ref.10
          2. Antipodes to MC3+ vs. Macbook to SOtM Stable to MC3+
          3. Antipodes to MC3+ vs. Macbook to SOtM Stable
          4. Antipodes to SOtM Stable to MC3+ vs. All of the above
          5. Antipodes to SOtM Stable to MC3+ vs. Antipodes to UltraRendu to MC3+  
          6. Habst 75ohm cable vs. Pasternack RG216 75ohm cable 

 

 

1. Cybershaft OP14 vs. Mutec Ref.10
Khun Aek @Kritpoon brought over his equipment over the day before and so it would be warming up for around 30 hours prior to our tests. We figured this should be enough for the clock and other components to sound at it’s best.

 

And of course, I couldn’t help but to have a listen the night before the tests alone after around 8 hours of warming. This was probably the simplest test to perform as I had my SOtM Stable already set up and it was only a matter of switching the Cybershaft 50ohm cable from the OP14 to the Ref.10 50ohm connector. 

 

 

With the track Live from Las Vegas, Louis Prima & Kelly Smith track 15, Embraceable you, the Ref. 10 can be heard as more natural with more focus. More separation with the instruments and voices without being thinner. The separated sounds have a touch better textures and more details, the combined effect is that the sonic picture is more real.

 

 

With Breathe by Jax Jones, the Ref10 sync beats more clearly focused in space, tight and bouncing around the soundstage. Voice more texture and natural. Who would have known a club/dance electronic pop song like this can have vocals that sound audiophile?!

 

 

Tower of song by Leonard Cohen, voice larger with OP14 made it more enjoyable to listen to but perhaps less accurate. 

 

 

Tears in Heaven, Eric Clapton Unplugged. OP14 looser imaging and each sound image larger. Fun to listen to but Ref10 more focused and more accurate. Overall balance of sound comes across as balanced and complete, voices correct sizes detailed and textured, guitars clear and defined, backing vocals gently at the back left. The thumping bass is resolved enough to imagine heel stamping and you can hear the shoe tapping on the floor too. Nice!

 

That said, it’s not as if the Cybershaft OP14 is a slouch, it’s already very good and the best my system had ever been prior to listening to the Ref.10. The improvement over the clock in the tX-USBultra is substantial and well worth the upgrade. It’s just that the Mutec Ref10 provides even more of a ‘clock’ effect. A little bit better in all aspects. In addition to lower a phase noise in the Ref.10, I wonder if that has to do with their use of a square wave form instead of the sine wave in the OP14? 

 

The next day when my guests arrived, I suggested that we used just the Tears in Heaven, Eric Clapton Unplugged track as our test track. It’s an excellent sounding track with guitar strings, vocals in a live environment. Everyone arrived at the same conclusion which was the external clock matters and greatly so with the Ref 10.

 

From this point on, we decided to stick with the Ref10 clock for any external clocking requirements.

 

 

2. Antipodes with MC3+ vs. Macbook through SOtM Stable with MC3+
Next up a pretty obvious comparison was to use the Antipodes by itself to my Vinnie Rossi DAC 2.0. Unfortunately we couldn’t get the Antipodes to recognise the DAC 2.0 and eventually we tried passing the Antipodes through the MC3+ (Ref10 reclocked with Habst cable) and using a coaxial (Digiflex Gold II) to the DAC2.0 and this worked fine. So the rest of our comparison tests had to pass through the MC3+.

 

The first surprise of the day - the Macbook setup sounded better. Same Eric Clapton track was used. The Antipodes setup sounded thicker and less airy between the soundscape. Eric Clapton’s voice with less detail and emotion.

 

We had expected that the Antipodes with it’s specially customised Asus motherboard and optimised power source would be better but we didn’t hear it that way. Could it be that the SOtM dCBL-Cat7 cables and iSO-Cat6 device from the Macbook were doing it job and providing the Stable with a sufficiently cleaned signal to best the Antipodes? Looking at the wired connections more carefully, we noticed that the Antipodes was connected to the Orbi router with a standard Blue Jeans ethernet blue cable. I had an unused Ghent JSSG cable lying around and so tried it. Ah and there’s the difference, once the cable was swapped, the Antipodes’ sound noticeably tightened up. Now the sound from both systems were about matched. *

 

* I revisited this setup after the meet by myself as it's pretty unexpected and came to the conclusion again. 

