Jump to content
IGNORED

Building a DIY Music Server


Recommended Posts

6 hours ago, ASRMichael said:

In the review it states the Duelund capacitors are $1000 each, hifi collective $210. Quite a difference. Are they the same specification? 

 

FYI - here's the official statement in 2019

 

https://www.facebook.com/permalink.php?id=209346612428600&story_fbid=2557407504289154

Quote

Due to significant investments from our major distributors - a selection of the most popular CAST Silver 630v capacitors can now be offered at much more attainable price leves. 

 

And then here's the verdict

 

https://www.instagram.com/p/B3RS56aJzuz/

Quote

In contest are
- pure silver cast 1uf
- tinned copper cast 1uf
- copper cast 1uf

 

https://www.instagram.com/p/B3Wa_vkJpY3/

Quote

I would rate; - Pure Copper cast 9 out of 10 - Tinned Copper cast 9.5 out of 10.
...For the top end separation.

 

And the best affordable capacitor out there at the moment. ...So what is 10!!!? ...someone has asked me that question before. ;-),
I honestly don't know but at the moment I would say the Silver cast is 10 or even more...,

 

https://www.instagram.com/p/B3Y9gvtpMf3/

Quote

Duelund Pure Silver Cast capacitors. ...It's actually unbelievable what this silver cap does to the music!

 

It's like sitting in front of a massive wall of multilayered sound that comes from very deep black and quiet background and riches somewhere behind your head, sound of cymbals or piano has that beautiful glow around it and top-end notes sound so 3d as there are some ghosts around my head whispering to my ears,

 

it makes me listen to well-known by me tracks over and over again! wow!

 

I know it sounds crazy but the music becomes so tangible, involving and hallucinogenic that I have become a silver cast addict!. You get massive soundstage, deep and wide! super quiet background, big and well shaped bass and best top-end ever, never bright or annoying, ... I actually worried about that, thinking silver is known to be quite bright! But not in this case, Duelund silver sounds big and well balanced and takes you right into the show!

 

It's a real step up from an amazing copper family as the bigger price would suggest.

 

Unfortunately, they are not cheap, but you definitely get what you paid for.

 

I rate it 10 out of 10!

 

So far the reference! and best audio capacitor that I know of.

 

Link to comment

FYI - Marcin just spilled the beans

 

http://jplay.eu/forum/index.php?/topic/3021-new-audiophile-build/#entry52870

Quote

If you don't want any switching regulators, you will have to wait for JCAT streamer/server... ;)

 

USBridge Sig was the very first but it's only good for Compute Module 3 Lite

 

https://allo.com/sparky/usbridge-signature-pcb.html

Quote

To lower the noise of the USB, we had to lower noise on EACH rail. In total, USBridge Sig uses more than 30 LDOs. Even the SD card power rail has a dedicated LDO. There are zero switching elements on board.

 

At last we'll have something with x86 instead of ARM, could that actually meet / beat ASUS WS C621E SAGE or what?

Link to comment

LOL, you just made a certain claim about X without bothering to specify what you actually did.

 

How about stating what kinda apples-to-apples comparison was done? Volume matching? Same distro? Same kernel? Same software player on the client / server side? Same network tweaks? Same PSU with the same voltage? Same cables? Same input on the same DAC?

 

https://community.roonlabs.com/t/ropiee-on-allo-usbridge-signature/79652/27

Quote

You‘ll enjoy the USBridge Sig! I‘ve got mine, along with the Shanti LPS. The combo beats my Aurender N100H hands down (this unit costs 3,000 €!)! In my system, Roon sounds very sweet with this unit!

 

https://community.roonlabs.com/t/ropiee-on-allo-usbridge-signature/79652/48

Quote

I can confirm Paul_Jeno’s comment. I also have the USBridge Sig + LPS with a Gustard U16 DDC. It now replaces a $4k streamer… so much so I have put these with more parts and built a streamer around it inside a slick ali case with screen as a single unit. Looking at a OXCO clock and this will play with the big brands.

 

https://audiophilestyle.com/forums/topic/59219-new-dd-converters-from-denafrips/page/6/?tab=comments#comment-1071704

On 8/3/2020 at 4:16 AM, richard_crl032 said:

Just listened to the allo's usb bridge signature via the gaia and T+ and compared the Antipodes directly via the T+  with various same tracks on both. 

 

Without the gaia for both allo's and Antipodes to the T+ that I am very familiar with, it was like subjectively 60-70% for the allo's if Antipodes is at 90% ....

 

With the gaia for allo's, it gets to 75-80% of the Antipodes which is quite impressive and nice as a backup server since I have the gaia but it also tells me just how nice the Antipodes is .. crisp, clear and extended in comparison.

 


 

FWIW running a (WAY) stripped-down OS with the right CPU affinity etc. turned out to make a HUGE difference

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/693/?tab=comments#comment-1073768

https://audiophilestyle.com/forums/topic/58164-building-a-diy-music-server/page/28/?tab=comments#comment-1077902

 

IMHO Allo could only ship their products with a distro that's relatively easy to support, meanwhile plenty of tweaks are available for ARM out there and quite a few examples could be found below

 

http://albumplayer.ru/linux/english.html

https://www.symphonic-mpd.com/forum/

https://github.com/thanhtam-h/rpi4-xeno3

https://www.patreon.com/posts/snakeoil-os-33602597

http://www.headphoneclub.com/thread-738525-1-1.html

http://www.erji.net/forum.php?mod=viewthread&tid=2192344

https://drive.google.com/drive/folders/1j9c6utgwrStLYD6xy5Yati63RcMlMp8S

https://vnav.vn/threads/raspberry-pi-music-server.38335/page-184#post-2600011

https://github.com/sam0402/pcp-44.1KHz/blob/master/xenomai3/pcp-44100Hz_xe.patch

https://www.my-hiend.com/vbb/showthread.php?13157-piCorePlayer6-1-0-Xenomai-44-1-48KHz雙機入門簡易安裝教學&p=242868#post242868

Link to comment

http://jplay.eu/forum/index.php?/topic/3594-optimo-atx-truly-linear-atx-power-supply/#entry52874

Quote

Initio and NET Card XE will be announced later this month

 

Now that NET Card XE is coming soon, it's only natural to ask a question like this — could some sorta isolation be achieved by binding audio-related process(es) like HQPlayer / NAA etc. to that specific network adapter exclusively? Meanwhile everything else will be forced to go through the on-board Ethernet and then we're making sure that no processes are allowed to access both adapters under any circumstances.

 

Is it really as simple as something like this?

  • On-board Ethernet — IPv4 = Enabled / IPv6 = Disabled
  • JCAT Net Card XE — IPv4 = Disabled / IPv6 = Enabled

Maybe we could even unplug the network cable from the on-board Ethernet at some point, though not really sure about how to keep the remote control part going.

 


 

Let's say it ain't so easy, could these programs do the trick by any chance?

 

https://github.com/falahati/NetworkAdapterSelector

https://www.raymond.cc/blog/bind-windows-application-to-specific-network-adapter-with-forcebindip/

 

And then we've got something else for Linux as well

 

https://nsjail.dev

https://github.com/google/nsjail

 


 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/656/?tab=comments#comment-1053701

On 5/22/2020 at 3:11 PM, darkfrank said:

I must say that even I'm already using the Telegartner M12 switch, by using ROON to play local file only, unplug the network cable still improve the SQ for about ~20%.

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/699/?tab=comments#comment-1077821

On 9/4/2020 at 12:56 AM, austinpop said:
Disconnected is best: My preferred mode of listening is local playback (from local SSD) using HQPlayer, with the network disconnected. Not only does this sound sublime, but it forms a reference for what the network chain needs to achieve.

 

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/page-206#post-655430

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/676/?tab=comments#comment-1063711

On 6/27/2020 at 1:04 PM, Nenon said:

Also, the recent developments at Taiko got me intrigued. They are into something that made Roon much better. 

This Asus Sage motherboard has KVM access. I can remote to my server while the NICs are disabled. What I have noticed is that disabling the NICs is an improvement.

I think there are two types of network noise. One is the electrical noise that comes over the network cable, which everyone is talking about and trying to clean up, isolate, reclock, etc. The other one is the noise that the computer generates when the Operating system is processing packets. We don't talk much about the latter. And I think that is the more harmful noise (at least after a good number of network tweaks). Every packet that comes to your network needs to be processed. That's why we have discovered that isolating your server on a seperate VLAN sounds better - your server does not have to process random broadcast packets from other devices on your network. This is also one of the reasons why we found that Stylus sounds better than Roon. Stylus caches the track before playing and tries to minimize any network activity. Roon constantly does something with the network. There is noise that is generated every time a packet is processed. Those are also CPU interrupts that are distracting the CPU from its main purpose - music playing. And no switch can do anything about that. I discovered this when I connected to my computer with KVM and disabled the NICs.

 

I expect that not everyone would actually play something on YouTube etc. with command-line based software (i.e. MPV Player + FFmpeg + youtube-dl = no ads whatsoever) like I do, though it's still quite interesting to try something like this

 

https://www.stereo.net.au/forums/topic/298457-ethernet-switches-for-audio-part-a-list-of-switches-related-info-experiences/page/7/?tab=comments#comment-4598801

Quote

Probably not the right place but related nonetheless. @rmpfyf do I remember correctly that you were able to get your network port to switch off automatically when a playlist starts and then switch back on when it has finished?

 

https://www.stereo.net.au/forums/topic/298457-ethernet-switches-for-audio-part-a-list-of-switches-related-info-experiences/page/7/?tab=comments#comment-4598833

Quote

Create a file e.g. 'nano NetworkDown.sh'

 

In the editor

 

sudo ifconfig eth0 down
aplay filename
sudo ifconfig eth0 up

 

save that and exit

 

Make it executable e.g. 'chmod +x NetworkDown.sh'

 

Then run it e.g. './NetworkDown.sh'

 

Should do it. If you replace 'filename' in the file with "$1" then you can run the script and pass the filename as an argument e.g. './NetworkDown.sh myfile'

 


 

Now the $64,000 question is, do we have anything out there that could (automatically) detect the beginning / the end of track so that all NICs will be disabled / re-enabled accordingly?

Link to comment

FYI - here's the latest from Soundaware and their flagship product AMC-D1 looked very so close to ASRock J4205-ITX

 

https://weibo.com/1458427661/JqEti7OHf

https://weibo.com/5544444327/JqED2oyF8

http://www.soundaware.net/newsinfo/459500.html

1syZg0C.png

9QeLm8b.png

ASRock J4205-ITX

https://www.asrock.com/mb/Intel/J4205-ITX/index.asp

YLivzcNl.png

 

Supposedly they spent like 3 years on the development. Basically they evaluated thousands of motherboards and less than 10 of them were able to meet their requirements, finally they chose to work with ASRock instead of Asus and Gigabyte etc.

 

https://card.weibo.com/article/m/show/id/2309404563632272048236

https://card.weibo.com/article/m/show/id/2309404563892683538661

https://card.weibo.com/article/m/show/id/2309404563901151838380

r77J6O2.png

Two different flavors of AMC-D1 will cost 14,999 RMB (Value Edition) and 14,999+1?,??? RMB (Deluxe Edition) respectively, the latter one should come with custom BIOS / stripped down version of Windows 10 IoT Enterprise LTSC etc. They mentioned that even ASIO driver was optimized for USB audio output.

 


 

Obviously that Pentium J4205 wouldn't be any good when compared to stuff like Xeon Scalable etc. since a more powerful CPU should sound (much) better under most circumstances

 

https://ark.intel.com/content/www/us/en/ark/products/95591/intel-pentium-processor-j4205-2m-cache-up-to-2-6-ghz.html

 

Though their J4105-ITX (w/ Celeron J4105) still sounded better than Supermicro X10SBA-L (w/ Celeron J1900) inside Innuos ZENith Mk.II Std

 

https://systems.audiogon.com/systems/2598

https://www.asrock.com/mb/Intel/J4105-ITX/index.asp

https://audiophilestyle.com/forums/topic/55916-euphony-os-wstylus-player-setup-and-issues-thread/page/58/?tab=comments#comment-1086740

On 10/27/2020 at 7:34 AM, Anwar said:

I played with different J4105 BIOS settings but settled with the following: Turbo Boost: OFF, SpeedStep: OFF, C-State: OFF.  With the Celeron J4105 only having 4 cores, I settled with "0-1 gstp 2 stylus 3" CPU isolation setting.

 

Wow, the SQ is so close to the SQ of my DAC's own rendering.  It even beats my Innuos ZENith Mk2.Std.  It surprises me as I don't need to spend crazy money on expensive USB reclocker.

 

Link to comment
On 10/11/2020 at 8:31 AM, Exocer said:

You're welcome, all I did was look up LVCMOS OCXOs @ 25Mhz on Digikey and @Soul Analogue confirmed the results. 

 

@tgb, looking forward to your project with the OH4510LF. Best of luck, and hopefully the results are worthwhile!

 

OH4610LF-025.0M = ±10ppb @ $55.1 (Minimum 1)

 

https://www.digikey.com/en/products/detail/connor-winfield/OH4610LF-025-0M/5641634

 

SIT5711AI-KW-33N-25.000000N = ±5ppb @ $59.68 (Minimum 500 = $29,840)

 

https://www.arrow.com/en/products/sit5711ai-kw-33n-25.000000n/sitime

https://www.sitime.com/part-number-decoder

atIhA0I.png

 


 

Any SiT5711 OCXO @ 25 MHz = 1,080 RMB / $165 (Minimum 1)

 

https://item.taobao.com/item.htm?id=585481394277

 

SIT5711AI-KW-33N-25.000000F = $106.25 (Minimum 25 = $2656.25)

 

https://www.arrow.com/en/products/sit5711ai-kw-33n-25.000000f/sitime

https://www.sitime.com/part-number-decoder

bELIQU4.png

 


 

A wider temperature range (-40 to 85 as shown above) should be superior to what JCAT picked for their USB Card XE (20 MHz) and NET Card XE (25 MHz) as follows

 

https://audiophilestyle.com/forums/topic/58899-usb-card-xe-xtreme-edition-usb-audio-output-with-ground-breaking-performance/page/2/?tab=comments#comment-1065007

On 7/3/2020 at 1:57 PM, Marcin_gps said:

The current lead time information for the OCXO SIT5711AC-KW-33E-20.000000F-ND is that they are expected to arrive at Digi-Key on July 8, 2020.

tO7Ma2v.png

 


 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/644/?tab=comments#comment-1039516

On 3/31/2020 at 6:40 AM, romaz said:

According to Marcin, it uses a new OCXO rated at 5ppb but what he found to be the more important characteristic is that this clock is somehow immune to vibrations.  Apparently but not surprisingly, Marcin and his OEM (Adnaco) found that a high performance clock in a vibratory environment has the potential to perform more poorly than a lower performance clock in a vibration-free environment and so while others may be touting OCXOs with better lab measurements under ideal conditions, this card was designed to function optimally under real world conditions.

 

Minimum 500 would be quite a stretch unless we're able to split $29,840 among 50 or so participants, though $2656.25 wouldn't be THAT bad for 25 pieces of high caliber OCXO.

 

AFAIK plenty of stuff could take advantage of an OCXO @ 25 MHz since both Ethernet (NICs / switches etc.) and USB 3 could perform much better with an upgrade.

 


 

BTW, here's their latest DCOCXO that's even more interesting

 

https://www.digikey.com/en/products/detail/sitime/SIT5721AC-KW333-T-24-57600F/11696570

https://www.arrow.com/en/products/sit5721ai-kw033-t-50.000000/sitime

https://media.digikey.com/pdf/Data Sheets/SiTime PDFs/SiT5721_Web.pdf

Quote

The SiT5721 can be factory-programmed to any frequency between 1 and 60 MHz. In addition to industry leading performance, the DCOCXO is offered in a 9 x 7 mm package, the smallest Stratum 3E OCXO available. 

 

Theoretically we could get 45.1584 MHz / 49.152 MHz for upgrading DACs etc. but their lead time might be a concern.

Link to comment

FYI - interesting post

 

https://cafe.naver.com/jazzkfpa

https://audiophilestyle.com/forums/topic/60871-korean-no-1-high-end-linear-power-supply/

Quote

Nayuta Flagship LPS is 3 rails / 4 rails / 5 rails.

 

- 3 rails :  40Kg / Price (Included Shipping Fee) : $ 7,000

- 4 rails :  50Kg / Price (Included Shipping Fee) : $ 8,500

- 5 rails : 62Kg / Can't sell this version to the world

 

Nayuta Audio Linear Power Supply

https://www.monoandstereo.com/2020/02/nayuta-audio-linear-power-supply.html

https://www.facebook.com/permalink.php?id=1800464690274533&story_fbid=2572476489740012

 


 

https://www.custom-hifi-cables.co.uk/power-supplies/dc4-power-supply

Quote

£3000 (4000 USD for North America)
(1 output) e.g. Chord M-Scaler

 

£3700 (5000 USD for North America)
(2 outputs) e.g. a switch/clock combo

 

£4400 (6000 USD for North America)
(3 outputs) e.g. Chord DAVE or BLU mk2

 

Link to comment

The thing is, not THAT many of us could / would have access to enough SR7 DR / DC4-tier (maybe OPTIMO_ATX down the road, too?) rails for obvious reasons

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/614/?tab=comments#comment-1011955

On 12/10/2019 at 4:25 PM, austinpop said:

This super-combo just smokes both the red and blue combos! I have a hard time even articulating how good this sounds. The HDPlex Linear ATX supply is already very good, but using the SR-7 to power the server just takes it to another level. SQ improves across the board, in dynamics, transparency, bass articulation, tonality and imaging. I finally got to hear what @romaz has been raving about, and it is truly something!

 

And then there's something about having the right software that could actually take advantage of those substantial investments in hardware

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/674/?tab=comments#comment-1062580

On 6/23/2020 at 12:43 AM, austinpop said:

What my experience showed me is that the 8700K sounded better than the 8700T at the same operating frequency:

  • IF the PSU is big enough to easily energize both CPUs for all the factors above, and
  • when using the performance governor (default) in Euphony that sets and holds the CPU frequency to the max value specified in the settings.

Does this prove anything? Nope. And if asking on any other OS than Euphony, without knowing how the OS will control CPU frequency dynamically, I wouldn't be able to say. But certainly for Euphony, my inclination would be to go with the K CPU, if you're confident of your PSU's capacity.

 

https://audiophilestyle.com/forums/topic/58164-building-a-diy-music-server/page/28/?tab=comments#comment-1077902

On 9/4/2020 at 6:07 AM, Nenon said:

It's the sum of all parts that matters. When I built my dual Xeon CPU server and ran Euphony it did not sound better than my single CPU AMD Ryzen with pinkFaun ultraOCXO motherboard clock. 

But after spending a few months to tweak a custom Windows LTSC, and after isolating the server and the streamer part on different CPUs, the dual Xeon is on another level. 

So there is no simple answer to your question. A single CPU server if tweaked well may sound better than a not so-well tweaked dual CPU server. There are so many variables, that it's hard to generalize. 

 

It's kinda like putting the right person, in the right place, at the right time.

 

Hardware, power, cables, server / client combos, software player, bit-perfect versus upsampling, CPU affinity, process isolation, ramroot / RAM drive, BMC + disabling NIC, CPU offload with RDMA / SMB Direct, kernel timer frequency, BFS / MuQSS CPU scheduler, Xenomai kernel with RTDM, compiling codes with GCC settings for a specific CPU type, so on and so forth.

 

Since it's a thread for subjective discussions, just about any kinda truth / falsehood / truthiness etc. could be both right and wrong IMHO.

Link to comment

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/702/?tab=comments#comment-1084936  

On 10/16/2020 at 2:27 AM, Nenon said:

I may try multiple runs of it in a power cable one day (with some vibration damping and Bocchino or Furutech NCF connectors).

 

No kidding, even RJ-45 went NCF for real

 

http://www.furutech.com/2020/10/05/20075/

http://www.furutech.com/ja/2020/10/02/20066/

EVVKrRe.jpg

zAv6aca.jpg

Quote

Main conductor: 24 AWG Silver plated α (Alpha) OCC conductor
Insulation: Special Foamed PE OD.: 1.35±0.05mm, 4 Twisted Pairs
Blue/White; Orange/White; Green/White; Brown/White
Shield 1: AL foil Conductive layer to the outside
Shield 2: AL foil Double-sided conductive
Shield 3: α (Alpha) Pure copper wire (16/10/0.10) braiding layer
Jacket: UL/CL3 approved flammability grade; RoHS Compliant Flexible PVC (Black)
Sleeve: High quality nylon yarn braided (Purple/Black)

 

Really interesting cable for this Intona POEsy switch then?

