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

bit01 · August 22, 2019

I have done a few permutations.

server/endpoints - PC/mR+IR/uR NUCs+Akasa fanless, Win7/AL JRiver/LMS/Squeezelite/Roon, various LPSs

big hits - ethernet bridging, good power supplies, AL RAMroot/extreme

as for best all around b4b for me (the magical spot for sound, musicality & $)- NUC playing local files:

J River MC25/AL LXQT3x RAMroot/Extreme (16GB RAM)> NUC7i7DNH1E + WD myPassport local music files, PH SR4 with custom DC cable & 3rd party AC cord.

Superb, and thanks to all contributors to AS and especially to those champs of this amazing thread!

sig8 · August 22, 2019

On 8/20/2019 at 6:20 PM, austinpop said:

Hi all,

It seems this thread has gone quiet in terms of posts by some of the "regulars" who have followed it from the beginning. My assumption, or at least hope, is that people are just happy with their systems now, and hence have no reason to post.

So as a way to revitalize this thread, I have a proposal. Please consider posting a description your system in its current state. If possible, please showcase which of the optimizations we've covered in this thread's long history you've tried, and what you found. You can be as terse or verbose as you want. For example, here's an update I posted on July 31: https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/?do=findComment&comment=977429

BTW - this isn't just for old-timers. New followers of this thread are welcome to join in. I just want the focus to be on:

changes that have resulted in SQ improvement, as experienced by direct listening

changes in the general area of the digital transport chain upstream of the DAC, which is the primary focus of this thread - although we do stray from time to time!

Feel free to share where you think you might optimize next, or if you're done optimizing (i.e. "happy"), say so.

To jog your memories, I'm listing the categories of optimizations that we have covered over the lifetime of this thread. I very likely may have missed a category or two. But you get the idea.

Bridging Ethernet interfaces on music server

Better PSUs (upstream of DAC)

Better quality system clocks for USB, Ethernet, and system

Better quality word clocks - for example in DDCs like Mutec MC-3+, SOtM dX-USB HD Ultra, Singxer DC-1

Reference 10MHz clocks

USB regenerators

Ethernet regenerators (switches)

Standalone vs. distributed - where did you land?

Endpoint optimization - NUCs etc

Server optimization - mobos, cards, etc

OS optimization - AL, Euphony etc

Music SW optimization - Roon, HQPlayer, SL, Stylus etc

Assorted:

Digital cables - USB, Ethernet, clock, etc

Power cables

Power optimizations

Dedicated circuits

Regenerators, conditioners, isolation transformers etc.

Cable shielding - JSSG, JSSG360

Isolators, filters

Grounding - JSGT, etc

Fuses

Other accessories

One biggest contributor to improvement in SQ has been transitioning from Windows 10 to AL, nothing comes close (tried Euphony as well).

I have tried all sorts of PS upgrades (all DC except the server is on two PH SR7's), NUC's, cables, etc.

Second was surprisingly upgrading server from an i7 to an i9-9900K.

Gavin1977 · August 22, 2019

14 minutes ago, sig8 said:

One biggest contributor to improvement in SQ has been transitioning from Windows 10 to AL, nothing comes close (tried Euphony as well).

I have tried all sorts of PS upgrades (all DC except the server is on two PH SR7's), NUC's, cables, etc.

Second was surprisingly upgrading server from an i7 to an i9-9900K.

What audiable improvement did you experience when moving to i9-9900K? What associated components also changed?

Johnseye · August 22, 2019

On 8/20/2019 at 5:20 PM, austinpop said:

Hi all,

It seems this thread has gone quiet in terms of posts by some of the "regulars" who have followed it from the beginning. My assumption, or at least hope, is that people are just happy with their systems now, and hence have no reason to post.

So as a way to revitalize this thread, I have a proposal. Please consider posting a description your system in its current state. If possible, please showcase which of the optimizations we've covered in this thread's long history you've tried, and what you found. You can be as 💻terse or verbose as you want. For example, here's an update I posted on July 31: https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/?do=findComment&comment=977429

BTW - this isn't just for old-timers. New followers of this thread are welcome to join in. I just want the focus to be on:

changes that have resulted in SQ improvement, as experienced by direct listening

changes in the general area of the digital transport chain upstream of the DAC, which is the primary focus of this thread - although we do stray from time to time!

Feel free to share where you think you might optimize next, or if you're done optimizing (i.e. "happy"), say so.

To jog your memories, I'm listing the categories of optimizations that we have covered over the lifetime of this thread. I very likely may have missed a category or two. But you get the idea.

Bridging Ethernet interfaces on music server

Better PSUs (upstream of DAC)

Better quality system clocks for USB, Ethernet, and system

Better quality word clocks - for example in DDCs like Mutec MC-3+, SOtM dX-USB HD Ultra, Singxer DC-1

Reference 10MHz clocks

USB regenerators

Ethernet regenerators (switches)

Standalone vs. distributed - where did you land?