 

This indicates to me that in my particular setup the server is not as critical as the renderer part of the source. A humble Macbook Pro (with power adaptor plugged in, eek!) and not software optimised still sounded comparable to a top quality server. But to achieve this it needed to be software bridged along with, I suspect, the triple cleaning effect of the SOtM ethernet filters to removed whatever noise that shouldn’t be there.

 

 

3. Antipodes with MC3+ vs. Macbook with SOtM Stable
Ok, then how about we remove the MC3+ from the Macbook setup? The sound becomes thinner and not as engaging as with the MC3+ in the path. It looks like the tx-USBultra, while doing a good job cleaning the USB signal, as is confirmed by so many others, can still be improved upon by yet another USB cleaner after it. Khun Aek informs me that up to 4 MC3+ devices can be chained together to further clean a signal noticeably! 

 

 

4. Antipodes with SOtM Stable with MC3+ vs. All of the above
So we established that the Stable with MC3+ added for double cleaning is necessary but what about adding the Antipodes as the source instead of the Macbook? This we haven’t tried yet as the (3) comparison was with the Antipodes with the MC3+ alone. And here the Macbook/Stable/MC3+ matched it.

 

 

Yes, this was the optimal configuration. The best of the lot. Beautifully natural sound, balanced of bass, voice details and texture with an open airy sound. There is no noticeable emphasis in the music, all parts of Eric Clapton’s ensemble are playing/singing in what feels like great teamwork, a musical sync without one aspect overwhelming another. Depth, width and even height of the music nicely rendered. When you hit a sweet spot, wow! The emotion really comes through. The differences of the various combinations of the equipment may not be all that much measured by percentages but when measured by the emotions of the musicians bared so clearly to us and the effect it had on us, it’s truly a 'goosebump' moment as Khun Aek put it. 

 

 

5. Antipodes to SOtM Stable to MC3+ vs. Antipodes to UltraRendu to MC3+  
Khun Aek recently discovered that his bridged Antipodes sounded better with the UltraRendu added as the renderer. It was the best his system sounded prior to coming over, so we decided to see how the UltraRendu would sound by swapping it in to replace the Stable.

 

Nope, the soundstage collapses quite a bit, several steps behind the last iteration in (4). Vocals are also slightly recessed. In my particular setup the Antipodes played through the Stable and extra cleaned with the MC3+ has some truly special synergy and here the UltraRendu cannot replace the Stable.

 

 

6. Habst 5N cryo silver 75ohm cable vs. Pasternack PE33403-36, RG216 75ohm cable 
Given the silver Habst cable was so expensive and since I had a new Pasternack RG216 double shielded cable around, we though it would be interesting to see how much of a change swapping the cable would show. This is from the Ref10 to the MC3+. The Stable was still connected to the Ref 10 via the Cyershaft 50ohm cable. 

 

So using the best sounding combination of equipment in (4) we tried this. Amazing change in sound. The Habst is far better. With the Pasternak the sound becomes unfocused, the bass and lower mids a smeared blob. The difference was almost as large as not using a clock. So the clock BNC cable matters - to the extent that an external clock is basically pointless without a good cable.  So much so that I am considering seeing if Habst can make up a 50ohm cable for my Cybershaft.

 

The Habst cable has 2 BNC connectors on the Ref 10 side - a thicker twin cable for signal and a thinner single for ground. We decided to fit the Habst back to the MC3+ but disconnect the ground BNC alone. With just the signal cable, the Habst is good but not great. The grounding BNC is essential but takes an additional port on the Ref.10

 

 

At this time, we decided to lay back and enjoy music for the rest of our meet. Some of the other tracks we listened to were:

 

• Livingston Taylor, Isn’t she lovely
• Nojima Plays Liszt (Minoru Nojima), Etude No3, La Campanella
• Jax Jones, Breathe
• Hilary Hahn/ Natalie Zhu, Mozart: Violin Sonatas K301
• La La Land (original motion picture soundtrack), Someone in the Crowd
• Imagine Dragons, Evolve, Believer

 

 

Man, I heard things I never heard before. In the Liszt track, the piano notes and pedal movements has always been fantastic sounding but for the first time, I could see the piano size and shape. This is one of my favourite piano tracks with beautifully captured attacks and ambience. Norjima's playing here is OK but makes 4 mistakes are even clearer now that I can count. ;-)

 

 

Imagine Dragons just came to Bangkok recently and with that high energy Believer track, they were back in town in my hifi room and sounded like we were on stage with them; soundstage width, depth and height extending far beyond the physical boundaries of my room.

