Music Server Project

PNWAS

New Member
Aug 4, 2010
16
0
0
#1
For those of you participating in our music server project. This post will be used to communicate back and forth about the project.
 

fork

New Member
Aug 3, 2010
188
0
0
Bellevue, WA
#2
I think we should have an initial meeting to help everyone who needs to help assembling and getting their system up and running and a second meeting where we can definitely use Gary and Joe, to setup the OS and the finishing touches. Those of you have already got your server running might want to post a list of hardware, so readers have an idea what kind of components will work well together.
 

fork

New Member
Aug 3, 2010
188
0
0
Bellevue, WA
#3
My PC Specs

Here are the specs on my server:

OS: Windows 7
Motherboard: Super Micro C2SEA
Power Supply: Silverstone Nightjar Fanless
Memory: 4GB DDR3 1066
Display: 58" Samsung by HDMI
Case: GMC Toast
Heatsink: Scythe Ninja
Processor: Intel 2.8, 65W, Dual Core
Thermal Grease: MX-3
Hard Drive: Samsung 1TB F3
Disc Drive: Pioneer 12x Blu-ray burner
Player Software: Foobar
USB Cable: Entreq Konstantain
DAC: Wyred4Sound DAC 2 (24/192)


Misc Details:

* The case interior is sprayed with EMI/RFI coating.

* The system is fanless, except for one fan on the CPU heatsink (because of the Blu-ray processing demand).

* The OS is tweaked from the Black Viper website.

* The PC serves as a music and home theater server. Blu-ray discs can be played from the drive, using the PowerDVD 9 player, but I also have Slysoft software, which can transfer the Blu-ray to the hard drive or a blank disc. 40 Blu-rays takes up 1TB of drive space, so it's not quite practical to store movies on the PC; 3TB drives will be out soon and a RAID server with 2TB drives could store hundreds of discs (with some hardware cost).

* I recommend using FLAC or ALAC, instead of WAV, because WAV does not store the tags in the file, which is a problem with some software. Many people have performed serious blind testing and found no audible difference between the compressed and uncompressed formats. If you can afford a solid state drive, you'll want the extra drive space.

* The memory has an input voltage of 1.5V. Check the specs for your motherboard to make sure it will take the memory you choose; some boards don't take all voltages.
 

fork

New Member
Aug 3, 2010
188
0
0
Bellevue, WA
#4
Don't buy an SSD right now. Information recently leaked about the new 25nm SSD technology. These will come out at the end of 2010. The capacity will be double, up to 600gb.

http://www.legitreviews.com/news/8792/
 
Apr 3, 2010
16,022
0
0
Seattle, WA
#5
There has been formal presentations by Intel/Micron at ISSC on their 25 nm flash development for a while. So the fact that they are coming to market next year is not a big surprise :).
 

fork

New Member
Aug 3, 2010
188
0
0
Bellevue, WA
#6
Yes, I realize the 25nm is not a concept that was revealed last week, but practically, for someone who wants to move to a SSD, knowing the approximate release date of the newer technology takes a lot of the uncertainty out of the decision of whether to wait, so in that respect, the news is important and useful.

Actually, by 2011 can more accurately be stated as during Q4 2010.
 

FrantzM

Member Sponsor & WBF Founding Member
Apr 20, 2010
6,469
0
0
#7
Hi

In stereotypical audiophile .. I also use an SSD, albeit a small one. Most of my Music is in a NAAS anyway, I am interested in a scheme where I would use the SSD or a cheap USB thumbdrive to buffer whatever I am playing, perhaps a script ... Has anyone fond such .. I feel lazy and wouldn't want to start from scratch ...:)

Additionally, what are the real advantages of an SSD if I am using Windows 7 with enough RAM (8) .. I am willing and able to strip Windows from almost everything save networking so .. Why would one want to use those (expensive) SSDs besides the absence of noise?
 
Apr 3, 2010
16,022
0
0
Seattle, WA
#8
The difference is hard to demonstrate. But in principal, an SSD puts a much quieter load on the PC power supply than a mechanical disk drive with its motors and servo control. Spikes on power supply can cause increased jitter on digital outputs of PC. But again, without instruments before and after, this is a hard point to prove.