 

https://intona.eu/en/products/poesy

https://intona.eu/en/products/buy/poesy/

 

FYI - an audiophile in China compared Intona to stock Melco S100 recently, the latter one didn't seem to be THAT much better despite such a significant price difference between the two

 

http://www.headphoneclub.com/forum.php?mod=viewthread&tid=740410&page=3#pid5228491

Link to comment
  • 1 month later...
1 hour ago, Dev said:

The reason why Horrender has stolen software from the start

 

Re: [mpd-devel] Request for GPL source code of your Aurender A10 product

https://www.mail-archive.com/[email protected]/msg00669.html

Quote

I already told you twice that your MPD source doesn't match your binary. The rules of the GPL mandate that you must provide the matching source code for all binaries you distribute. If you can't follow those rules, then your license is terminated automatically, making your product illegal.

 

Now I could tell you how I verified the mismatch, but then you'd change that one piece in the source, and we're going to play endless whack-a-mole where I can only lose. That's a poor way of solving something that is naturally only your problem. So no, sorry, I can't let you know right now. It's your turn, not mine. You need to resolve your copyright problem. I am the victim, not the defendant.

 

Arch Linux, MPD and Kernel GPL violations by Euphony Audio

https://bbs.archlinux.org/viewtopic.php?pid=1778399#p1778399

Quote

The problem with Euphony is that their written offer is fake. They never intended to give their modifications to the kernel to those who received the binary version.

 

They put up a fake GPL written offer page, in order to silence package authors (such as Max Kellermann), but they never honored the written offer. They used all kinds of delay tactics, and when we finally provided an address to them to send the CD to, they suddenly claim they no longer offer the trial version and therefore don't need to provide us anything. This is not a valid excuse. The GPL does not work this way.

 

https://bbs.archlinux.org/viewtopic.php?pid=1779602#p1779602

Quote

They finally did send me a link with the kernel source, so I consider this case closed.

 

For those who want to optimize audio playback under Arch, here's the link of their kernel:

 

http://audiokernel.com/linux41src.tar.gz

 


 

Somehow things fall apart whenever commercial interests are involved, while no products could be sold without involving any commercial interests whatsoever.

Link to comment
4 hours ago, MarcelNL said:

Storing all content in RAM is a bit hard, streaming means investing in more hardware as I also listened to the effects of network cables routers etc...choices...

 

https://audiophilestyle.com/forums/topic/58164-building-a-diy-music-server/page/43/?tab=comments#comment-1088701

On 11/9/2020 at 3:00 PM, Nenon said:

I am using this one at the moment - https://www.amazon.com/gp/product/B07776HSPW/

Moved to PCIe based Optane and NVME storage with the dual Xeon CPU, because I can attach them directly to the CPU of my choice, bypass the chipset, use more capacity, etc. 

The P4800X sounds better than the 900P! And the PCIe version sounds better than the M.2 version (P4801X).

I can only imagine what an externally powered PCIe P4800X with an upgraded OCXO clock could do!

 

1.5 TB version of P4800X turned out to cost 22,988 RMB (about 3,500 bucks) on TaoBao or roughly $5,000 on Amazon

 

https://item.taobao.com/item.htm?id=603137014053

https://www.amazon.com/SSDPED1K015TA01-P4800X-2Heigh-PCIex4-XPoint/dp/B077BMV9DD

 

Dual Xeon E5-2600 V2 (Ivy Bridge) motherboards with 24 DIMM slots are available for 500 RMB / $150 / £90 a pop

 

https://item.taobao.com/item.htm?id=632593953534

https://unixsurplus.com/x9dri-ln4f-v1-20a-motherboard-only/

https://www.bargainhardware.co.uk/supermicro-cse-815-v2-configure-to-order

 

Those "ancient" Ivy Bridge-based Xeon processors are worth next to nothing these days

 

https://www.serversupply.com/PROCESSORS/Intel Xeon 6-Core/2.4GHz - 7.2GT QPI/INTEL/SR1AZ_198913.htm

https://www.serversupply.com/PROCESSORS/Intel Xeon 6-Core/2.4GHz - 7.2GT QPI/INTEL/CM8063501376200_198863.htm

 

