Jump to content
IGNORED

NVMe SSD designed for audiophiles


Recommended Posts

Yes I got confused yesterday when I typed earlier response. I still have the Audiolinux on a USB / ramroot and all for my endpoint fanless minipc. Didnt like it at all. I can hear why others like the clarity and the dynamic SQ but to my ears it is not to my liking. Still prefer the Windows Server SQ. 

Trying out Euphony now. Strange that it sounds 'not good' = checking my settings and trying basic 44.1 khz CD Wave files to check. Hires 705.6 sounded "not good".

 

Link to comment
6 hours ago, kelvinwsy said:

Trying out Euphony now. Strange that it sounds 'not good' = checking my settings and trying basic 44.1 khz CD Wave files to check. Hires 705.6 sounded "not good".

 

I too have a Euphony license but I find Windows to be more musical and involving. Euphony sounds euphonic but muffled (low fidelity) and very hard to understand music lyrics. It seems to me that notes are truncated (decays are shorter). The music is so vidid and intonation and singer inflection is easily rendered with Window. I use the same hardware (swapping same capacity Optane drive to do the comparison.)

Link to comment

In response to your observations about Windows OS =

Win 10 Pro, Win LTSC, Windows Server 2012, 2016, 2019 then GUI and Core

With AO2 and AO3. Fidelizer 6.0 to 8.0

 

Let me summarize my impressions and opinions.

Windows 10 Pro is Noisy, cluttered. - Obviously because of too many processes running, and the electrical noise generated. - even with programs like AO and Fidelizer, you will need to have Project Lasso/ MinorityClean(? is that the name) to reduce and prioritize the processes.

 

Windows LTSC - a step in the right direction but still sounds bland - But it is a lot quieter and less cluttered sounding than 10 Pro - 10 Pro is unlistenable on a good transparent system

 

Let me bring in JRiver MC 23 -26 (I am on 26 now) = PLaying from RAM mode the entire song brings a great improvement to the sound. Listenable - all other Modes - Yuks! Only good for casual background music playback.

 

Audiolinux - good for End point. But it gave a high frequency tilt to the sound = Not using it any longer

 

Euphony - on trial mode now.

My OS for the Main Rig >>>

What I prefer are the Server OS's of Windows - 2012 was OK, 2016 was quite a bit improved.

But 2019 Win Server OS even in GUI mode Without Fidelizer and AO is quite pleasant to listen to. 

It is quiet, neutral in the bass/mids/highs - No nasties and like you quoted above, intonation and pronunication by vocalist are clearly heard.

 

It is when you switch to CORE with AO3 / Fidelizer 8.0 Pro that you get the processes down to 20 in the most extreme case- No internet, No remote network in, no WIndows Defender... almost everything is switched off. 

 

The issue is that it is SO transparent that you run into expose' of the weakness of the music playback chain.

On USB - the digital sound of your bad PC power, USB hub noise, Ground Voltage different levels, all are heard exactly for what they are. 

These have to be sorted out.

 

Using ethernet /switches/Clocks - I have gone 2 switches and Ether-regen in series with 2 clocks 50ohm/75 ohm and SOTM Ultra Neo - Cleans up the noise but the Vividness and Clean Vitality and Dynamism of the Sound is scrubbed and filtered by all those transformers in the chain of switches.

 

I have since switched back to USB direct connection to my DAC using many tricks including a double headed LH Labs USB cable which allows the injection of Uptone Regen's LPS1.2 power into the Isoregen separate from the data link. This already comes from a Clones Audio SHAAR USB PCIE card. LT3045 and Supercharged battery packs power everything of 5V and 12V. No SMPs at all in my system. Internet Broadband banished entirely from my music room. Music Server is cut off from all outside electrical/network interference

 

The components I use are not up to the JCAT Femto, Taiko Extreme, Paul Hynes level but I find that synergy and matching and extreme care in balancing the Power/Isolation/Filtering arrangements on my playback system is giving me a sound that I can enjoy music on.

 

Back to the FEMTO M.2 card testing topic

Just continued to test the card with further burn in. The SQ is stabilizing at a very high level. It is a very good neutral sounding drive ( Ha! NVME drives have SQ characteristics - who would believe - Well I have the lot, Intel Optane 64, Samsung 970, and this FEMTO M.2 - they sound different)

 

Tried to boot it from a M.2 USB adapter card - My Gigabyte Aorus Mobo Bios just would not recognize the drive. Only the Samsung 970 NVME on the adapter can be boot up on my Lenovo 6700k PC not the Gigabyte 10700k, So that test had to be abandoned.

 

Now I will try the Grounding of the USB port on the LPSU driving the Femto M.2 card tomorrow when another of those Grounding boxes arrives. I am using one at present but I do not wish to contaminate the ground with the other circuits I have grounded to this first box - rather use a new grounding box just for the M.2 drive. The M.2 drive is that transparent.

 

Will report back if improved Grounding can elevate the sound of this M.2 card further.

 

Happy Listening to All

Link to comment

 

2 hours ago, Lukasluis said:

I too have a Euphony license but I find Windows to be more musical and involving. Euphony sounds euphonic but muffled (low fidelity) and very hard to understand music lyrics. It seems to me that notes are truncated (decays are shorter). The music is so vidid and intonation and singer inflection is easily rendered with Window. I use the same hardware (swapping same capacity Optane drive to do the comparison.)

What OS/player are you using now? With 300gb on boot drive I can entertain revisiting  Server 20xx although the license cost after trial is pretty steep.

Regards,

Dave

 

Audio system

Link to comment
27 minutes ago, davide256 said:

 

What OS/player are you using now? With 300gb on boot drive I can entertain revisiting  Server 20xx although the license cost after trial is pretty steep.

Windows Server 2019 Core, AO 3.0, Fidelizer Pro 8.0, Minority Clean, Jplay Femto. Controller is Bubble Upnp. Local music is on Usb-attached HDD with external LPS. I also turn off Hyperthreading and Turbo when using Windows. Also cap the Frequency to 1.2Ghz.

Link to comment

I’m trying out Euphony on a 10i7 NUC powered by a 19V 8A LPS… got an internal Samsung EVO 970 plus M.2 SSD and was considering to order an Optane 32GB ( or the apparently better sounding 16GB version?) to replace the 970 plus but after reading this I wonder if I should hold off and try the 970 plus first… got about 20 days of trial left… any else can comment on the 970 plus sound quality?

Link to comment
1 hour ago, Smaragdhk said:

I’m trying out Euphony on a 10i7 NUC powered by a 19V 8A LPS… got an internal Samsung EVO 970 plus M.2 SSD and was considering to order an Optane 32GB ( or the apparently better sounding 16GB version?) to replace the 970 plus but after reading this I wonder if I should hold off and try the 970 plus first… got about 20 days of trial left… any else can comment on the 970 plus sound quality?

Mine has been relegated to WIN10 OS drive in desktop... Optane was better, fewer digital irritants. Speed is not necessarily your friend in an audio USB out machine,

stable and electrically quiet is. The FEMTO drive builds on that and is superior to running using Optane drive or OS loaded to RAM

Regards,

Dave

 

Audio system

Link to comment
  • 2 weeks later...
On 6/14/2021 at 8:09 PM, Exocer said:

Brief update:

First of all, thanks @ASRMichael for the tag! I totally would have missed this thread otherwise.

 

I am now providing external power to the Femto NVME and this has taken it to new heights. As I told @darkforce, all of my cables were built using Oyaide 2.5mm/5.5mm straight barrel DC outputs. With the location of this card in my machine, the barrel connector did not fit, nor was it the proper size. So, I went ahead and ordered a 90 degree Oyaide in 2.1/5.5m size and modified one of my cables to use it.

 

I thought about going with an adapter but my OCD didn't let me do so although that would have been a much easier approach.

 

Summary:

  • There is great improvement in transparency and realness to the sound. The improvement is shockingly large. There will be no going back to the regular Optane at this point.
  • The sound is more transparent and detailed without fatigue.
  • The presentation is even more effortless
  • Noise floor has dropped to some degree
  • Timing and PRAT are fantastic and a bit more "real" sounding
  • Greater sense of being there for the performance.
  • Decay of notes, especially in some Jazz piano recordings, is so much more realistic sounding than before
  • Larger and more precise soundstage

I'll just conclude this by saying the drive was an improvement to my system across the board. Without the external power, there were no notable negatives to the sound (although the sound was a bit subdued and polite at first), and obvious improvements to vocal detail. With external power it really is something special. This bodes well for the power supply and the drive itself I suppose.

 

Context:

The above feedback pertains to the Femto NVME drive vs my memory of the Optane. Many of the above benefits were obtained by simply switching to the Femto NVME without dedicated power, but, to a lesser degree.  The major benefits came after adding external power.

 

My listening spans across multiple genres but I really enjoy Jazz pianists. My playlists usually contain a lot of live Jazz performances. Other genres also displayed improvements I did not expect. For example, I stumbled upon Madonna's "Vogue" and was gobsmacked by the soundstage. I had not listened to this song since he earlier part of my computer audiophile journey and it sounded like an entirely different song this time around, in a good way.

 

Cheers,

-Rob

 

1461189709_NVMEpower.jpg.90195b5ac5ecf94f3fa80870b9204843.jpg

Wow! Great review! Did you compare to Ramroot? 

 

@darkforce when will these be available for sale? 

Link to comment
1 minute ago, ASRMichael said:

Wow! Great review! Did you compare to Ramroot? 

Did you mean buffering?

