[doraymon]

How does one go about purchasing a holo audio cyan...

You can message Tim Connor of Holo Audio USA aka @bimmer100
He'll take care of you.

 

[ahmadfaizadnan]

You can message Tim Connor of Holo Audio USA aka @bimmer100
He'll take care of you.

It's nowhere to be found on their website but yeah maybe Tim can help with that.

 

[Energy]

Not really. If I were you, I would pick microrendu/ ultra-rendu without a second thought. It's not about Roon, it's about HQplayer NAA (network audio adapter). Have a look at the HQplayer site or join the CA community, lots of sensible nerds discussing such there.

In short, by having NAA, you will literally galvanically isolate your HQplayer DSP PC from your DAC since ethernet transmission is inherently galvanic isolated. So you will be less worried about the RFI/EMI your PC generates during intensive HQplayer processing transmitting to your DAC/interface via usb cable. You can literally place your PC and microrendu at two difference places. Disadvantage of it is: your cannot stream your youtube or movie through your system, but these things have crappy SQ, so I never bother streaming them from PC to my HIFI system.

NAA as a second setter, only do the bit-perfect transmission without doing heavy processing. So it can be a compact, small, low-power consumption, software optimized NUC (I regard microrendu as a linux-based NUC as well), you can have better signal integrity USB output and less noise problem. It's also easy for you to upgrade lots of accessories such as LPSU for your NUC (since they are low power consumption, find a good LPS-1 is much cheaper and easier than big linear power supply in your PC).

You can go for ultra-capacitor-based LPS like LPS-1 from Uptone or mini-dc-4ever from VR and your microrendu powered by these things are basically off the gird. You can also put a optical ethernet switch in between, and install a NIC pice card on your PC, so you transmit your HQplayer DSD512 data through your PC via optical Ethernet cable which is a step further of isolation compared to normal copper Ethernet cable. Thinking about these up-gradable and flexible options.

Only one thing you need to be careful is some DAC supports DSD512 (44.1k or 48k) only on windows OS like my T+A DAC and your DAC supports DSD512 at 48K base only on linux. I prefer a linux-based NUC since it save me the trouble to optimize OS and product like microrendu has optimised its USB output. But unfortunately, i cannot use them since my DAC's firmware does not support it.

You can also upsample DSD64 file to DSD512 one without PCM conversion, this is also one of the unique selling points of HQplayer, which also yield wonderful result. Just untick the DirectSDM in the DSF/DSDIFF settings.

I want to start off by really thanking you for the time and efforts that you put into your replies in order to educate the lesser few out here.

Starting off, I'm aware of the NAA mode mRendu and uRendu both offers when using HQPlayer and how the Ethernet port offers galvanic isolation. Unfortunately where I currently have the computer is for multifarious uses such as some of what you've mentioned (YouTube, Movies, Photoshop, Internet Browsing, Music Listening, Bitcoin Mining). On the bright side, USB transmission won't be an issue due to the ISO REGEN that's under my disposal. It offers true galvanic isolation and cleans up the USB signal. I also have a Singxer SU-1 converting that USB signal to the superior balanced I²S before making it's way into the DAC. All ultra low jitter and noise methods.

I see why you would recommend the two items from Sonore as it's an easy solution in solving the USB output noise/signal issue as well as easily being powered by something like the well regulated LPS-1 with the addition of being free from AC Leakage Currents. Your method for the optical ethernet switch also makes total sense and I wish I could do something similar if only I didn't have so many other uses for this setup.

Wait a minute. Your T+A supports 48K on Windows OS? Drats! I wish mine did...