 

 

And so after our meet, I can say I've reconfirmed some of my earlier thoughts about equipment contributions but perhaps reordered their position in the significance table. To summarise our findings :

 

• Clocks and clock cables - they matter greatly perhaps the most
• USB cleaners, more than one in series works and significantly
• Much can be done to improve the Source. The renderer part more so than the server part
• The Macbook Pro makes is a great Roon server when bridged and ethernet output cleaned
• But the best sound is to have an excellent source and further re-clock it and clean the output multiple times.

That’s it, hope this is interesting reading for all. 

 

 

A big thank you @flkinfor a detailed review of different types of signal treatment. Magazine reviewers are far too kind and tend to gloss over details which are important, I prefer the user reviews and epsecially where consideration of the system is in play. The clocks especially and the cables are an eye opener, and even more the source itself, which was pleasing for the wallet and the ear  

I did have a choice on the Antipodes, but decided to make my own server instead from regular computer parts, the process started earlier this week. There are a few treatments I would like to share once all the parts arrive, but won't be for at least a month or more. Thank you again for the write up and the experiences you found, well done.

[lmitche]

FYI - I just posted part 2 of the custom motherboard exercise which can be found here:

 

[austinpop]

3 hours ago, rickca said:

Here is the explanation.

 

Next up a pretty obvious comparison was to use the Antipodes by itself to my Vinnie Rossi DAC 2.0. Unfortunately we couldn’t get the Antipodes to recognise the DAC 2.0 and eventually we tried passing the Antipodes through the MC3+ (Ref10 reclocked with Habst cable) and using a coaxial (Digiflex Gold II) to the DAC2.0 and this worked fine. So the rest of our comparison tests had to pass through the MC3+.

 

Understood, for the direct attach case, from DX-3 to the DAC 2.0. In fact, this is painfully reminiscent of the problem I had with the Zenith SE handshaking with my DAC.

 

However, the MC-3+ ought not to be necessary between the tX-USBultra and the VR DAC 2.0.

[austinpop]

Also, @flkin and @Kritpoon - as I read your write up, the other scenario I would ask you to try is to use the Antipodes with a partial SOtM stack like Eric & I did as indicated below. In all cases, with the Ref 10 clocking the tX and the MC-3+.

 

- Antipodes > tX-USBultra > MC-3+ > DAC.

 

How does this compare to:

 

- Antipodes > modded sMS-200 > tX-USBultra > MC-3+ > DAC

 

and

 

- Mac Mini > modded sMS-200 > tX-USBultra > MC-3+ > DAC?

 

In other words, does removing the modded sMS-200 from the chain improve or hurt the SQ?

 

This is your follow up homework!  

[rickca]

1 hour ago, austinpop said:

However, the MC-3+ ought not to be necessary between the tX-USBultra and the VR DAC 2.0.

I agree and it would be nice to get it tested.  I think the MC-3+ USB was in all the tested scenarios to keep 'other things being equal'.

 

So you're asking for 

Antipodes > tX-USBultra > DAC

[rickca]

3 hours ago, LTG2010 said:

9 rails

One less and we could have named it the Octopus!  What's your rationale for the amps on the ATX rails?  Especially the 12V rails ... 1.5A seems awfully light.  How are you going to power the CPU (and what motherboard/CPU do you intend to use this PS with)?  How are the rails split between the two transformers?  Thanks.

[LTG2010]

4 minutes ago, rickca said:

One less and we could have named it the Octopus!  What's your rationale for the amps on the ATX rails?  Especially the 12V rails ... 1.5A seems awfully light.  How are you going to power the CPU (and what motherboard/CPU do you intend to use this PS with)?  How are the rails split between the two transformers?  Thanks.