What I like about SSD is the fast boot up time. So I am able to restart my PC almost as fast as CE device. Likewise, windows updates go extremely fast. More RAM does not solve either.

The application for larger SSD is to put the entire library on it so that you don't have to mess with caching. My library is over 200 gigabytes right now. So larger SSDs would be nice from that point of view. Once there, you can also turn off networking reducing any impact that might have on performance.
 

garylkoh

WBF Technical Expert (Speakers & Audio Equipment)
#10
I think that amirm is spot on the button with his comment that SSD's put a much quieter load on the power supply than normal spinning hard discs. With the servo control of the seek arm, the power could be the reason why we all think a SSD sounds better. I modified my home PC using external power supplies for all the drives, and it did make it sound far better. Realizing this, there is no need for large SSDs if the actual music store is external, on a disk array that has an external power supply.

For those of you who want your music server "diagnosed", bring them early to the next meeting (Sept 9). I will be there early (before 6pm) and we can have a mini-session. I'm also in the process of building another server RMAF, if anyone's interested, I'll share my component list, and we could build a few together. It's not done yet - some components didn't work well together, but it will be sub-$500 and run Windows 7.
 

FrantzM

Member Sponsor & WBF Founding Member
Apr 20, 2010
6,469
0
0
#12
I think that amirm is spot on the button with his comment that SSD's put a much quieter load on the power supply than normal spinning hard discs. With the servo control of the seek arm, the power could be the reason why we all think a SSD sounds better. I modified my home PC using external power supplies for all the drives, and it did make it sound far better. Realizing this, there is no need for large SSDs if the actual music store is external, on a disk array that has an external power supply.

For those of you who want your music server "diagnosed", bring them early to the next meeting (Sept 9). I will be there early (before 6pm) and we can have a mini-session. I'm also in the process of building another server RMAF, if anyone's interested, I'll share my component list, and we could build a few together. It's not done yet - some components didn't work well together, but it will be sub-$500 and run Windows 7.
Inquiring mind wants to know .. Did you write sub $500? Also interested in what you did to the power supplies
 
#13
Interpretation

Recently Home Theater and High Fidelity tested the jitter of a blue-ray player with and without the optical disk spinning.
As you can see it is measurable on the SPDIF out.
Sound plausible to me that a HD has the same impact or probably more.
Interpreting the result of an experiment is never a slam dunk. Borrowing a result done by other people for a different purpose than yours is even less of a slam dunk.

If 'Player in stop mode' means that nothing was playing, then there are more differences between the two traces than just a motor running. For example, the bit pattern being output in "Player stop mode" might have been silence. Big chunks of the DVD player hardware might have been idle in one case and not in the other case.

From the review article that you linked to

"Jitter was very low both in the stopped mode and when playing an 11 kHz test signal. When playing, jitter was between 7 and 10 picoseconds, which is the data-dependent jitter. The peaks are called periodic jitter and are there whether a disc is playing or not."

Is the difference between a ~5ps jitter across the frequence spectrum and a ~10ps level audible?

The periodic jitter peaks seem higher and might produce more audible effects. It is hard to see how those peaks differ for "player in stop mode" versus playing the test signal.

Results like this are great for generating a hypothesis. However, they aren't necessarily sufficient to allow a difference to be ascribed to a specific cause or to decide whether it is significant.

Bill
 

garylkoh

WBF Technical Expert (Speakers & Audio Equipment)
#14
Inquiring mind wants to know .. Did you write sub $500? Also interested in what you did to the power supplies

Yes, I intend to keep it under $500 by leaving out things that are not necessary to good sound. As PNWAS members know, I have a bit of experience in building these things. I demo'ed one during CES2005 when we were at Alexis Park. I was universally laughed at, and even when I showed that the server sounded better than the kilo-buck CD transport + DAC we were using then, no one wanted to believe their ears.