Currently to get the Singxer SU-1 working in DSD512 one must use the Holo Audio Spring v3.40 driver. Soon there's news that Singxer is coming out with a new driver for their device to support DSD512 natively so since it sits between the PC and the DAC, perhaps by some miracle it will offer the 48KHz rate.. If it's doesn't then I'm really going to look into installing a virtual machine for Linux on my Windows OS to get around this issue. Even then, I find it difficult to be able to run the required core count for POLY-SINC-XTR on a virtual machine. I have no knowledge of how to use Linux so it would be extremely challenging. I can't just wipe out Windows OS for Linux since the majority of productivity and all it's needed programs are Windows based (video editing programs, etc). Considering all the gear I already have, grabbing a Sonore device to be put in the same room will be a lot just for a small improvement. Unless I sell the ISO REGEN in exchange for the uRendu? I hear the uRendu has better USB output than the mRendu. Or should I get the cheaper mRendu and keep the iSO Regen. Or do you think the mRendu alone is good enough? I'm sorry to hear that your T+A DAC doesn't support them. Why didn't you switch over to the Spring instead?

Yes I love the DSD to DSD upsampling without PCM conversion. Jussi surely has my loyalty as a customer, great stuff.

P.S.
@elan120 & @Yeang
I found HQPlayer to perform better with SMT turned off (Multi-thread or Hyperthreading).
Each core utilizes 2MB of L3 cache so when multithread is on there's a lot of back and forth that happens where the L3 cache is split up within the algorithm and for me, it doesn't play well. Call it placebo, but somehow it sounds better. I've tested a bunch of my well known songs and each listen has lead me to the same conclusion.

 

[louisxiawei]

. On the bright side, USB transmission won't be an issue due to the ISO REGEN that's under my disposal. It offers true galvanic isolation and cleans up the USB signal. I also have a Singxer SU-1 converting that USB signal to the superior balanced I²S before making it's way into the DAC. All ultra low jitter and noise methods.

I see why you would recommend the two items from Sonore as it's an easy solution in solving the USB output noise/signal issue as well as easily being powered by something like the well regulated LPS-1 with the addition of being free from AC Leakage Currents.

IIRC, John Swenson mentioned that although ISO REGEN has a good galvanic isolation using high-speed isolation chip, it cannot isolate the noise completely. That's why the microrendu/ultra-rendu has a different approach using more solid way, since ethernet transmission is galvanically isolated by standard. Otherwise, these companies, hospitals, schools will have trouble with many ethenret cables in their buildings.

Or should I get the cheaper mRendu and keep the iSO Regen. Or do you think the mRendu alone is good enough?

I think one of the merit of microrendu is its compact size, so you can put it right back of your dac and being stealth while ultrarendu is bigger and claimed to have better main board design, clock and stuff like that. But microrendu also has its uprgrade option, and the total price will be the same as ultrarendu, who is better? I have no idea.

I would pick ultrarendu and keep the iso regen as well. So the logic is like this: Ultrarendu does the galvanic isolation from your HQplayer PC the first time via ethernet, and then ISO-regen does the galvanic isolation from the ultrarendu the second time via its high-speed chip. (since ultrarendu is also a NUC, it has its own noise interference, but much smaller than your HQplayer PC again)

Regarding the performance, in my case, i put a windows based NAA, a NUC (Logic CL100) between my audio system and my HQplayer PC, and its USB has not been optimized like microrendu, I can already feel the obvious improvement. The sound is much cleaner.

I can't just wipe out Windows OS for Linux since the majority of productivity and all it's needed programs are Windows based (video editing programs, etc).

I won't encourage you to use LINUX system for your HQplayer PC as well if you have so much other fun on windows OS. In my case, I will also stick to the windows OS. My i9 7980XE and Titan xp is on the way to me, for that combo, installing linux just for HQplayer is a waste I think. VR gaming, data processing, video editing, etc these things will equally make you life productive and happy.

However, linux based HQplayer is quite different from windows desktop one, it's a embedded one, having some more advanced function. I might consider build another Linux server for HQplayer PC to give Linux HQplayer a go in the future. But I doubt whether I can leave Roon's UI and Tidal (I'm using Roon+HQplayer +Tidal+ lots of DSD files in my local drivers), and do I need that many PCs at home?

Why didn't you switch over to the Spring instead?

I like my T+A DSD512's sound and T+A is working on linux firmware at the moment, so I'm going to wait instead of purchasing other DACs.