I've got a couple of boards and the supply is meant to be flexible to suit a few lower power boards along with SOTM's boards. Currently i'm using a Supermicro X10SBA. The processor is a celeron j1900 and only needs 10 or 12 watts. There's a 24 pin atx connector and a 4pin 12v atx connector, this (4 pin) supplies the processor when in atx mode, this is what I understood when discussing it with Supermicro's tech help.

Sean's designs use a booster module to increase power from 1.5 to 5 amps we only need one for the 5 v rail on this board, since there's a bit more going on at 5v.  With the 2 transformers one will have the 2 x 12v rails, the other one the rest. I think it best to isolate the CPU from the other components. The more rails the cleaner the supply to each board / component.

Currently I'm powering it with just 12v to the board with a meanwell medical grade smps, 9v to the SOTM boards from a sPS500 via the txUSB-exp and 5v to the ssd from an lps1, the hdd powered from the board. (3 supplies)

May at SOTM felt a good supply to the txUSB-exp is filtered and passed on to the sCLK-ex and again filtered & onto the txUSB Hubin, but I think 3 seperate rails will boost their performance. I'll be able to test this out in 2 or 3 weeks.

[austinpop]

3 hours ago, rickca said:

So you're asking for 

Antipodes > tX-USBultra > DAC

 

Yes, compared to:

• Antipodes > tX-USBultra > MC-3+ USB > DAC

[BigAlMc]

7 hours ago, One and a half said:

big thank you @flkinfor a detailed review of different types of signal treatment. Magazine reviewers are far too kind and tend to gloss over details which are important, I prefer the user reviews and epsecially where consideration of the system is in play. The clocks especially and the cables are an eye opener, and even more the source itself, which was pleasing for the wallet and the ear 

 

Hi @One and a half I agree in general and especially around magazine reviewers who seem obliged to write reviews saying everything is great.

 

Not so sure I agree about pleasing for the wallet tho.  

 

Awesome work and review by @flkin and @Kritpoon - truly great stuff.

 

But in terms of my wallet, one my assumptions has been shattered. I was assuming up until now that it was kind of a choice between SOtM Trifecta (or similar endpoint based stack) versus high end server such as Antipodes Gen 3 DX or Innuos Zenith SE. 

 

The conclusions from the Bangkok listening sessions seem to say:

 

Trifecta stack AND kickass server is better than the kickass server on its own. 

 

Trifecta stack AND external clock AND kickass server is better still. 

 

My wallet is taking little comfort in these findings and the apparent conclusion that adding a clock to my stack is recommended. I assumed it was job done when I invested all that money in my Zenith SE. Sigh.... 

 

Great stuff nonetheless guys, even if my wallet is giving you both dirty looks! 

 

Cheers, 

Alan 

[flkin]

19 hours ago, zoltan said:

@flkin Wow! Great system! Thanks for the review. I wonder how hi-fi magazines survive if we can read such fantastic reviews online. 

My front-end is very similar to what you found best, except the Antipodes and power supplies. I use a passive PC (in Streacom DB4) Win 10 I built myself. It is not optimized so my conclusion from your review is that perhaps it is something worth looking at. At the same time, given the price of an Antipod, do you think that difference was significant enough (compared to a Macbook) to justify $7500? 

 

I have also experimented with some network isolation to avoid the need to optimize the server. I have a few Acoustic Revive RLI-1s that work nicely but the best so far was a FMC between the SOtM modded switch and the SMS-200 ultra where the receiving side is powered by the SPS-500. It seems to isolate noise prior in the chain so well that putting the D-Link switch back to its original SMPS power supply didn't even matter.  

Recently, I friend brought over two more MC3+USBs and thus we cascaded three. It is an incredible difference compared to one and I am torn between getting one more MC3+USB or a Habst clock cable, nearly the same price. I am just going to order Canare clock cables that are optimized up to 12GHz https://www.sportsvideo.org/2017/04/25/nab-2017-canare-intros-12g-coax-for-4k-uhd-television-signals/
I wish I (or anybody else) had a chance to test this against the Habst. Speaking of the Habst, I technically can't understand what that separate gound does. A digital 75 ohm signal is supposed to be coax in geometry and have a 'signal' as the centre core and 'shield' as gound. Does anyone have a clue how it works in Habst?