The server I'm currently using has evolved for the past 5 years, and it's time to put it out to pasture. Hence, building a new one to bring to shows. I built one for the home with expensive bits - big SSDs, quad-Core processor, etc. but it didn't sound any better than my 5-year old. So, the new one won't have lots of processing power. About the only processing needed is for foo_hdcd. And that's not much of a load.

It won't play BluRay, and it will have a small SSD. The main music store will be off a RAID drive (not included in the sub-$500 BOM). The music store will be separately powered. It may be on NAS or on USB - haven't decided yet. I'll have to determine which sounds better.

It will be headless, keyboardless, mouseless. I will use remote desktop, and an iPad to control Foobar.

I don't know yet what I will do with the power supply. That will come later when I tweak it. Likely I will build a linear power supply for it.

Preliminary bill of materials (including the reasons for picking these parts):
Intel DG43GT motherboard (for on-board Firewire, HDMI and S/PDIF optical out) - $85
Intel Pentium E6300 Wolfdale (Enhanced SpeedStep allows slow-down to reduce heat) - $80
OCZ Gold 4GB DDR2 800 (high performance gaming memory with heat sink) - $80
Kingston SSDNow 64GB SATA II SSD (MLC, cheapest SSD with Win7 TRIM support) - $125
CoolerMaster Elite 100 case (tiny, tiny case as my server travels a lot) - $60
Sony AD-7700S slim CD-reader/DVD-burner (this will be for ripping CDs) - $40
Silverstone NT07-775 90mm CPU Cooler (runs really quietly, I've owned a number) - $20
Total: $490

Just as a by-the-by, I picked a mother board with an optical S/PDIF out because most of my source are redbook CDs. Using that output, through a de-jittering buffer like the Monarchy DIP or Genesis Digital Lens, and then AES/EBU to a good DAC sounds fabulous.

What is even more interesting is that the optical S/PDIF cable electrically isolates the ground of the server (noisy) from your hifi system. And while the Toslink standard specifies plastic fiber, I found that using a glass cable (I use the Van den Hul and the Sonic Wave) sounds far better.

Unfortunately, a lot of on-board chipsets are 48k only. So, Windows upsamples 44.1 to 48. I don't know about this one - will have to finish building it to find out. The last motherboard I bought couldn't fit the case, so I'm waiting for the new board to arrive.
 
Last edited:

Vincent Kars

WBF Technical Expert: Computer Audio
Jul 1, 2010
860
0
0
#15
Results like this are great for generating a hypothesis. However, they aren't necessarily sufficient to allow a difference to be ascribed to a specific cause or to decide whether it is significant.

Bill
Thanks for your replay Bill.
As usual hard to disagree with your thoughtful analysis.

I looked at the matter from a different perspective.
Up to now I have only heard claims (You are familiar with AudioAsylum) about the great improvements when…..
Well almost everything that in someway or another reduces electrical activity inside the PC. This is the first time I see some measurements indicating that this relation indeed exist, is measurable.
As we audiophiles have very revealing gear and such an acute hearing that we can even hear inaudible differences, I was glad to find some evidence that at least there is a measurable effect.

The graph is a funny one.
Anyone arguing that the jitter level DOUBLES has a point.
Anyone saying it remains well below a very severe threshold (20 ps at 20 kHz, Julian Dunn) has a point.
 
#16
Thanks for your replay Bill.
As usual hard to disagree with your thoughtful analysis.
Thanks.

I looked at the matter from a different perspective.
...
This is the first time I see some measurements indicating that this relation indeed exist, is measurable.
As we audiophiles have very revealing gear and such an acute hearing that we can even hear inaudible differences, I was glad to find some evidence that at least there is a measurable effect.
I think you were right to spot the useful graphs of jitter levels in the DVD player and think what they might tell you.

I don't know for sure what this graph is showing. I'd guess that it illustrates the result of a calculation about the maximum level of jitter that allows for 16 bit or 24 bit accuracy in the D/A process. If so, it doesn't represent the outcome of actual experiments on audibility.

As I recall, there were assumptions made about how the D/A process works that may not apply equally well to the range of DAC chips used in modern D/A implementations.