Endpoint optimization - NUCs etc

Server optimization - mobos, cards, etc

OS optimization - AL, Euphony etc

Music SW optimization - Roon, HQPlayer, SL, Stylus etc

Assorted:

Digital cables - USB, Ethernet, clock, etc

Power cables

Power optimizations

Dedicated circuits

Regenerators, conditioners, isolation transformers etc.

Cable shielding - JSSG, JSSG360

Isolators, filters

Grounding - JSGT, etc

Fuses

Other accessories

I'll show you mine if you show me yours.

Bruce Orr · August 22, 2019

Rajiv -- thanks for all you do! I'll be glad to share what I've done, as I am enjoying sounds I never would have imagined (compared to where I was) without the help of this forum's contributors and you. Know I am not a DIY (although I did manage to create add the JSSG shielding to my Lush cables!), so my process has been to read the discussions in this topic, learn and pick my spots. I got engaged Summer 2017 and the attached schematic is where I am presently.

My system has been developed almost entirely from information uncovered in this topic thread, personal messages with Rajiv and a few others, and one fun trip at the outset to meet and learn from @austinpop and Eric in Austin. I have adopted SQ recommendations from most of the areas discussed in this thread. Importantly, in many areas I have ratcheted up SQ by adopting second and third-step improvements all of you have uncovered in your never-ending quest for the best. Examples would include servers, endpoints, power supplies, cables, shielding, sequencing of devices, location of cores relative to endpoints, the move from lower to higher power voltage, and more. I wish I could say my system benefitted from all the items in Rajiv's list, but it doesn't. Having said that, I do believe I am hearing SQ at levels that make me very glad I started down this rabbit hole and invested a not insignificant amount of time and money to get here. It's my opinion every person goes through a SQ Continuum. Let's say there are 6 stages from poor SQ to the "best". At each stage of relative improvement to my system, I was blown away by the change in SQ, and couldn't really imagine how things might sound better still, and no idea how to get there. However, the inertia and contributions of the members on this forum topic kept teasing me. I read over and over again about higher levels of SQ I couldn't imagine, so decided I better stay tuned to this channel! I'm sure I've only made it thru 2 or three of the SQ improvement phases, but I can't believe how good the music sounds. Thank you to all, most of whom have unwittingly helped me get to this point.

I don't do much A/B testing for lack of interest (I have all of you for that!), and to do so is very difficult with the cramped quarters in which my system is housed. My observations in terms of SQ equipment mirror what I've heard here, but I dare not try to identify them in rank order because I have not done the critical testing. I also believe what most you say here - every system set up will have different points of inflection.

- Power supplies: Who among the uneducated would guess this?! Whether it has been the LPS-1, LPS 1.2, or now the PH SR-4, these devices make a tremendous difference. Also, moving from 7v to 12v (or higher) also provides greater torque and SQ, or so it seems.

- USB regenerator: My modded SOtM tx-ultra has provided a nice bump in SQ from the get go. But things get better with better power, as I'm learning.

- Reference Clock - I've only had one: the Mutec, and when combined with Habst cables, it definitely makes a difference. I wish more devices had ports for a reference clock.

- Cables: I was slow jumping on this one, but having made a leap to more significantly-priced power cables (Cardas Clear) in particular, I hear nice improvements in SQ. Ditto for the Lush USB interconnects and the Ghent/JSSW 360s.

- Server/Endpoint/Player - I've enjoyed the MicroRendu, UltraRendu, and Zenith SE as endpoints. All gave me great improvements, but equipment like the SE really move you to a different phase of the SQ continuum. I've had my Roon Core and Player both in the Zenith, but separated them, as I felt even if the iMac is sub-optimal vs. Epiphany and NUCs, SQ is better when there is a switch in between the Core and Endpoint. And as many have pointed out, I apparently can take that still further, and add a second switch in between for more upside.

I've gone on long enough. Thanks to all, and I'm looking forward to adding another SR-4 with Cardas Clear in early October. I hope to have a period of prolonged "listening" at that point, but I know all you will make that hard to do!

My Music Topology 08.22.19.pptx

austinpop · August 23, 2019

It's always interesting to read about how people have optimized their system. It's usually not practical to try everything, so these reports are fascinating, and it's great to hear how much people are loving the sound quality.

Keep 'em coming!

Confused · August 23, 2019

14 minutes ago, austinpop said:

It's always interesting to read about how people have optimized their system. It's usually not practical to try everything, so these reports are fascinating, and it's great to hear how much people are loving the sound quality.

Keep 'em coming!

OK - Here is my contribution🙂. I am currently running a version of what was known in as the "Trifecta". SOtM kit and spaghetti does feel a bit like ancient history in this thread now that the serial tweakers have move onto NUC's and software optimisations, but for me, it is my system. Also, it is not a case that I have settled on a solution a year or so ago and stopped there, I am still tweaking myself in pursuit of some kind of ultimate goal, kicking the ball at some ever wondering goal posts I guess.