 

Lastly, can I ask what software you used and what settings? Upsampling? Filters? Dither? etc,
 

 

Wow, thanks for your compliment, this is my first attempt, was trying not to come across as too much fluff and not enough data while remaining entertaining. 

 

In my particular setup, I've tried to get the renderer and cleaning side well setup so it would require a less optimised server. Looks like this approach works, but it seems there is still benefit to a better server. I probably wouldn't spend on an Antipodes for the server but a little server work may go a long way to getting that additional benefit I heard during our tests. 

 

Agreed, cascading cleaning devices seems to be beneficial so long as the distortions from cable connections or additional power supplies (what yet another one?!) don't make the overall worse. That's something to consider too I suppose.

 

My Vinnie Rossi DAC 2.0 uses the AKM AK4497EQ chip has a zero filtering setting and allows the chip to read and play what was sent to it verbatim. And from Roon, I have found that zero upscaling sounds the most natural and offers the largest soundscape in all cases. I have tried both Roon upscaling and HQPlayer with all their myriad of settings. Ultimately I still prefer to play the files as they come. Re-clocking is a good way to get the effect of upscaling - I find similar benefits of focusing the sound but doesn't make the images thinner nor shrink the soundscape while doing so.

 

 

14 hours ago, austinpop said:

What an excellent review, @flkin, in the finest tradition of this thread! I am envious of the collective quality of equipment you had on hand, although perhaps less so than others, as we have a lot of this gear in Austin as well!

 

In general, I am glad to hear that your findings are quite consistent with @limniscate's and mine, and many others on this thread. Of course, the general finding is that everything matters - clocks, PSUs, cables, and everything else. But more specifically, it points to an approach with clocking that I really hope manufacturers are paying heed. And that is:

1. pay attention to the quality of both your word and your system clocks
   • consider a design that enables an external reference clock like the Ref 10, Cybershaft et al.
     • My shallow understanding of clock design suggests that this means a frequency synthesizer approach to internal clocks, rather than fixed frequency clocks
     • However, perhaps there are other ways to do this - I don't know
   • Bottom line, I would love to see components that both accept external reference clocks, and apply this reference to both word and system clocks. 
2. provide the ability to override the internal PSU with an external DC supply
3. Imagine an end to end chain that could be natively (i.e. without mods) clocked by a reference clock, and powered by the best PSUs

Let me comment on a few things, specifically.

 

 

I noticed you never tried running USB direct from the tX-USBultra to the DAC 2.0? Is there a reason? Or did I just miss it in the write up? Is the S/PDIF input of the DAC 2.0 known to be much better than USB?

 

I say this because I learned from um, a friend, that he at one point tried the MC-3+ USB between his tX-USBultra and his Blu2/DAVE, all clocked by the Ref 10, and he found that the MC-3+ USB actually detracted from the SQ. Of course, different DAC, and his MC-3+ was not modded for LPS. Anyway, I was just curious. This is the same reason I asked @Confused to try going direct to his Devialet.

 

 

Well, this is timely, as @limniscate are going to do the Ref 10 vs. OP-14 shootout at my place in a few days. Now I really look forward to this comparo!

 

 

Sigh, I was afraid of this! Please let me know what you learn from Habst about the 50 Ω cable. Since the Cybershaft has just the one BNC, I am not sure how the separate ground concept is possible.

 

 

As I've pointed out in a recent post, I think the lines between "server" and "renderer" are quite blurry, and best not to fixate on function. See: https://www.computeraudiophile.com/forums/topic/38110-innuos-zenith-se-mkii/?do=findComment&comment=786922

 

IMO, the quality of the device that starts the USB chain seems to be of extreme importance. Getting the clocking, PSU, and overall HW design is critical. I think looking forward, it will be nice to work towards filling in the obvious information gaps:

1. Does the trifecta approach regain ascendancy when powered by SR-7 grade PSU?
2. How does the Antipodes DX3/Zenith SE compare to bespoke custom sCLK-ed servers like @romaz, @ElviaCaprice, and @Johnseye's? Apologies if I missed others.

 

It certainly was interesting reading. Thanks again for taking the time!

 

Thank you @austinpop for your kind comments. 

 

We were having trouble getting the Antipodes to connect to the VR DAC2.0 and in order to proceed we had to pass the signal through the MC3+ for the connection. This made it necessary to follow the same route to continue the comparisons. In test (3), I did a direct Mac/Trifecta tx-USBultra connection to the DAC via a Lush USB and found the sound to be as good as the Antipodes/MC3+ connected via the Digiflex Gold II coax. Actually even better until a cable swap improved the Antipodes and matched the SOtM equipment. Perhaps I should revisit this and see if I can get the MC3+ connected to the DAC2.0 via USB and listen again. I'll have to ask Khun Aek @Kritpoon for help with this as I'm not familiar how the MC3+ settings work. Also I should try just this with more tracks as our comparisons only used a single one due to time constraints.

 

Look forward to reading your findings about the Ref10 vs OP14. For us it was a clear step up from what was already a good place. But at double the price it's as it should be.

 

I did try using the Habst 75ohm cable from Ref10@75ohm to the tx-USBultra and guess what, it works fine! Didn't connect the ground connector and only used the signal one. And sounded fine too, better than the standard Cybershaft black cable. More refined and clearer placements of instruments and a touch more width to the stage. I wonder if one can hear that much difference between 50 and 75ohm cables? But for sure there is a large difference between the Habst (0.5m) and the Pasternack RG216 (36 inches). Unfortunate but true.

 

As for how a Paul Hynes would affect the Trifecta - and I'm of the opinion it will drastically, well I can answer that in about 6-8 months if I'm lucky...   queue #29

 

 

9 hours ago, One and a half said:

A big thank you @flkinfor a detailed review of different types of signal treatment. Magazine reviewers are far too kind and tend to gloss over details which are important, I prefer the user reviews and epsecially where consideration of the system is in play. The clocks especially and the cables are an eye opener, and even more the source itself, which was pleasing for the wallet and the ear  

I did have a choice on the Antipodes, but decided to make my own server instead from regular computer parts, the process started earlier this week. There are a few treatments I would like to share once all the parts arrive, but won't be for at least a month or more. Thank you again for the write up and the experiences you found, well done.

 

Thanks for your comments. 

 

Kin

 

 

 

[flkin]

8 hours ago, rickca said:

I agree and it would be nice to get it tested.  I think the MC-3+ USB was in all the tested scenarios to keep 'other things being equal'.

 

So you're asking for 

Antipodes > tX-USBultra > DAC

5 hours ago, austinpop said:

 

Yes, compared to:

• Antipodes > tX-USBultra > MC-3+ USB > DAC

 

Ok, that's an interesting test, I'll have a look. Will it matter to you if I keep the mod switch there? I'm not sure if I can remove the mod switch easily as the cabling is rather tight and short and SOtM's stiff cables are a real bi*ch to deal with 

 

1. Antipodes > mod switch > tX-USBultra > DAC
2. Antipodes > mod switch > tX-USBultra > MC-3+ via USB > DAC
3.           Mac > mod switch > modded sMS-200 > tX-USBultra > MC-3+ via USB > DAC

 

1. May or may not be possible since Antipodes doesn't recognise the DAC2.0. Not sure if the tX-USBultra can help.
2. Should be doable 
3. Should be doable

 

2 hours ago, BigAlMc said:

 

Hi @One and a half I agree in general and especially around magazine reviewers who seem obliged to write reviews saying everything is great.

 

Not so sure I agree about pleasing for the wallet tho.  

 

Awesome work and review by @flkin and @Kritpoon - truly great stuff.

 

But in terms of my wallet, one my assumptions has been shattered. I was assuming up until now that it was kind of a choice between SOtM Trifecta (or similar endpoint based stack) versus high end server such as Antipodes Gen 3 DX or Innuos Zenith SE. 

 

The conclusions from the Bangkok listening sessions seem to say:

 

Trifecta stack AND kickass server is better than the kickass server on its own. 

 

Trifecta stack AND external clock AND kickass server is better still. 

 

My wallet is taking little comfort in these findings and the apparent conclusion that adding a clock to my stack is recommended. I assumed it was job done when I invested all that money in my Zenith SE. Sigh.... 

 

Great stuff nonetheless guys, even if my wallet is giving you both dirty looks! 

 

Cheers, 

Alan 

 

hahaha, just think of the savings you get from concerts you won't have to go to. 

 

But seriously, we didn't hear the Zenith SE and it's supposed to be near as good as it gets. So unless you get a case of "clockatitis" perhaps the job is done. 

 

Kin

[BigAlMc]

23 minutes ago, flkin said:

hahaha, just think of the savings you get from concerts you won't have to go to. 

 

But seriously, we didn't hear the Zenith SE and it's supposed to be near as good as it gets. So unless you get a case of "clockatitis" perhaps the job is done. 

 

Hi @flkin - yeah you are probably right and I should stop the foolishness and enjoy the amazing SQ I have.

 

But this hobby (or should that be sickness) being what it is there is a tiny cog in my tiny mind wondering! 

 

So I recently added a TX-USBultra to my SE because the likes of Roy, Rajiv and Eric all said it improved the USB output from the SE further. And it did! 

 

I then swapped the LPS-1 powering the TX-USBultra for an LPS-1.2. And guess what. It improved again. 

 

So the direction of travel is looking like a clock shaped rabbit hole! 

[LTG2010]

1 hour ago, flkin said:

Ok, that's an interesting test, I'll have a look. Will it matter to you if I keep the mod switch there?

Firstly thanks for the detailed write up, especialy listing all your musical choices.

I think the key here is how good is the Antipodes USB output? Keeping the switch should not matter.

How good is the USB out when it is re-clocked by the txUSB Ultra. Can the Mutec components be added to this re-clocking chain somehow, and then on to the DAC?

The ultimate question is then which sounds better, ethernet out to Ultrarendu and or the trifecta/ reclocking chain?

Or USB out to reclocking chain?

Some of us have ditched our renderers for USB, do we need to go back to ethernet now? it never ends

[magnuska]

2 hours ago, flkin said:

 

Wow, thanks for your compliment, this is my first attempt, was trying not to come across as too much fluff and not enough data while remaining entertaining. 

 

In my particular setup, I've tried to get the renderer and cleaning side well setup so it would require a less optimised server. Looks like this approach works, but it seems there is still benefit to a better server. I probably wouldn't spend on an Antipodes for the server but a little server work may go a long way to getting that additional benefit I heard during our tests. 

 

Agreed, cascading cleaning devices seems to be beneficial so long as the distortions from cable connections or additional power supplies (what yet another one?!) don't make the overall worse. That's something to consider too I suppose.

 

My Vinnie Rossi DAC 2.0 uses the AKM AK4497EQ chip has a zero filtering setting and allows the chip to read and play what was sent to it verbatim. And from Roon, I have found that zero upscaling sounds the most natural and offers the largest soundscape in all cases. I have tried both Roon upscaling and HQPlayer with all their myriad of settings. Ultimately I still prefer to play the files as they come. Re-clocking is a good way to get the effect of upscaling - I find similar benefits of focusing the sound but doesn't make the images thinner nor shrink the soundscape while doing so.

 

 

 

Thank you @austinpop for your kind comments. 

 

We were having trouble getting the Antipodes to connect to the VR DAC2.0 and in order to proceed we had to pass the signal through the MC3+ for the connection. This made it necessary to follow the same route to continue the comparisons. In test (3), I did a direct Mac/Trifecta tx-USBultra connection to the DAC via a Lush USB and found the sound to be as good as the Antipodes/MC3+ connected via the Digiflex Gold II coax. Actually even better until a cable swap improved the Antipodes and matched the SOtM equipment. Perhaps I should revisit this and see if I can get the MC3+ connected to the DAC2.0 via USB and listen again. I'll have to ask Khun Aek @Kritpoon for help with this as I'm not familiar how the MC3+ settings work. Also I should try just this with more tracks as our comparisons only used a single one due to time constraints.

 

Look forward to reading your findings about the Ref10 vs OP14. For us it was a clear step up from what was already a good place. But at double the price it's as it should be.

 

I did try using the Habst 75ohm cable from Ref10@75ohm to the tx-USBultra and guess what, it works fine! Didn't connect the ground connector and only used the signal one. And sounded fine too, better than the standard Cybershaft black cable. More refined and clearer placements of instruments and a touch more width to the stage. I wonder if one can hear that much difference between 50 and 75ohm cables? But for sure there is a large difference between the Habst (0.5m) and the Pasternack RG216 (36 inches). Unfortunate but true.