 

[zolkis]

All isolation/purification/decrapification are partial, be it capacitive, inductive, or opto-coupled. They just spread noise in different ways and you just choose the taste combination of the poison. No doubt you hear a difference, but a new product comes and you hear different difference. IMHO albeit convenient, high frequency serial comms are not really good for audio, and that includes I2C as well... I'd vouch for low EMI parallel interfaces, but where are those, and then on the comms side it would be subject to more interference than we want. If I'd have to choose a serial poison, I'd vouch for Ethernet/RTP audio comms, too, but with low clock parallel HW interfaces (increase bandwidth with memory size, not clock speed). Also, I prefer the sound of parallel DACs (a matter of constant discussion with Jussi ). But that's just a preferred poison, too . Anyway, to bring some balance to the force, if the Holo Cyan comes out, I prefer the PCM version .

 

[Energy]

IIRC, John Swenson mentioned that although ISO REGEN has a good galvanic isolation using high-speed isolation chip, it cannot isolate the noise completely. That's why the microrendu/ultra-rendu has a different approach using more solid way, since ethernet transmission is galvanically isolated by standard. Otherwise, these companies, hospitals, schools will have trouble with many Ethernet cables in their buildings.

I think one of the merit of microrendu is its compact size, so you can put it right back of your dac and being stealth while ultrarendu is bigger and claimed to have better main board design, clock and stuff like that. But microrendu also has its uprgrade option, and the total price will be the same as ultrarendu, who is better? I have no idea.

I would pick ultrarendu and keep the iso regen as well. So the logic is like this: Ultrarendu does the galvanic isolation from your HQplayer PC the first time via ethernet, and then ISO-regen does the galvanic isolation from the ultrarendu the second time via its high-speed chip. (since ultrarendu is also a NUC, it has its own noise interference, but much smaller than your HQplayer PC again)

Regarding the performance, in my case, i put a windows based NAA, a NUC (Logic CL100) between my audio system and my HQplayer PC, and its USB has not been optimized like microrendu, I can already feel the obvious improvement. The sound is much cleaner.

I won't encourage you to use LINUX system for your HQplayer PC as well if you have so much other fun on windows OS. In my case, I will also stick to the windows OS. My i9 7980XE and Titan xp is on the way to me, for that combo, installing linux just for HQplayer is a waste I think. VR gaming, data processing, video editing, etc these things will equally make you life productive and happy.

However, linux based HQplayer is quite different from windows desktop one, it's a embedded one, having some more advanced function. I might consider build another Linux server for HQplayer PC to give Linux HQplayer a go in the future. But I doubt whether I can leave Roon's UI and Tidal (I'm using Roon+HQplayer +Tidal+ lots of DSD files in my local drivers), and do I need that many PCs at home?

I like my T+A DSD512's sound and T+A is working on linux firmware at the moment, so I'm going to wait instead of purchasing other DACs.

You make some very solid points. For that I plan to make my next purchase the Sonore ultraRendu (and LPS-1). For awhile I imagined the ISO REGEN to be doing more than what it actually did. I thought galvanic isolation was complete isolation so I did no bother with any other forms of NUC's and focused heavily on the 16-core computer for HQPlayer's xtreme filter.

So if Ethernet transmission offers complete isolation, does this mean I simply put the Sonore ultraRendu right after the computer and connect the two together using Ethernet CAT6 cable? You said there would be more isolation if I got an optical ethernet switch box as well as an optical Ethernet output PCI-E card for the computer? Does these two transfer data using light thus why there is better isolation? But I thought you said Ethernet was already isolation why need for light? Wouldn't light result in timing delays like with other optical connections as seen within DAC's or does it get reclocked again once the data reaches the NUC?

I think I will start small and grab the Sonore ultraRendu + LPS-1 first. Since you recommend keeping the ISO REGEN, is it okay if I use the iPower to power it? My other LPS-1 is already being used on the Singxer SU-1. I think it would be crazy expensive to have three LPS-1's for such minimal gains considering the SU-1 already consist of digital isolators within it. The Singxer SU-1 im using already draws LPS-1 power and so the iPower only powers the ISO REGEN as well as the USB input side of the Singxer SU-1.

Furthermore, with this addition I should be able to play DSD512 in 48KHz right? I'm pretty sure my 16 core computer can do it since it's running at 4.1GHz overclocked.