My streaming system currently consists of a Windows 10 PC, typically running Roon, HQPlayer via router, switch, sMS-200Ultra Neo, tX-USBulta, Mutec MC3+USB AES/EBU to my Devialet amp. The SOtM kit and the MC3+USB are all fed from the Mutec REF10, the SOtM kit has the benefit of PH SR4 power. Am I happy with this set up? Well yes and no if I am honest. In terms of functionality, the sMS-200Ultra suits me perfectly. It is one of a very short list of commercially available "NAA" devices that run HQplayer, and for reasons not really relevant to this thread, I also make good use of the other Eunhasu modes, it suits me well. But in terms of sound quality, I never really bonded with the sMS-200Ultra. In terms of detail and dynamics it is tremendous, but there is something a little thin in it's presentation, which gives rise to me sometimes being slightly annoyed with the tonal presentation of some recordings. This in turn means that on some occasions I am sat listening to the system, being distracted by what some might call a slight "digital edge", so end up "listening to my system" rather than actually enjoying the music. This is not anyone's goal. Going back in time, there was a point on this thread where the action had moved away from kit like the sMS-200Ultra to the various NUC configurations that were being investigated. In my case, I had been more than happy with the results from the Mutec MC3+USB, REF10 and tX-USBultra, it was just the sMS-200Ultra that was niggling me. So I think I was probably a prime candidate to ditch the sMS-200Ultra and join in the fun with a NUC. However, I am currently extremely busy with both work and life, and frankly I do not have the time. If anything replaces the sMS-200Ultra it will need to be very much a "turnkey" off the shelf item.

But this leads me to a thought. Despite the fact that I do not have the time to go the NUC route, I have been following this thread closely over the last couple of years. Recently there was a time when the main action in this thread related to NUC's and "low latency" AudioLinux software. At a similar time, we also had a series of firmware updates from SOtM. It felt a little bit like there was an arms race between the NUC tweakers on this thread and SOtM firmware updates, as if the SOtM guys were picking up some pointers here and using this to drive their own development forward. Maybe just wild speculation, putting 2+2 together and making 5, but the fact remains that there were a lot of SOtM firmware updates at this time, and the sMS-200Ultra, to my ears at least, was sounding better than ever.

This takes me to where I am now. With the Neo upgrade and the latest firmware, the sMS-200Ultra has gone from a component that would be top of my list to replace, to something that I am now more than happy with, and overall, my system is sounding better that it ever has. Then we had the Blu-ray player incident......

I have mentioned my Pioneer BDP450 Bluray player in a number of old posts. The poor old Pioneer was becoming a bit troublesome, a case of becoming mechanically very noisy, and vibrating quite severely during some phases of disc playback. Hearing the little Pioneer audibly vibrating during key sections of a movie was not the best way to become immersed in whatever was on screen, so I figured a replacement was needed.

Looking at what was on the market, I quite fancied the Cambridge Audio CXUHD. Trouble is that I do not have a 4K TV, so I then thought about getting a second hand 1080P Oppo. With Oppo recently ceasing production of Blu-ray players, the second hand prices of the better Oppo players were a little high, to be honest, I would have preferred the new Cambridge Audio for similar money. Am I planning to get a 4K TV? In a way yes, I have two 1080P plasma TV's, one is about 6 years old, the other pushing 15 years old. I have been told that Plasma TV's only last about 10 years, so presumably one of them will keel over before two long, at which time the next TV will be 4K. But that might not happen for years, and if I did buy a 4K TV, I would want to buy a latest spec 4K player at the same time, not 3 years or something in advance. So I then started looking on eBay for reasonable looking second hand 1080P Blu-ray players, just anything that looked nice, I was not too fussy. I started putting lowish bids on a number of items, just to see what might happen. The next thing I knew, I had bought an Arcam BDP300 for £200. (they were £999 when new)

https://www.arcam.co.uk/products,FMJ,BD-...#techspecs

Anyway, the thing turned up. The box was immaculate, the remote, manual, cables etc. had all been put back in their original plastic bags with cable ties, the player itself looked new. Connecting it up, it works fine, picture quality and sound quality with movies seem fine. It is a big heavy beast too, probably over twice the size and weight of the Pioneer. I am happy I have got a reasonably good buy for the money, it definitely looks like it has been well cared for.

The Arcam BDP300 never had the best reputation, certainly a step down from the equivalent Oppo players. It does use a lot of "parts bin" components from mainstream players, with a little of Arcam's audio circuitry and LPSU's on top. Anyway, this got me wondering, what would it sound like in two channel mode? The Arcam will play music files from a USB drive, and will of course play CD's. I read something on-line that claimed that the BSP300 sounds slightly better from USB, maybe a quirk of the "parts bin" CD / Blu ray drive. So out of curiosity, I loaded some music files onto a USB thumb drive, connected the Arcam to the REF 10 clocked Mutec MC3+USB, thus allowing me to do a direct head to head comparison with the SOtM streaming kit. I have done this before with the old Pioneer BDP450, and the sound quality was quite a step behind the SOtM kit. I have also tried this with my near 20 year old Musical Fidelity CD player, which as I mentioned in a previous post, got pretty close to the SOtM kit.

I was expecting similar things from the Arcam, that it would sound ok, but the SOtM kit would edge ahead for ultimate sound quality. Listening to the first few tracks, the Arcam sounded good, a few switches between the Arcam and SOtM kit I found that it was very hard to discern much difference, this was a close one. I think expectation bias was getting too me, I was telling myself that the SOtM kit was better, after all, it is dedicated audio kit, it has the advantage of being fed with up-sampling HQPlayer, it has SR4 power supplies, etc, etc. Ultimately, and much to my surprise, I had to come to the conclusion that the Arcam was the better performer. Just that touch more real, less "digital", slightly better ambiance and clarity. Not what I expected.

So a big surprise. I guess I have never been 100% convinced by the SOtM kit, but I did not expect my "£200 off eBay" universal device to edge it out. Maybe something to do with taking the PC and network out of the chain, and instead using a simple bit of audio circuitry to send a digital file to the MC3+USB? Maybe a result of taking USB out of the chain altogether? Doing a bit of research on the Arcam, it does appear to be very well designed from a audio perspective, very good electrical isolation and shielding , this kind of thing. But I remain staggered that it has edged out the REF10 clocked SOtM kit. In a way this reminds me of the guy that was experimenting with very simple sources playing from SD cards. A similar concept, and he seamed to be getting good results too. I guess this shows that with audio it is not always the complex and expensive solutions that come out on top, in some areas simplicity can completely nail it.

The plan now is to use the Arcam as a "reference", a second source that is almost identical the SOtM kit for sound quality. This could be very useful for verifying a few tweaks. For example, a while ago when changing my rack and layout of the kit, I took the SOtM modified ethernet switch out of the chain, and it has never gone back. To compare the system with or without the SOtM kit would be a case of listening to a playlist, then very much faffing around swapping out the switch, getting Roon and HQPlayer to talk to Eunhasu again, and so on. Usually when doing this, Eunhasu crashes, much hassle getting everything to work again, by which time I am in a bad mood and have forgotten exactly how things sounded with the previous switch. This is not a good way to look for those subtle differences. But now I have the Arcam as a kind of benchmark, I could install a new switch or whatever, and then do immediate A/B swaps between the two methods of playback. Immediate and hassle free playing the same song, much better, and much easier.

I mentioned functionality at the start of this post, and in this regard the Arcam is utter rubbish! Not in terms of it's ability to spin Blu-ray discs, this is fine, but to actually live with as a music server, it is hopeless. A case of slow and clunky on TV screen navigation. So I am not using it for general listening, but as a reference and a tool for A/B tests, the thing is gold dust.

On my "to do" list is to try this with the SOtM switch. I also have a plan to try the SOtM kit direct to the Devialet's USB, taking the Mutec MC3+USB out of the chain. In this scenario I could try direct A/B swaps between Arcam +Mutec (which sounds virtually the same as SOtM to Mutec to Devialet) with the SOtM kit direct by USB. Depending on the results of this, I could try running HQPlayer, DSD64, SOtM USB, direct to the Devialet, with direct swaps between the "benchmark" Arcam + Mutec route. I am sure you get the idea. The SOtM kit is now powered by a pair of Paul Hynes SR4's, maybe this will provide an advantage over the SMPS powered MC3+USB? Now I can do direct A/B comparisons, these checks will be much easier to do as it eliminates relying on aural memory. It surprises me to think that I have had the SOtM kit for over a year and never tried it direct to the Devialet's USB input, but now I can do these direct A/B comparisons I have become a lot keener to try it. Other things to try now I can do this direct A/B testing is to run my old microRendu again, mR to tX-USBultra to Mutec MC3+USB. Another would be to try the sPS-500 to power my router. I suspect this will make zero to very little difference, but now I have a direct A/B reference it might be worth a go. Currently the sPS-500 is sat on a shelf doing nothing, so worth a try I guess.

In terms of my medium term plans, the EtherRegen looks interesting, and I am also currently researching Room treatments, but that is a topic for another thread.

Meanwhile I will keep following the many adventures in this thread with great interest, ever optimistic that it will eventually trickle down into my own system, by some means or another. Keep up the good work folks! And a final nod to @austinpopfor his good work keeping this thread in good shape, it is appreciated.