There is a good article on jitter in one of the Audio Critic back issues. Of course, like the Julian Dunn articles, this article is pretty old so D/A implementations have moved along.

Bill
 

fork

New Member
Aug 3, 2010
188
0
0
Bellevue, WA
#17
Music Server Meeting this Thursday 9-9-2010

Everyone who will be attending would benefit from reviewing Gary's server document on the front page of the Genesis website. Last I heard he had 20,000+ downloads. About half of the information is out of date, but it's still a good document to learn the ropes. Also, you will find some useful information on this site in the "Computer Based" forum.

I will be interested in hearing more opinions (from experience) on the SSD debate, I've tried comparing files played from a thumb drive and my Samsung F3 hard drive, without much improvement. My current system (First Watt F5, Modwright 36.5, ASI Liveline) is made for resolution, excepting the Zu Druid speakers which don't excel at treble extension. However, to tell you the truth I haven't gotten the biggest gains from trying to limit potential jitter causing activities on my server. For example, I added a switch to my motherboard pins that can turn the Ethernet circuit on and off, but it doesn't have much impact.

Joe Pittman from Kosmic offers Weiss DACs and with his DAC 202, which has a feature to confirm bit-perfect playback, he has determined that just bypassing the kmixer (even when using his SSD drives) is not enough to achieve bit-perfect. I'll have to find out if he has finally figured it out. As a rule I limit the electrical, hardware and software activity, as much as can be done without creating too much inconvenience.

In his server document, Gary mentions a noticeable improvement with turning off Plug and Play, but neither of the DACs I've owned would make any sound without PNP, including the Wyred4Sound, which has a proprietary driver. I can tell you, it's possible to overdo it with shutting down Windows Services, from the perspective of audio quality. Using Boot It software I created separate Windows 7 partitions, with one that had all but about 10 services disabled. Unfortunately, there was an unmistakable loss of quality in the minimized version. The tweaks suggested on the Black Viper website don't seem to have any negative impact on sound.
 
Last edited:

garylkoh

WBF Technical Expert (Speakers & Audio Equipment)
#18
Fork, let's not take over the meeting for a Server meeting. We have a guest presenter and we should accord him the respect for taking the time to come in. I will discuss servers before the meeting starts ONLY!

Regarding your comments above, I have not found that playing off a USB thumb drive with a USB dac to be an improvement. It might even be detrimental. When we did the A/B comparison demo during the meeting with an external SSD, it was plugged into an ExpressBus slot. Also, turning the Ethernet on and off does not make a difference. It is the WiFi that made the biggest difference.

Turning off Plug and Play is with WinXP. Vista and Win7 are totally different architecture, and I don't know if it works the same. With Joe's Win7 server, we are also using WASAPI and the Weiss ASIO. I haven't checked out the Black Viper website, but I will before the meeting.

See ya all tomorrow.
 

fork

New Member
Aug 3, 2010
188
0
0
Bellevue, WA
#19
Gary, I don't see your motherboard on Newegg. I believe you have two of the #s reversed and it should be DG43GT.

Also, not listed in Gary's parts list is, a copy of Windows 7 (don't accidentally buy the upgrade version), a keyboard, mouse and display (if you're not connecting to an iPhone or iPad) and MG Super Shield RFI/EMI coating (optional).
 
Last edited:

garylkoh

WBF Technical Expert (Speakers & Audio Equipment)
#20
Oops, you are right. Thanks Darrin.

Correction, folks - it is DG43GT.

As I mentioned in my post, I intend for it to be headless, keyboardless and mouseless. Just borrow them from another computer you have when you are doing the set up.

For Win7, you should be able to easily get the OEM version off the Internet for around $75. Not Win7 Home - you need the Pro version so that you have remote desktop.

Joe is organizing for a group buy and hoping to get some discount from NewEgg. He's got a commercial account and will pass through any savings to club members.

Cheers
Gary
 

About us

  • Founded in 2010 What's Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals we enjoy learning about new things and meeting new people and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing