Building a DIY Music Server

[jean-michel6]

On 12/13/2020 at 4:40 PM, ASRMichael said:

The best sata cable is no data cable! I recently turned on sata controller to listen to internal storage. Wasn’t on for long. Even with good LPS, it’s just too noisy. 

Fully agree .... sata turned off is really the way to go . Sata is so noisy .... also turned of usb 3 if possible ! 

[jean-michel6]

On 12/13/2020 at 4:45 PM, MarcelNL said:

You'll need to store the data somewhere, I don't envision going the server client route if I can help it...I hear what you are saying, I listened to a couple of SATA cables and PSUs and was amazed how much that affects SQ

 

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...

A 2,5” hdd on usb is better than ssd on sata  . 

[MarcelNL]

haha, so that is where I am, a HDD on USB...

 

I found the audio files on the SSD, they are in one of the user subfolders. 

 

I will compare after I listened to the Optane...tonight when the grid is cleaner and the house is quieter...

[jean-michel6]

19 hours ago, MarcelNL said:

thought the Optane supposedly sounds better due to lower noise than SSD running the OS, I already have 64GB RAM installed and Daphile fits easily in a tiny portion of it as it is.

I am also using an optane 32 go card with the Daphie os on it , a nice sq bump compared to running from a usb stick .

Have you ever tried to put music files on the optane drive and running from there ? 

[deama]

3 hours ago, StreamFidelity said:

 

I also consider low latencies to be a key to increasing SQ. The setting in the ASIO USB driver depends on the sample rate. With DSD 256, 32 samples can be specified as the lowest buffer. Then it is 1.18 ms input latency and 1.50 ms output latency for USB.

 

With LatencyMon the process latency can be measured. I have measured the sensationally low value of 12.20 µs = 0.0122 ms since using the XILINX Solarflare Flareon Ultra SFN8522 PCIe card.

 

 

My tips:
- Check the process latency with every change. Disturbing drivers are displayed with LatencyMon. A driver update usually helps.
- Select and configure everything that transports data according to low latencies (CPU, USB, Ethernet, RAM, SSD, etc.)
- Disable unnecessary processes

I'm not sure what you mean, I've set my buffer size to 16 samples and I get a 70ms delay when measuring, but in latencymon I get similiar results to you. Latencymon I think only measures the latency between the usb and computer, which is super fast, even when using usb 2.0; the problem seems to be that windows does some processing or something and that adds lots of delay.

[MarcelNL]

35 minutes ago, deama said:

I'm not sure what you mean, I've set my buffer size to 16 samples and I get a 70ms delay when measuring, but in latencymon I get similiar results to you. Latencymon I think only measures the latency between the usb and computer, which is super fast, even when using usb 2.0; the problem seems to be that windows does some processing or something and that adds lots of delay.

From what I see you are measuring ALL of the period between creating an impulse and the microphpne registering it, if I'm not mistaken there is little you can do once the impulse leaves your computer. So I assume most folks here are focussing on latency within the computer, everything thereafter is a composite of other gear and a different phenomenon.

[MarcelNL]

1 hour ago, jean-michel6 said:

I am also using an optane 32 go card with the Daphie os on it , a nice sq bump compared to running from a usb stick .

Have you ever tried to put music files on the optane drive and running from there ? 

I plan to do so tonight....IF I get Daphile to copy a file to another location ....probably IRL not very practical as I'd need to copy a couple of albums to the Optane drive (16Gb) remove the HDD and listen-repeat...

I expect the results will be almost similar to playing an album from memory with the external drive removed and Optane only used as OS ROM, but we'll see....

 

The hunt for the evasive Apacer ECC Ram continues....what is it that I so frequently end up on a quest for unobtanium...hope this one is easier than getting a Shigefusa Kitaeji Santoku in 200mm.

[ASRMichael]

16 minutes ago, MarcelNL said:

I plan to do so tonight....IF I get Daphile to copy a file to another location ....probably IRL not very practical as I'd need to copy a couple of albums to the Optane drive (16Gb) remove the HDD and listen-repeat...

I expect the results will be almost similar to playing an album from memory with the external drive removed and Optane only used as OS ROM, but we'll see....

 

The hunt for the evasive Apacer ECC Ram continues....what is it that I so frequently end up on a quest for unobtanium...hope this one is easier than getting a Shigefusa Kitaeji Santoku in 200mm.