 

If so, yes. When buffering, the files are run from RAM as opposed to OS disk.

 

Caching to OS disk sounds better than Buffering for playback.

 

Either way, I have still decided to ramroot the OS to keep some separation between OS and music files.

 

I will be sticking with Ramroot enabled and caching of files before playback.

Link to comment
4 hours ago, Exocer said:

Did you mean buffering?

 

If so, yes. When buffering, the files are run from RAM as opposed to OS disk.

 

Caching to OS disk sounds better than Buffering for playback.

 

Either way, I have still decided to ramroot the OS to keep some separation between OS and music files.

 

I will be sticking with Ramroot enabled and caching of files before playback.

Running OS in RAM helps reduce I/O bottleneck if you do DSD512 upsampling, I was getting pauses with OS on disk

Regards,

Dave

 

Audio system

Link to comment
  • 2 weeks later...
On 5/9/2021 at 3:03 PM, badK said:

I figured that using a storage layer that is isolated by a physically separate PS (actually four of them), and fibre cables for data, I would be about as isolated as I could get.  Has anyone else used FC-based storage?  These are available on ebay for pennies, relatively speaking.  Prices this morning like $149 for a 24 drive array with FC-to-SATA controller.

 

Fibre channel has been dirt cheap for awhile. Nowadays the good NICs are "converged" ie. they can support FC or IB or Ethernet and then there's FCoE etc.

 

The latest greatest technology is NVME over Ethernet (NVMEoE) and that's where 100 Gbe starts making more broad sense. Modern NICs support these protocols on controller.

Custom room treatments for headphone users.

Link to comment
On 6/15/2021 at 9:05 PM, davide256 said:

Running OS in RAM helps reduce I/O bottleneck if you do DSD512 upsampling, I was getting pauses with OS on disk

Linux generally runs in RAM, at least the kernel does. Ive booted Ubuntu over the network eg PXE or iSCSI and have the network down (so no "disk access") ... it runs happily along until you try to run a new program etc. Given generous RAM, HQPE for example, only accesses disc to do logging, as all music data is otherwise mounted via SMB3.

 

Windows ... that's another story

Custom room treatments for headphone users.

Link to comment
  • 2 weeks later...

I run the entire windows server 2019 and music software in ram. The Firadisk driver is used to trick the system to use ram instead of a hard drive. There's only one sata disk installed and it only holds music. The system boots into ram from a usb stick that's removed after booting. Here's the original how-to on the Jplay forum RAMdisk with Windows Server 2016/2019 - Configuration Guide 

 

If your music is on an external usb or a nas you can disable the sata controller for even more benefit. The best disk drive is no disk drive ;-)

Link to comment

Is it feasible to hack a PCIe to NVME adaptor to accept external power? Barring upgrading elements on an NVME SSD that might be a doable upgrade, just thinking out loud..

ISP, glass to Fritz!box 5530, another Fritz!box 5530 for audio only in bridged mode on LPS, cat8.1, Zyxel switch on LPS, Finisar <1475BTL>Solarflare X2522-25G, external wifi AP, AMD 9 16 core, passive cooling ,Aorus Master x570, LPSU with Taiko ATX, 8Gb Apacer RAM, femto SSD on LPS, Pink Faun I2S ultra OCXO on akiko LPS, home grown RJ45 I2S cable, Metrum Adagio DAC3, RCA 70-A and Miyaima Zero for mono, G2 PL519 tube amps. 

Link to comment

basically, a PCIe riser card like this one should work, allowing external Power: 

https://www.amazon.nl/ADT-Link-Gen3-0-sleutel-Pci-Express-verlenging/dp/B07Z4RCDKZ

ISP, glass to Fritz!box 5530, another Fritz!box 5530 for audio only in bridged mode on LPS, cat8.1, Zyxel switch on LPS, Finisar <1475BTL>Solarflare X2522-25G, external wifi AP, AMD 9 16 core, passive cooling ,Aorus Master x570, LPSU with Taiko ATX, 8Gb Apacer RAM, femto SSD on LPS, Pink Faun I2S ultra OCXO on akiko LPS, home grown RJ45 I2S cable, Metrum Adagio DAC3, RCA 70-A and Miyaima Zero for mono, G2 PL519 tube amps. 

Link to comment

Moving the FEMTO NVMe to a PC I built in December (i5 9600K-Coffee Lake/MSI PRO Z390-A) as I've learned through comparison the i7 NUC was causing

distortion in choral music. Will be trying out  a PCIE x4 to M.2 adaptor  as I want to be able to switch  boot between WIn10 on the EVO 970 + and Euphony on the FEMTO drive,

While the MOBO has 2 M.2 slots, the second is for boards shorter than 2080. Would have preferred an extension cable or elbow adaptor option for the second M.2 slot but

what I could found looked sketchy for quality.

Regards,

Dave

 

Audio system

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...