Blackmorec · August 23, 2019

Quite a few of you now have implemented the ‘guidance’ found in this thread and managed to create some really high performing systems that quite possibly exceed your previous expectations. In the meantime I’ve continued to refine my system, which is entirely off the shelf bar a couple of key mods to the network to try and achieve similar levels with I believe some degree of success. I’m just wondering if you’ve noticed that as things got better, you started to get improvements that were clear to hear but quite difficult to analyse and describe, other than a rather lame “the music sounds more real” or “everything sounds better”. Beyond that there isn’t particularly more of anything.....it just sounds better.....more natural, more real, more alive, more vibrant, more there but almost indescribable in the usual hi-fi parlance. I can’t actually pick out what got better, just that it did and sometimes by a not trivial amount. Anyone else feeling that their recent upgrades brought great, but difficult to quantify and describe sonic improvements?

Ralf11 · August 23, 2019

On 8/20/2019 at 3:20 PM, austinpop said:

* * *categories of optimizations that we have covered over the lifetime of this thread. ...

Bridging Ethernet interfaces on music server

Better PSUs (upstream of DAC)

Better quality system clocks for USB, Ethernet, and system

Better quality word clocks - for example in DDCs like Mutec MC-3+, SOtM dX-USB HD Ultra, Singxer DC-1

Reference 10MHz clocks

USB regenerators

Ethernet regenerators (switches)

Standalone vs. distributed - where did you land?

Endpoint optimization - NUCs etc

Server optimization - mobos, cards, etc

OS optimization - AL, Euphony etc

Music SW optimization - Roon, HQPlayer, SL, Stylus etc

Assorted:

Digital cables - USB, Ethernet, clock, etc

Power cables

Power optimizations

Dedicated circuits

Regenerators, conditioners, isolation transformers etc.

Cable shielding - JSSG, JSSG360

Isolators, filters

Grounding - JSGT, etc

Fuses

Other accessories

you might ask CC to put this post in a more place as a comprehensive list of possibilities

it's too long to go at the top but a pinpoint link to would fit in a single line

sig8 · August 24, 2019

On 8/22/2019 at 4:24 PM, Gavin1977 said:

What audiable improvement did you experience when moving to i9-9900K? What associated components also changed?

Contrary to conventional wisdom; sound got little more relaxed. Nothing else changed, even used the same RAM modules, only the board and processor were different.

Johnseye · August 24, 2019

1 hour ago, sig8 said:

Contrary to conventional wisdom; sound got little more relaxed. Nothing else changed, even used the same RAM modules, only the board and processor were different.

Could have been something with the board physically or chipset related as well.

Johnseye · August 24, 2019

On 8/23/2019 at 12:20 PM, Blackmorec said:

Quite a few of you now have implemented the ‘guidance’ found in this thread and managed to create some really high performing systems that quite possibly exceed your previous expectations. In the meantime I’ve continued to refine my system, which is entirely off the shelf bar a couple of key mods to the network to try and achieve similar levels with I believe some degree of success. I’m just wondering if you’ve noticed that as things got better, you started to get improvements that were clear to hear but quite difficult to analyse and describe, other than a rather lame “the music sounds more real” or “everything sounds better”. Beyond that there isn’t particularly more of anything.....it just sounds better.....more natural, more real, more alive, more vibrant, more there but almost indescribable in the usual hi-fi parlance. I can’t actually pick out what got better, just that it did and sometimes by a not trivial amount. Anyone else feeling that their recent upgrades brought great, but difficult to quantify and describe sonic improvements?

The removal of noise from components will allow the music to be heard without that influence layered into it. Noise will come from many places such as power, clocking, EMI/RFI, etc. As you remove those sources of noise from your system you may hear more of an influence from a different noise source that was combined with or mask by another noise source. Descriptions of what you hear are based on this and interpreted by layman's terms as best as possible.

fas42 · August 25, 2019

On 8/24/2019 at 3:20 AM, Blackmorec said:

Quite a few of you now have implemented the ‘guidance’ found in this thread and managed to create some really high performing systems that quite possibly exceed your previous expectations. In the meantime I’ve continued to refine my system, which is entirely off the shelf bar a couple of key mods to the network to try and achieve similar levels with I believe some degree of success. I’m just wondering if you’ve noticed that as things got better, you started to get improvements that were clear to hear but quite difficult to analyse and describe, other than a rather lame “the music sounds more real” or “everything sounds better”. Beyond that there isn’t particularly more of anything.....it just sounds better.....more natural, more real, more alive, more vibrant, more there but almost indescribable in the usual hi-fi parlance. I can’t actually pick out what got better, just that it did and sometimes by a not trivial amount. Anyone else feeling that their recent upgrades brought great, but difficult to quantify and describe sonic improvements?

Yes, this is the onset of what I call convincing, or competent sound - for me, the source has always been CDs, but that is completely irrelevant to the situation - recordings are stored as digital data, somewhere, and then read by the appropriate mechanism to be delivered to the DAC; it's from here on in that things get interesting ... .

What the factor is in causing this perceived transformation is that all the key audible distortions are now under control; they no longer disturb the brain, and force one to constantly compensate for their presence - amongst other negatives these being too audible is the cause of, notorious, "listener fatigue" ...

One very useful approach is not to think that the sound "is better" - rather, that any standard that is less than that is perceptible flawed, and with some practice it becomes trivially easy to pinpoint what the poorer SQ is getting wrong - the issues "stand out like sore thumbs".

FredM · August 25, 2019

On 7/29/2019 at 2:20 AM, seeteeyou said:

The rationale behind choosing that ASRock gaming motherboard (instead of Supermicro) could be found here

https://audiophilestyle.com/forums/topic/55873-shootout-at-the-linux-corral-audiolinux-vs-euphony/page/9/?tab=comments#comment-961583

For X570 motherboard and Ryzen 3000 Series, this particular motherboard is gonna be fanless while beating ASRock (14 + 2 phase with 70A instead of 5 + 2 phase with 60A) handily

https://www.reddit.com/r/Amd/comments/bxl3rm/gigabyte_x570_xtreme_motherboard_analysis_true/

https://pokde.net/system/pc/motherboard/gigabyte-unveils-true-16-phase-amd-x570-motherboard-pcie-4-0-ssds/

However, its form factor is E-ATX and therefore only one fanless chassis is available so far

http://www.turemetal.com/product_en.html

https://item.taobao.com/item.htm?id=590791895684

https://www.fanlesspro.com/shop/product/fanless-cases/turemetal-up10-fanless-case/

The CPU alone could use 2 separate 12V rails (i.e. two 8-pin EPS connectors) to begin with, getting quite expensive already and then we'll also need one more rail for 24-pin ATX connector. Quality rails cost quite a bit while it also takes at least several months to build each LPSU.

There's also C621 AORUS XTREME with 32 phase VRM and we could even feed 4 separate rails to the CPU itself, followed by 2 more rails for a pair of 24-pin ATX connectors. Could such a beast beat that €20,000 Taiko Audio SGM Extreme?

The XTREME has impressive temp scores

Some more info here, sub titles/translate is your friend 🙂

https://m.youtube.com/watch?v=hZrCHUwzd1Y

https://nl.hardware.info/artikel/9411/15-amd-x570-moederborden-review-op-de-proef-met-een-3900x

Anyone already had the change to try the Gigabyte X570 Extreme?

Gavin1977 · August 25, 2019

1 hour ago, FredM said:

The XTREME has impressive temp scores

Some more info here, sub titles/translate is your friend 🙂

https://m.youtube.com/watch?v=hZrCHUwzd1Y

https://nl.hardware.info/artikel/9411/15-amd-x570-moederborden-review-op-de-proef-met-een-3900x

Anyone already had the change to try the Gigabyte X570 Extreme?

I’ve not tried x570, but will be trying 32 phase MSI XPower motherboard shortly based on the same logic. Will be interested in your results as well as I’ll be moving from a small 6 phase MSI Z97 Gaming ITX. The XPower has tantalum chokes which are apparently better as blocking high frequency noise then the super ferrite chokes.

FredM · August 25, 2019

2 hours ago, Gavin1977 said:

Will be interested in your results as well

I’m affraid I won’t have any results to share any time soon (*). Server wise I’m considering three options at the moment, all with their pros and cons:

- DIY server (X570 build?)

- Pinkfaun server

- new SOtM server

(* unfortunately I’m back at square one: https://audiophilestyle.com/forums/topic/43611-tls-the-one-server/page/4/?tab=comments#comment-970079)

FredM · August 25, 2019

2 hours ago, Gavin1977 said:

I’ll be moving from a small 6 phase MSI Z97 Gaming ITX.

A whopping 24 phase difference, looking forward to your impressions.

rickca · August 25, 2019

3 hours ago, Gavin1977 said:

https://nl.hardware.info/artikel/9411/15-amd-x570-moederborden-review-op-de-proef-met-een-3900x

What an excellent article! I will check this website in future. Thanks for the link.

Gavin1977 · August 28, 2019

Hello All! OK I have completed my VRM comparison. The results suprised me somewhat, but perhaps my findings shouldn't have been such a suprise after alll.... read on.

So, I was waiting for a Gigabyte Z87 XPower motherboard with 32 Phase VRM to be delivered. Much to my disappointment the eBay vendor sent me the wrong motherboard. So anyway, my testing was now done on a Z87 Mpower motherboard with 16 Phase VRM, vs my older 6 Phase MSI ITX motherboard. Both motherboards use the same capacitors and ferrite chokes.

Previously I had ranked some of my earlier system configurations here. My system following the server is Naim DAC V1, Ncore NC500 Monoblocks and IPL Transmission line speakers which remained unchanged during the testing. Playback is direct from the server via USB to the Naim V1 (i.e. not a server/endpoint config). I do have a separate Synology NAS with a load of music on it elsewhere, but to my ears it sounds noticeable poorer than playback via Tidal using Euphony Stylus. I use WifFi, not Cat5.

So, subjectively, how would I rank now (ranked according to software used, as I find this to be one of the largest determiners):

BEST

Euphony (DNHE NUC with Paul Hynes SR7)

.

Euphony (MSI Gaming Z97 ITX 6 Phase VRM, 4790k, Corsair SF450 20-25mV ripple)

.

Euphony (MSI Z87 MPower ATX 16 Phase VRM, 4790k, Super Flower Leadex III 550W Gold 5mV ripple)

.

Euphony (DNHE NUC with stock power brick)

Audiolinux RAMBOOT (DNHE NUC with stock power brick)

.

Roon Rock or Audiolinux without RAMboot (DNHE NUC with stock power brick)

.

MacBook Pro, iTunes playback within MacOS

WORST

This pretty much surmises my journey so far. The audiable differences between each configuration are easy to discern, so this is not a subtle thing.

I have stopped using AudioLinux, and favour Euphony, so I haven't and won't be doing any further testing with AudioLinux unless there is some new amazing development. That's why you see it lower down the list.

So, what I have found with the VRM experiment is that a very high quality VRM - 16 Phases in this case - when combined with a low ripple power supply sounds really very detailed, I would say it is close to the DNHE / SR7 combination in terms of detail retrieval. The 16 Phase VRM on the MSI Z87 MPower is apparently able to control voltages with an accuracy of 0.001v. So high phase VRM when paired with a low ripple power supply = more detail than a 6 Phase VRM and a higher ripple power supply. This would make sense and is completely logical!

However, I have lost tonal colour and dynamics/transient response compared to the Z97 ITX 6 Phase with the Corsair PSU (this is why I rank the Corsair solution higher). I have not tested, but I am pretty sure this is due to the differences in PSU design. Perhaps to achieve the low 5mV ripple the SuperFlower PSU has more voltage regulation applied in its design and this has damped it's performance as a power supply for audio purposes when compared with the Corsair.

So what have I learnt. It's all about system design. Paul Hynes SR7 not only sounds good because of it's low ripple, it's also down to Paul knowing how different circuits and components sound (e.g. capacitor brand X vs capacitor brand Y). This is were it becomes a matter of experience and time, and 'perhaps' were commercial companies succeed compared to the DIY route due to the knowledge and resources available. I'm now of the presumption that it not will not be easy to match the performance of a Pink Fawn or Sound Galleries Music Server as you need a specially designed PSU (this was not my aim anyhow!). The Superflower SMPS has the lowest ripple in it's class, so I think we're pretty much maxed out before adopting something like a HDPLEX PSU. Since the combination of a lower power DNHE NUC and a SR7 achieves lashings of detail, dynamics and tonal colour I don't think it would be worth my while investigating different combinations of ATX motherboards and power supplies. I do not have the experience (e.g. capacitor brand X vs capacitor brand Y), time or resources to evaluate lots of different combinations to get the right balance of characteristics.

I was lucky to get hold of an SR7 second hand, prior to this I was going to retire the NUC and use the MSI Gaming Z97 with the Corsair PSU. It really does sound great, and provided you don't mind the reduction in detail you find it very cost effective and musically satisfying. I'm going to sell the PC's now, I might try a SOtM tx-USBultra with the proceeds as my next step as I'm lucky to have a spare rail on my SR7. This would be logical as I completely trust the experience of AustinPop and he has done something similar, although I intend to limit my digital frontend to no more than two boxes :-)

I hope this helps others.

George Hincapie · August 31, 2019

I am still waiting for my Farad server supply. The power connector needed for my Audio Store Prestige MX server couldn't be sourced in Europe so Mattijs has had to order it from the US. The delay is doing my nut, but Farad have been great. Looking forward to getting it.

seeteeyou · September 1, 2019

FYI - here's yet another flavor of Ubuntu with RAM disk (activation required) support

https://www.snakeoil-os.net/wiki/Configuration/snakeoil#RAM_Disk

https://www.snakeoil-os.net/forums/Thread-Snakeoil-Firmware-1-1-7-Blind-Testing-Update-7

Activation codes will be sent after (one-time / monthly for whatever amounts) donations are received

https://www.snakeoil-os.net/wiki/Configuration/snakeoil-activation

https://www.patreon.com/snakeoilos

austinpop · September 2, 2019

Followup on Farad Super 3 PSU

A few weeks ago, I wrote about my impressions of the Farad Super 3 PSU here. I listed my next steps in the summary.

On 7/24/2019 at 10:26 PM, austinpop said:

Next Steps

Mattijs will be shipping me a 15V unit soon to validate whether the Farad @ 15V is a better fit for the TT2. I plan to compare this with an SR7 @ 15V. The TT2 improves only modestly even with the SR-7, so the question here is whether the TT2 powered by the Farad is "close enough" to the SQ of the TT2 powered by the SR-7 DRXL.

I also expect to receive, in due course, Mattijs level 3 cable once he's perfected it. I'll be comparing this to my benchmark Audio Sensibility and Oyaide DC cables, and reporting back.

Stay tuned for an update in a few weeks.

As promised, here is my followup. This time, I'll describe the 15V unit of the Farad Super 3, and the soon-to-be-unveiled Level 3 silver cable. After having put the 15V unit and the L3 cable through the same burn in procedure as the previous 12V unit, I was ready for some comparisons. I used the same power cable (Cardas Clear) and DC cable (Farad Level 3) on both the 12V and 15V versions of the Farad Super 3.

Comparison of the Farads powering the Chord Hugo TT 2

I was actually a bit taken aback at how big an improvement I got going to 15V! Well, I knew the TT2's SQ improved when powered at 15V vs. 12V, but with the Farads I could barely believe I was listening to the same power supply, just at different voltages. The improvement centered around 2 areas:

Bass: I was listening to Quia Fecit from Arnesen's Magnificat. The movement starts with a deep bass drone. This drone was startlingly more physical and visceral with the 15V supply.
Image size: This track - indeed, the whole album - does an excellent job projecting the spacious ambience of the Nidaros Cathedral, Norway where it was recorded. While the 12V supply does an excellent job, the image just grows in all 3 dimensions with the 15V supply.

Now, keep in mind, this comparison was primarily applicable to the TT2. Not all components accept a range of voltages, and even if they do, they don't necessarily exhibit big SQ differences with higher voltages.

Comparing the Farad 15V with other PSUs powering the TT2

I compared the Farad 15V to the following supplies:

Stock TT2 15V SMPS: Just like my initial comparison with the 12V unit, this was a significant step up in SQ. The Farad just sounds cleaner, more relaxed, and richer than the stock PSU.
Power Add Pilot Pro battery: this battery only has 12V and 16V settings, and the TT2 will not accept 16V. While the battery @ 12V is a nice step up from the stock SMPS, the Farad 15V takes SQ up by another quantum leap. Again, the main areas of improvement are bass and image size, along with more incisive transients.
SR4-15: this comparison is pending until Paul has tested the SR4-15's ability to handle a 3A inrush transient.

Comparing the Farad 15V with an SR-7 SR (single regulator) 15V rail powering the Chord Hugo TT 2

The SR-7 (using the Paul Hynes DC3FSXLR cable) still held an SQ advantage over the Farad Super 3, but the difference was quite modest. Since SR-7 rails are typically a scarce resource for those who have them, I would be quite happy to use a 15V Farad on the TT2, and allocate the SR7 rails to devices for which it matters more.

Impressions of the prototype Farad Level 3 silver DC cable

This is a clear step up from the Level 2 cable. In my listening, I found it very, very similar to the Oyaide FTVS-910 cable I mentioned in my prior report. Compared to the Audio Sensibility Signature, the difference was subtle. The L3 cable was more visceral in the bass, while the AS cable had a hint more air and transparency. I'd be hard pressed to prefer one over the other, and it would come down to a question of system synergy.

Final Conclusions

The Farad Super 3, in addition to being an excellent PSU as I found in my previous report, has the added advantage of a continuous current capacity of 3A. This puts it in a unique position compared to lower powered supplies like the LPS-1.2 and the SR-4. The field of high-quality PSUs with current capacity over 2A is quite limited, and the Farad is a welcome addition. I can see it being a great fit for DACs, and other current-hungry devices like NUCs.

For my use case of powering a Chord Hugo TT2, the Farad Super 3 was not only up to the task, its SQ while doing so was excellent.

I next look forward to Mattijs' higher current Farad supply, now in active development.

Tommd64 · September 2, 2019

Dear Austinpop, thank you for your long review and comparison between other PSUs in my opinion and I also tested a lot, only the SotM sPS500 came very close. My test case was a Mytek Brooklyn + It made a world of difference between some PSUs. Although I have never experienced the Paul Hynes. The Farad3 is sofar the best i, ve ever tested.

RickyV · September 2, 2019

2 minutes ago, Tommd64 said:

Beste Austinpop, bedankt voor je lange beoordeling en vergelijking tussen andere PSU's in mijn opinie en ik heb ook veel getest, alleen de SotM sPS500 kwam heel dichtbij. Mijn testcase was een Mytek Brooklyn + Het maakte een wereld van verschil tussen sommige PSU's. Hoewel ik de Paul Hynes nog nooit heb meegemaakt. The Farad3 is sofar the best i,ve ever tested.

Thats Dutch try English works better😁. Maybe you can edit the post

Tommd64 · September 2, 2019

Sorry i wrote it in original English i didn,t know googler translate did the job ill look for it..

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

Recommended Posts

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Create an account or sign in to comment

Create an account

Sign in