Right now 64GB of DDR-3 LRDIMM costs only 480 RMB, two dozens of them would be more like 11,520 RMB / approximately $1,750 for 1.5 TB of RAM

 

https://item.taobao.com/item.htm?id=623606011431

FqhoQDs.jpg

 

It's getting closer and closer to a dollar / GB of DDR-3 RAM now, though it's hard to tell if that were better sounding than Optane (PCIe) or otherwise

 

https://www.crucial.com/articles/about-memory/difference-between-speed-and-latency

https://pcper.com/2018/12/intels-optane-dc-persistent-memory-dimms-push-latency-closer-to-dram/

u0gpL2G.jpg

 

As usual it also depends on the software that determines the performance of RAMDISK

 

https://thessdguy.com/an-nvdimm-primer-part-2-of-2/

Quote

But this is all housekeeping, and doesn’t take full advantage of the NVDIMM’s speed advantage over an SSD or HDD.  More software needed to be developed to prevent NVDIMM accesses from being bogged down by slow I/O routines that were developed for HDD and SSD.  At the time that they were written these routines were significantly faster than the drives themselves.  With NVDIMMs the opposite is true: The storage hardware (the NVDIMM) is significantly faster than the I/O routines accessing it.

 

Link to comment

https://buy.advantech.com/968TW19HL0-968TW19HL0/968TW19HL0/model-968TW19HL0.htm

https://audiophilestyle.com/forums/topic/59727-windows-ltsc-install-for-music-server/?tab=comments#comment-1063212

On 6/25/2020 at 7:10 PM, dminches said:

After that I purchased an LTSC high-end license which was required since I am running a Ryzen 3700.  They have to mail the license to you since the CLA agreement requires you to put a sticker on the case.

 

Cost was about $150 in total.

 

Or simply do a search on Google etc.

Link to comment
  • 2 weeks later...

http://www.mics.ne.jp/~cdorya/#download

http://jplay.eu/forum/index.php?/topic/3063-pink-hq-minorityclean/page-42#entry53432

Quote

Holy mother of god. The depth, the low-level details, the clarity and purity of MC81 is breathtaking. MC81>>MC79

 


 

BTW, it could be fun to steal some components from this OneCoreUpdateOS / Windows 10X

 

https://forums.mydigitallife.net/threads/windows-10x-sdk-and-emulator-image-download.81144/page-2#post-1577888

Quote

6) UPDATEOS.wim located in MainOSDisk looks to be an incredibly slimmed down version of WinPE, but its edition ID is OneCoreUpdateOS. Seems to be completely new, and contains a select few standard Win32 programs like command prompt.

 

https://www.reddit.com/r/Windows10/comments/je7tk8/minimal_windows_10_onecoreupdateos_but_with/

Quote

This is OneCoreUpdateOS but the UpdateOS "Customizations" have been removed, leaving only OneCore with win32kmin, gdi32min, and other "minimal" OS components left, and BootShSvc was added to enable launching CMD at startup.

 

https://store.rg-adguard.net

https://www.microsoft.com/en-us/p/windows-10x-emulator-image-100195780-preview/9pjss0px0j6k

Quote

Microsoft.Windows10XEmulatorImage10.0.19578.0Previ_1.0.1.0_x64__8wekyb3d8bbwe.msix

 

https://betawiki.net/wiki/Windows_10X

https://thecollectionbook.info/windows/10x-v21h1

https://twitter.com/stroughtonsmith/status/1228089162697453570

https://msbuilds.rg-adguard.net/GetInfo?id=cfe185fb-c5a3-43b1-91a8-b0706dfb1c49

https://osg.wiki/books/windows-10x/page/installing-windows-10x-(emulator-image)-on-real-hardware

Link to comment
On 1/8/2021 at 7:05 AM, Nenon said:

The purpose of this chassis is to go with a specific DIY recipe utilizing the dual CPU Asus SAGE motherboard for now. If we find something that sounds better, it can be tweaked in that direction. There is plenty of space inside for anything and plenty of cooling capacity.

 

I know this one is only good for single Xeon but just wondering if you or Emile were getting a chance to test it out?

 

https://rog.asus.com/motherboards/rog-dominus/rog-dominus-extreme-model/

Quote

Designed for extreme performance: 32 power stages, dual 24-pin, quad 8-pin and dual 6-pin 12V power connectors that deliver industry-leading power efficiency

 

Granted it's meant for that Xeon W-3175X monster, though Xeon Silver with 85W TDP could still work just fine with the latest BIOS release

 

https://rog.asus.com/motherboards/rog-dominus/rog-dominus-extreme-model/helpdesk_cpu/

 

12 DIMM for a single CPU Scalable instead of the same number for dual Xeon, just wondering if that were able to score some points when combined with insane power stages / dual ATX / quad EPS etc.

Link to comment

If an expensive LPS weren't providing enough current, better be safe than sorry then?

 

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/page-97#post-615725

Quote

While the Extreme consumes only about 60 watts during music playback, Emile tells me that the Extreme can consume up to 300 watts during bootup. This would fry any of the bespoke PSUs I have on hand.

 

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/page-132#post-623214

Quote

We have tested HQPlayer version 3.x with the Extreme, DSD upsampling using the most taxing filter back then, polysinc-xtr (non -2s), it can run on 1 CPU up to DSD256, for DSD512 you need both CPUs, at which point it draws about 330 watts from the wall. The EC modulators in HPlayer 4 use even more resources, this is really beyond reasonable assuming using a linear powered server.

 

Link to comment

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/695/?tab=comments#comment-1074185

On 8/17/2020 at 4:17 AM, romaz said:

The process has been educational and what I am finding is that the quality of the files that you generate with HQP Pro is very much dependent on the quality of the server used to create these files.  Don't think you can just create files on any server and end up with excellent results.

  

On 8/17/2020 at 4:17 AM, romaz said:

my experience with creating new masters via offline upsampling with HQP Pro strongly supports that noise created by your server becomes permanently embedded in the new recording and even play back of these files through the Extreme won't fix it.

 

Just thinking out loud here, and feel free to call me crazy if you will.

 

It's only good for those of us who are willing to purchase HUGE hard drives, and obviously that will only work for NAS / locally-stored music library but no love for (although there's actually a trick or two) music streaming services.

 

Instead of having to worry about high current draw of PCM / DSD upsampling on the fly, does it make sense if everyone who's interested in this thread were willing to share the costs of building something together? That's meant to be THE ultimate server for running HQPlayer Pro and there must be a way to automate the entire upsampling process.

 

Basically we're uploading part of our own music library to the cloud first, then those files will be scheduled to be transferred to that extra-special server mentioned above. A script with specific HQPlayer Pro settings will be executed automatically and a password-protected archive (containing all upsampled files) is uploaded back to the cloud for a number of days, of course it will be deleted later once the owner is done with the downloading.

 

FYI - the size of DSD1024 files should be roughly 330 MB / min

 

https://samplerateconverter.com/educational/dsd-files

 

There's always a catch, though. How do we find someone who's trustworthy + savvy enough to be in charge of a server like that? As usual time is a concern, there's only one server that's upsampling one file at a time. The queue would keep growing if more and more forum members were interested in joining the project. Maybe CUDA offloading could speed things up with offline upsampling, though?

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/694/?tab=comments#comment-1074085

On 8/16/2020 at 8:24 PM, romaz said:

HQP upsampling also benefits from CUDA cores with the Nvidia Pascal GPU as a minimum and the Volta/Turing GPUs being better still and so anyone doing serious DSD upsampling should consider incorporating such a GPU but note that these are power hungry and noisy devices.

 

Fortunately Turing T4 (with 2,560 CUDA cores) is passively cooled @ 70W TDP

 

https://www.dihuni.com/product/nvidia-tesla-t4-gpu-16gb/

https://www.etb-tech.com/dell-nvidia-tesla-t4-16gb-graphics-card-ppgxg.html

https://www.ebay.com/itm/NVIDIA-Tesla-T4-GPU-16GB-GDDR6-Accelerator-Card-new-/203069728865

https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/tesla-t4/t4-tensor-core-datasheet-951643.pdf

https://www.wiredzone.com/shop/product/10028713-supermicro-gpu-nvtt4-nvidia-tesla-t4-16gb-gddr5-pcie-3-0-x16-1915

 

Cloud storage / bandwidth / hard drives etc. do cost money, and then HQPlayer Pro also costs €2,390. Gotta find the right person, the right place, the right time then?

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...