Do you recommend any good performing Ethernet cables but on the cheap side? I've been going through financial problems lately and can't afford to spend hundreds on an Ethernet cable. The optical method you suggested I may try in the future once I better understand it and have learned it's advantages alongside its translation on obtaining higher sound quality.

 

[oneguy]

You make some very solid points. For that I plan to make my next purchase the Sonore ultraRendu (and LPS-1). For awhile I imagined the ISO REGEN to be doing more than what it actually did. I thought galvanic isolation was complete isolation so I did no bother with any other forms of NUC's and focused heavily on the 16-core computer for HQPlayer's xtreme filter.

So if Ethernet transmission offers complete isolation, does this mean I simply put the Sonore ultraRendu right after the computer and connect the two together using Ethernet CAT6 cable? You said there would be more isolation if I got an optical ethernet switch box as well as an optical Ethernet output PCI-E card for the computer? Does these two transfer data using light thus why there is better isolation? But I thought you said Ethernet was already isolation why need for light? Wouldn't light result in timing delays like with other optical connections as seen within DAC's or does it get reclocked again once the data reaches the NUC?

I think I will start small and grab the Sonore ultraRendu + LPS-1 first. Since you recommend keeping the ISO REGEN, is it okay if I use the iPower to power it? My other LPS-1 is already being used on the Singxer SU-1. I think it would be crazy expensive to have three LPS-1's for such minimal gains considering the SU-1 already consist of digital isolators within it. The Singxer SU-1 im using already draws LPS-1 power and so the iPower only powers the ISO REGEN as well as the USB input side of the Singxer SU-1.

Furthermore, with this addition I should be able to play DSD512 in 48KHz right? I'm pretty sure my 16 core computer can do it since it's running at 4.1GHz overclocked.

Do you recommend any good performing Ethernet cables but on the cheap side? I've been going through financial problems lately and can't afford to spend hundreds on an Ethernet cable. The optical method you suggested I may try in the future once I better understand it and have learned it's advantages alongside its translation on obtaining higher sound quality.

Ethernet offers isolation from computer noise but not network noise. There is s thread on computer audiophile which I’ll link at the bottom taking about this and describing how to prevent this. I’ll summarize it by saying the simplest way is to use a NetGear FS105, FS108, GS105 or GS108 switch (12v version 3 for the FS105, not sure about the others) and ground the negative negative output of the SMPS.

https://www.computeraudiophile.com/forums/topic/37034-smps-and-grounding/

Optical networks due offer 100% isolation from all other comments on the network EXCEPT the last switch that converts back from optical to digital. Also, as you mentioned optical network do introduce the potential for timing delays.

If you already have one LPS-1 on hand then buying a second will suffice for your system. You can use one set at 5v for the SU-1 and one set at 7v for the ISO REGEN plus the ultraRendu with a “Y” DC power cable.

http://www.ghentaudio.com/part/dc11.html

I don’t know about DSD512 in 48kHz but I run DSD256 upsample on my 2013 MacBook Air with no problems in conjunction with an ultraRendu. That is mostly an apples to oranges comparison but my network chain would be similar to yours and that’s the part that’s actually interfacing with the DAC to play the music. I am running 2 LPS-1’s (one for uR and the other for SU-1) along with a ground SMPS for both the LPS-1’s and the input to my NetGear FS105 is grounded.

I recommend Supra CAT8 network cables. They are not as cheap as generic CAT6A but they also won’t break the bank. You can get them for starting at $44 for a pre-made 1m cable or go the DIY rout and pay $8/meter for the cable and $12/ per pair of RJ45 plugs on eBay. I’m running them now.

 

[louisxiawei]

Furthermore, with this addition I should be able to play DSD512 in 48KHz right? I'm pretty sure my 16 core computer can do it since it's running at 4.1GHz overclocked.

I think oneguy pretty much answered most of your questions. Yes, if you use ultrarendu, then DSD512 at 48kHz will be available for your Spring dac.