 

As for how a Paul Hynes would affect the Trifecta - and I'm of the opinion it will drastically, well I can answer that in about 6-8 months if I'm lucky...   queue #29

 

 

 

Thanks for your comments. 

 

Kin

 

 

 

 

 

Hello!  Many thanks for your write up and  testing! 

 

I also use the VR Lio with DAC 2.0 and am a little concerned that it would not see the Antipodes DX-3. I am now using the sotM sms-200 ultra with Iso Regen direct to the DAC. My hope was to be able to slim down the nr of boxes and was thinking of the new Antipodes EX acting as the renderer. My current server will also be used with the new direct ethernet connection.  Making it a kind of budget CX-EX solution.

Wonder why Antipodes seem to focus on noise everywhere and starting on the MOBO - no advanced clocking there. It seems that they are working hard to remove noise before at its core and therefore minimising the need of better clocking but this is my own thinking.

 

 

[One and a half]

Please keep in mind the MC3+USB can be further improved by purchasing a DC only version at the time of order.

 

The meanwell smps that’s fitted in the Mutec reduces the sq of the unit. The main problem of the DC version is to maintain 6V and not exceed 6.2V. This is a challenge in itself.

 

My theory is you need more MC3 to squash the noise from the meanwell smps. Go DC and you can save a stage.

[Confused]

3 hours ago, flkin said:

I did try using the Habst 75ohm cable from Ref10@75ohm to the tx-USBultra and guess what, it works fine! Didn't connect the ground connector and only used the signal one. And sounded fine too, better than the standard Cybershaft black cable. More refined and clearer placements of instruments and a touch more width to the stage. I wonder if one can hear that much difference between 50 and 75ohm cables? But for sure there is a large difference between the Habst (0.5m) and the Pasternack RG216 (36 inches). Unfortunate but true.

 

One point regarding the Habst vs RG216 comparison.  My understanding is that the REF10, in particular due to it's square wave output, is very sensitive to cable quality and cable length.  So in this comparison, the Habst has the advantage of being the shorter cable. (0.5m vs 0.9m for the RG216)

 

This makes me wonder, how much of the improvement is due to the Habst cable itself, and how much due to the shorter length.  My hunch is that the cable length is less significant in this case, but you never know.  This begs the question, has anyone tried different lengths of identical cable between the REF10 and say an MC3+USB?  Personally, I have a whole list of things I need to try, but I do have RG216 cables in both 12" and 24" lengths, so this is something I can try once I've sorted my immediate niggles out.

 

Also, how much of an influence does the Habst ground connector make?  Are there better cables than the Habst that would actually not perform as well overall because the Habst ground connector is making a big influence?

 

 

[zoltan]

23 minutes ago, One and a half said:

Please keep in mind the MC3+USB can be further improved by purchasing a DC only version at the time of order.

 

The meanwell smps that’s fitted in the Mutec reduces the sq of the unit. The main problem of the DC version is to maintain 6V and not exceed 6.2V. This is a challenge in itself.

 

My theory is you need more MC3 to squash the noise from the meanwell smps. Go DC and you can save a stage.

Where did you see such an option? Can you share a link? I have removed the SMPS myself but I am not aware of any other option than doing it DIY.

[zoltan]

20 minutes ago, Confused said:

Also, how much of an influence does the Habst ground connector make?  Are there better cables than the Habst that would actually not perform as well overall because the Habst ground connector is making a big influence?

 

 

3

Can somebody point out what this ground connector does?? A clock signal has the 'hot' and the 'ground' both of which are already connected to the BNC plug of a Habst cable each end. This extra ground baffles me. 
 

[rickca]

2 minutes ago, zoltan said:

Can somebody point out what this ground connector does?? A clock signal has the 'hot' and the 'ground' both of which are already connected to the BNC plug of a Habst cable each end. This extra ground baffles me. 
 

Yes also please explain what does this ground connector connect to at each end?

[mozes]

Just a heads up for the followers of this thread. I will soon be listing some goodies including a custom build Nimitra and 3 LPS-1s.