Apacer ECC ram is here

 

model - D31.23185S.001

 

https://www.mouser.co.uk/ProductDetail/apacer/d3123185s001/?qs=Cb2nCFKsA8prqNusSnCNZw==&countrycode=GB&currencycode=GBP

[StreamFidelity]

19 minutes ago, ASRMichael said:

Apacer ECC ram is here

 

model - D31.23185S.001

 

Thanks for link. 

I don't know if it means anything. The RAM latency is rather average: CL19 / 2666 MHz = Calculated nanoseconds 14.25.

I use CL10 / 2400 MHz = Calculated nanoseconds 8.33.

[MarcelNL]

18 minutes ago, ASRMichael said:

Apacer ECC ram is here

 

model - D31.23185S.001

 

https://www.mouser.co.uk/ProductDetail/apacer/d3123185s001/?qs=Cb2nCFKsA8prqNusSnCNZw==&countrycode=GB&currencycode=GBP

TA, mate ;-)

 

I'll try check if it fits and will order, plenty of time to tinker now we are entering our second lockdown.

[MarcelNL]

1 minute ago, StreamFidelity said:

 

Thanks for link. 

I don't know if it means anything. The RAM latency is rather average: CL19 / 2666 = Calculated nanoseconds 14.25.

I use CL10 @ 2400 MHz = Calculated nanoseconds 8.33.

not everything works around latency, transistors like tubes all sound different even those with similar slew rates.

[ASRMichael]

Just now, MarcelNL said:

TA, mate ;-)

 

I'll try check if it fits and will order, plenty of time to tinker now we are entering our second lockdown.

If in the states, just click top right. US in stock also. 

[MarcelNL]

4 minutes ago, ASRMichael said:

If in the states, just click top right. US in stock also. 

Netherlands here, I had the Dutch subsidiary phone nr on my desk for tomorrow...this is easier!

 

biggest question now is...how many, 1 -2 or 4 *4Gb

 

 

I think it should work :-)

 

MB manual states:

4 x DDR4 DIMM sockets supporting up to 64 GB of system memory Dual channel memory architecture Support for DDR4 2933/2667/2400/2133 MHz memory modules Support for ECC Un-buffered DIMM 1Rx8/2Rx8 memory modules (operate in non-ECC mode)

 

[ASRMichael]

6 minutes ago, StreamFidelity said:

 

Thanks for link. 

I don't know if it means anything. The RAM latency is rather average: CL19 / 2666 MHz = Calculated nanoseconds 14.25.

I use CL10 / 2400 MHz = Calculated nanoseconds 8.33.

I’ve never tested the difference from CL10 to CL19 myself. If I recall correctly it was Marcin at JCAT & Nenon that reported the CL19 was a step up on previous versions. Is the CL10 a newer or older model? 

[ASRMichael]

If anyone interested This is the NUC version. Apacer SO-Dimm. 1 x 8gb
 

https://www.mouser.co.uk/ProductDetail/apacer/d2323240s004/?qs=GBLSl2AkiruWco00G6Jb1Q==&countrycode=GB&currencycode=GBP

[MarcelNL]

anyone had to take a dremel to their chassis to be able to use a proper power cord? I find that my nice Viborg pure copper Schuko connector won't fit due to the location of the connector in the chassis 🥵 

[MarcelNL]

5 hours ago, jean-michel6 said:

I am also using an optane 32 go card with the Daphie os on it , a nice sq bump compared to running from a usb stick .

Have you ever tried to put music files on the optane drive and running from there ? 

 

I am comparing the two, and with somewhat surprising, well baffling rather,  results. 

All fairly quick comparisons (30 sec or so, with not more than 5-10 sec between them), done using the same track by the late Johnny Cash, 'unearthed' track 14 (alternate take) of 'down there by the train'

 

first round: externally fed SSD vs Optane

SSD sounded more organic, better separation of layers and overtones and a better 'view' of the original acoustics than the Optane that sounded more mechanical, rougher. Mind you, small differences but at this level important ones.

 

Doing that same comparison four five times showed a change and therefore a preference for the Optance, as if it's performing better over time (it was already on for 10hrs).

 

next round; RAM buffer fed from SSD vs RAM buffer fed from Optane:

small difference, almost negligable, hard to tell, no real preference

 

last round; 

SSD vs Optane with SSD power removed

 

CLEARLY better, cleaner, blacker background, more depth in sounds, better dynamics etc., sort of the 'coming home'experience, more organic, rounded yet maintaining dynamics etc, as if a thin veil is lifted etc etc, the usual high end babble of superlatives.

 

So now I'm in a conundrum having a clear preference for playing from Optane....to invest $$$$$ in a huge Optane PCIe disc?

 

Or better, definitely for my wallet, how to automate buffering tracks to Optane when playing stored music ....much like a RAM disc?

 

A second small (affordable) Optane unit could sit on a PCIe riser (that is IF there is space on the MB, that Pink Faun I2S card is WIDE. Currently Daphile shows I have some 12Gb free on the 16 Gb Optane stick, should be plenty for buffering....but how.

 

 

Have to say that without a Linear PSU this bit of kit already sounds better than the NUC with a nice PSU

 

I will make the same comparison later, it's easy enough now I can find both locations with the same track, once the Optane has had some time to break in as it supposedly was brand new so I may be hearing that effect (doubt it).

 

I will also copy the track to another location on the SSD and the Optane from one and the same USB stick so I am comparing  same version....as there are 'some' that say each copy sounds slightly different (and I have reasons to believe that).

[Kal Rubinson]

4 minutes ago, MarcelNL said:

next round; RAM buffer fed from SSD vs RAM buffer fed from SSD:

small difference, almost negligable, hard to tell, no real preference

That should be reassuring. 

[MarcelNL]

5 minutes ago, Kal Rubinson said:

That should be reassuring. 

Just trying to understand your comment. Why would that be reassuring? If there are differences between the two they might 'imprint' like as if it was a 'sonic fingerprint', I expected nothing, was just listening.  

[dminches]

1 minute ago, MarcelNL said:

Just trying to understand your comment. Why would that be reassuring? If there are differences between the two they might 'imprint' like as if it was a 'sonic fingerprint', I expected nothing, was just listening.  

your post seems to be comparing the same thing: “next round; RAM buffer fed from SSD vs RAM buffer fed from SSD”

 

[MarcelNL]

6 minutes ago, dminches said:

your post seems to be comparing the same thing: “next round; RAM buffer fed from SSD vs RAM buffer fed from SSD”

 

ROFL...I SEE....will edit my post, the understanding reader knew that I intended to say RAM fed by Optane vs RAM fed by Optane ;-)

 

[MarcelNL]

4 hours ago, ASRMichael said:

Apacer ECC ram is here

 

model - D31.23185S.001

 

https://www.mouser.co.uk/ProductDetail/apacer/d3123185s001/?qs=Cb2nCFKsA8prqNusSnCNZw==&countrycode=GB&currencycode=GBP

on order :-)

[MarcelNL]

holy smoke, some tinkering in BIOS tells my my route may have to be changes....from a one size fits all headless music computer to indeed the two box solution....Just switched off all USB in BIOS (is switched off during boot, I am controlling the PC via Daphile anyway), a veil is lifted.

 

disabled some CPU power saving settings, manually controlled the CPU fan to start at higher CPU temps etc, now digging into the more obscure BIOS settings....is there a list with common settings and audio related preferences somewhere?  (mean stuff like XHCI and EHCI handoff, suspect many are the same across CPU platforms but for sure there are some AMD specifics in this BIOS)

 

So now up to disabling SATA, for a quick listen...  

[MarcelNL]

18 hours ago, StreamFidelity said:

 

Thanks for link. 

I don't know if it means anything. The RAM latency is rather average: CL19 / 2666 MHz = Calculated nanoseconds 14.25.

I use CL10 / 2400 MHz = Calculated nanoseconds 8.33.

can you post a link perhaps? I'm game for a comparison, a couple of sticks don't break the bank (pun intended)

[StreamFidelity]

4 minutes ago, MarcelNL said:

can you post a link perhaps?

 

Yes of course. But it is not an ECC RAM. I attach great importance to cooling, because I run a fanless PC and the lowest latencies.

 

DOMINATOR® PLATINUM 16GB (2 x 8GB) DDR4 DRAM 2400MHz C10 memory kit