Regarding the optical ethernet, I see people using a pair of FMC to do the optical isolation and claiming there is a huge improvement. However, I don't intend to use such. Quite troublesome and make the audio chain complicated, yet the final endpoint is still copper ethernet. I prefer optical out and optical in. Something like: (PC → Optical PCIE → Optical Cable → Optical switch → Optical cable → NUC/NAA → USB → DAC)

I can obviously build another PC as NAA and insert optical PCIE card on the motherboard, but then the NAA will not be as compact as NUC or microrendu and I believe more noise interference will be introduced since this NAA just becomes another conventional PC. Unfortunately, there is no small NUC motherboard so far on the market have optical SFP module integrated while still support windows OS. At the moment, my optical Ethernet stops at the NAA's side. (PC → Optical PCIE → Optical Cable → Optical switch → CAT6A → NUC). I really hope Sonore can have a rendu product that have optical SFP module instead of RJ45 Ethernet port, but they might have different opinions against optical connection. So, I have to wait for future available products or make compromises.

It is worth noting that I perceive no delay when using optical ethernet. On the contrary, I feel optical ethernet connection gives a bit more faster response. I'm using OM4 LC-LC cable. Another thing I like optical ethernet is: there is no "audio-grade" optical cable so far, so less BS/ snake oil.

 

[Energy]

@oneguy @louisxiawei

Interesting. So if Ethernet connection works as such a good isolator why is there a need for more such as using optical? Is Ethernet not 100%? Using optical fiber seems to have a lot of other issues like you said with the the last converting switch as well as timing domain issues that may be problematic. I'm not too knowledgeable but in order to use the ultraRendu a Ethernet switch must be put in-between the PC and the ultraRendu? There is no direct connection? So let's say I get a NetGear FS105. It's own circuit will create some noise but it's much less than my PC is what you're saying. The grounding of the negative lead helps to further reduce the SMPS leakage into the DAC?

You recommended I have one LPS-1 and connect it to both the ISO REGEN and ultraRendu using a Y-cable but that doesn't make sense for me. The ISO REGEN's job is to isolate itself from the NUC further and if both shared the same common ground the isolation will no longer be in affect. All it can improve at that point is USB signal integrity and power output which is not much if considering the ultraRendu already does a decent job.

I'm surprised you got DSD to work with MacBook. The last time i used Mac OS I could not upsample to Native DSD. The only thing it had available was DSD over DOP which was better than standard PCM but not as good as DSD Native. I'm guessing you can do it due to having the ultraRendu within your chain.

Supra CAT8 looks nice. Thanks for the recommendation. I can't buy the DIY materials as it seems you need an expensive crimper to make the cable.

I'm glad 48KHz will be available if I use the ultraRendu. I will give many back and forth comparisons between it and 44.1 to make sure.

What are pairs of FMC? I could not find anything online in regards. I see what you mean on why you prefer optical out and optical in. With that purpose one is able to further isolate the optical switch noise due to only light being passed through. It makes total sense to me however I feel like Sonore hasn't offered that for a reason. Perhaps timing issues plays a bigger role in things than we imagine? For that reason I am going to stick to a regular Ethernet switch and ground the negative lead. I also don't want to put an optical PCI-E card in my computer as that will restrict the airflow of the graphics card and ruin aesthetics. Have you heard a big improvement from using all optical up to switch other than a more speedy sound? Wait, so this isn't like the optical found in the back of DAC's but one that looks more dual ported like the OM4 LC-LC cable that you mentioned?

 

[oneguy]

Optical prevents having long runs of wire act as an antenna to potential EMI sources. With the setup I described you are talking about a really, really low values approaching zero of EMI getting through due to the cables and upstream network components vs zero EMI noise for optical (since that would be impossible). That last optical to wire converter could introduce some type of noise but no matter what your setup is your can’t get around that unless you go with a direct connection. I’ll let @louisxiawei tackle the other optical related questions since I’ve never looked too heavily into an optical network.

On the subject of direct connections, it is possible to connect your PC directly to the uR but it must be setup as a network bridge. This does simply things but you loose the noise isolation that switches offer through their internal transformer. Now whatever noise your PC and NIC are putting out will pass to the uR.

The NetGear switches mentioned offer less noise out than your standard PC when the negative is shunted to ground from what I understand. I’d have to find John Swenson’s original post to quote values but the audible differences have been noted by many. Without the shunting to ground only low-impedance noise is blocked by the transformer(s) internal to the switch. Grounding the negative allows high-impedance noise to be blocked as well.

Thanks for reminding me to turn on Direct DSD through the options in the uR firmware. When I sold my mR I forgot to adjust that setting in the uR, lol. That being said yes, the uR allows Direct DSD playback since the uR is what is actually interfacing with the DAC or SU-1 in my case.

EDIT: Replaced “optical switch” with “optical converter”

 

[Energy]

@oneguy . Talking to you has always been educating.

Optical route makes sense and I can fathom why people go as far as using an optical PCI-E to gaining both ZERO EMI as well as 100% galvanic isolation.

For my particular setup however the ultraRendu sits 1 feet away from the PC so Ethernet seems like the simpler option. So basically you're saying regardless of optical or ethernet switch, either switches make noise and should have it's negative lead grounded. Now if one were to use a direct connection from PC to ultraRendu then noise isolation is not available due to no network switch transformer pretty much. Now my question is, does the optical path offer better noise isolation or is it the same as the Ethernet method but only better due to EMI related reasons?

If you could do it all over again, would you still go for the NetGear FS105? Would making a linear power supply for the network switch be beneficial or is the grounding trick good enough?
How big was the sound improvement once you got an NUC? Did the Supra CAT8 cable help with sound or is the construction better in terms of emi related issues only?

 

[oneguy]

@Energy ,

Always happy to help! I think it best to link back to the source material that started these revelations:

https://www.computeraudiophile.com/...ndu-ultrarendu/?do=findComment&comment=724269

The link above refers to the NetGear switches by name and model number and Cisco switches by name. I’m not sure if that’s all or some Cisco switches but the NetGear is easy to get and cheap so
I didn’t look any further. I snagged mine for $12 off of eBay (new) and they can be had all day long on Amazon for $20 (new).

I would like to clarify a point you made if I could. You mentioned above “regardless of optical or ethernet switch, either switches make noise and should have it’s negative grounded.” I wouldn’t go as far as to expand the scope to say all switches will benefit from this grounding. As stated in the following link “The property of shunting leakage from OTHER boxes only seems to work with these [NetGear] switches. Grounding the SMPS with switches other than these doesn't seem to do this.”

https://www.computeraudiophile.com/...ndu-ultrarendu/?do=findComment&comment=738368

As a result of that comment and the low cost of the NetGear switches I would limit the pool of switches that you plan on using to the previously mentioned NetGear switches.

An optical converter’s noise isolation is 100% from all other network components on the network EXCEPT for the noise from the converter itself. The ethernet grounding (with the NetGear switches) allows a significant amount of noise from the entire network and the switch to be blocked. I haven’t seen any measurements but it boils down to all “all of the last optical converter’s noise” vs “a very small percentage of the total network noise.” No idea who would win but it’s probably manufacturer and design specific.

If I had to do it over again I’d still go with the FS105. I am maybe $70-80 into this tweak including the switch and additional ethernet cable so it’s dirt cheap as far as mods go for a small but noticeable improvement. The biggest change I heard was a blacker background.

You can power the switch with a LPS but it probably won’t net you much. My switch is actually powered by an LPS not out of necessity but out of convenience. I was only using one of four 12V output from my LH Labs LPS4 so it made sense to use another output for the switch. The switch would block the low-impedence noise from whatever is powering it so that’s why I think it wouldn’t matter whether you use an LPS or SMPS.

On the subject of cables, it is my opinion that the Supra CAT8 cables helped a tiny, tiny bit with the sound compared to my generic CAT6A cables but it honestly could have been the placebo effect. I’ll keep them in there though because buying the three cables didn’t break the bank and I like the piece of mind.

(To those of you reading this who are anti-network cable’s make a difference, I did say that was my OPINION. Please no flame spray.)

EDIT: replaced “optical switch” with “optical converter”

 

[sbenyo]

Great thread!

I own a T+A DAC8 DSD and I also have ISO REGEN. I currently use hqplayer with PCM upsampling to 352k. The sound is great.
I am planning to buy a PC for DSD512. From what I can see the AMD 1920x may be the best option. Can anyone confirm if it's good enough for top XTR modes? (both 44.1k and 48k)

 

[louisxiawei]

Great thread!

I own a T+A DAC8 DSD and I also have ISO REGEN. I currently use hqplayer with PCM upsampling to 352k. The sound is great.
I am planning to buy a PC for DSD512. From what I can see the AMD 1920x may be the best option. Can anyone confirm if it's good enough for top XTR modes? (both 44.1k and 48k)

I have i9 7980XE and Titan Xp intalled. Did a few experiments myself with T+A DAC8 DSD. I can confirm you a few things:

1. When doing upsampling using poly-sinc-xtr 44.1→44.1*512, Titan Xp is pulling 7980XE's leg, causing playback stutter. So in terms of GPU's cuda offload, adding any gaming graphic card won't help for the poly-sinc-xtr filter.
2. i9 7980XE alone can do 44.1 → 44.1*512 and even more extreme setting 48K*512 at poly-sinc
3. i9 7980XE alone can do 44.1 → 44.1*512 or 48K based file → 48*512 at poly-sinc-xtr but cannot do 44.1 → 48K*512, I tried to overclock the i9 7980XE and had a few chat with Jussi (desinger of the HQplayer). Turn out it to be 44.1 → 48K*512 at poly-sinc-xtr is not only hungry for CPU horsepower but also for cache/ram speed.

As for the sounding, it is always being personal preference. I prefer 44.1 → 48K*512 to 44.1 →44.1*512 at poly-sinc. Difference is not subtle.

Poly-sinc-xtr at DSD512 is just beautifully smooth, this is a very well-tuned filter, I especially use this filter for some vocal music while ussing poly-sinc for classical orchestral music.

Can't help you with the 1920X, i9 7980XE is quite faster than 1950X, so you can have a rough idea how the settings will go.

 

[Energy]

@Energy ,

Always happy to help! I think it best to link back to the source material that started these revelations:

https://www.computeraudiophile.com/...ndu-ultrarendu/?do=findComment&comment=724269

The link above refers to the NetGear switches by name and model number and Cisco switches by name. I’m not sure if that’s all or some Cisco switches but the NetGear is easy to get and cheap so
I didn’t look any further. I snagged mine for $12 off of eBay (new) and they can be had all day long on Amazon for $20 (new).

I would like to clarify a point you made if I could. You mentioned above “regardless of optical or ethernet switch, either switches make noise and should have it’s negative grounded.” I wouldn’t go as far as to expand the scope to say all switches will benefit from this grounding. As stated in the following link “The property of shunting leakage from OTHER boxes only seems to work with these [NetGear] switches. Grounding the SMPS with switches other than these doesn't seem to do this.”

https://www.computeraudiophile.com/...ndu-ultrarendu/?do=findComment&comment=738368

As a result of that comment and the low cost of the NetGear switches I would limit the pool of switches that you plan on using to the previously mentioned NetGear switches.

An optical switches noise isolation is 100% from all other network components on the network EXCEPT for the noise from the switch itself. The ethernet grounding (with the NetGear switches) allows a significant amount of noise from the entire network and the switch to be blocked. I haven’t seen any measurements but it boils down to all “all of the last optical switches noise” vs “a very small percentage of the total network noise.” No idea who would win but it’s probably manufacturer and design specific.

If I had to do it over again I’d still go with the FS105. I am maybe $70-80 into this tweak including the switch and additional ethernet cable so it’s dirt cheap as far as mods go for a small but noticeable improvement. The biggest change I heard was a blacker background.

You can power the switch with a LPS but it probably won’t net you much. My switch is actually powered by an LPS not out of necessity but out of convenience. I was only using one of four 12V output from my LH Labs LPS4 so it made sense to use another output for the switch. The switch would block the low-impedence noise from whatever is powering it so that’s why I think it wouldn’t matter whether you use an LPS or SMPS.

On the subject of cables, it is my opinion that the Supra CAT8 cables helped a tiny, tiny bit with the sound compared to my generic CAT6A cables but it honestly could have been the placebo effect. I’ll keep them in there though because buying the three cables didn’t break the bank and I like the piece of mind.

(To those of you reading this who are anti-network cable’s make a difference, I did say that was my OPINION. Please no flame spray.)

Great post as usual! Thanks for the link! It surely did not go to waste, I ended up going through all the pages to absorb myself with this stuff.
It makes sense to why John Swensen did not push for the Cisco SG100D. It would have been hard to find the same model SMPS.

I went along as you said and purchased a NetGear FS105. Realized my listening room is not properly grounded and only utilize floating grounds (receptacle, box) so will have to call an electrician. Not sure how much leakage can be stopped with a float, probably very little as I read that ground lines have to be earthed for them to be effective. Regardless of which, I went with the procedure and have both LPS-1's negative output grounded with a wire matching the thickness of a green wire.

So the shunted NetGEar FS105 removes a significant amount of the noise from the total network. I read an article where they recommended that if you use Ethernet for internet, then it's recommended to use two converters of fiber optic. One to convert Ethernet to fiber optic and another to convert fiber optic back to Ethernet.

Example below:

With what you said however it seems easier to just use an optical network switch. Anyways, I will follow your advice and stick to the shunted SMPS. I don't have any extra linear power supplies laying around. If having one does not improve much then I'll scratch that idea. Will go for the SUPRA CAT8 cable, it looks nice!

Great thread!

I own a T+A DAC8 DSD and I also have ISO REGEN. I currently use hqplayer with PCM upsampling to 352k. The sound is great.
I am planning to buy a PC for DSD512. From what I can see the AMD 1920x may be the best option. Can anyone confirm if it's good enough for top XTR modes? (both 44.1k and 48k)

Even without a fast computer I would still pick poly-sinc-xtr-2s any day as long as you had at least a quad core.
1920X should be enough, no guarantee though. The 7900X one i used before the 1950X had 10 cores but lacked L3 cache so it still suffered from stuttering problems. The extra 2 cores of the 1920X and same amount of cache as the 1950X should be sufficient. DSD512 over 48KHz is only available if you use Linux or have a NUC like the microRendu/ultraRendu over a network used for playblack.

I have i9 7980XE and Titan Xp intalled. Did a few experiments myself with T+A DAC8 DSD. I can confirm you a few things:

1. When doing upsampling using poly-sinc-xtr 44.1→44.1*512, Titan Xp is pulling 7980XE's leg, causing playback stutter. So in terms of GPU's cuda offload, adding any gaming graphic card won't help for the poly-sinc-xtr filter.
2. i9 7980XE alone can do 44.1 → 44.1*512 and even more extreme setting 48K*512 at poly-sinc
3. i9 7980XE alone can do 44.1 → 44.1*512 or 48K based file → 48*512 at poly-sinc-xtr but cannot do 44.1 → 48K*512, I tried to overclock the i9 7980XE and had a few chat with Jussi (desinger of the HQplayer). Turn out it to be 44.1 → 48K*512 at poly-sinc-xtr is not only hungry for CPU horsepower but also for cache/ram speed.

As for the sounding, it is always being personal preference. I prefer 44.1 → 48K*512 to 44.1 →44.1*512 at poly-sinc. Difference is not subtle.

Poly-sinc-xtr at DSD512 is just beautifully smooth, this is a very well-tuned filter, I especially use this filter for some vocal music while ussing poly-sinc for classical orchestral music.

Can't help you with the 1920X, i9 7980XE is quite faster than 1950X, so you can have a rough idea how the settings will go.

1. I can confirm. CUDA offload is useless even for 1080 Ti or Titan Xp. It does not handle the filter better than the CPU standalone.
2. I will give that filter a try over 48KHz. How often do you use it? With everything I've tried I still preferred the poly-sinc-xtr. Then again I haven't tried going to 48KHz nor do I have any classical music..
3. Even with 4.2GHz on the CPU & 3600MHz on the ram I am getting stutters playing 44.1 > 48KHz and 48KHz > 44.1 KHz in DSD512 with the poly-sinc-xtr filter.