Operating systems and their sound signatures

[sandyk]

Perhaps it's due to "software Bloat"  where with some OS's you need to do a lot more than just the basics ?

 Even Win 7 in Safe Mode sounded better, both when ripping and playing, with non essential start up programs not enabled, and is also the reasoning behind programs such as Fidelizer.

Peter St. has also had a great deal to say about the SQ of different Windows Operating Systems.

[sandyk]

20 minutes ago, ElviaCaprice said:

After modifying my mobo with the sCLK-EX, I find that any optimization has little to no effect.  I'm sure it would be the same for choice of OS.  Thus Windows 10 is fine.

 

That's a damn expensive way to try and correct a minor problem with  some set ups and OS's !

[sandyk]

56 minutes ago, Tomslin said:

A client OS sounds rarely as good as a server OS. And if you also run the server OS in Core mode, the difference is even greater. And in fact the same goes for hardware, serverboards, Xeon-CPU and ECC memory will also improve SQ further.

It was several years since I discovered these differences under all my tweaks with JPlay, AO and all possible computer-based hardware. Unlike before, now I dare say that mostly everything in some way makes a difference. As well as power supply last but not least.

 

And the 1s and 0s remain the same , despite the audible differences.

[sandyk]

Al

While LPS

3 hours ago, ALRAINBOW said:

All is lps powered I have tried Methods this is best to date.  

Reg lps while they can not be used and inprovmets are there. Using them every where gives a sound that is very close to vinyl having clean relaxed music with weight. 

Please comment on my observations 

Al

I can't comment on most of what you have done, because I haven't felt the need to do so when using ultra clean power as described later. 

 Clean  and isolated power to the various peripheral devices, including SSDs, HDDs ,USB, and Optical devices results in superior sounding Audio, whether exported via USB, or from Coax Out of a decent Soundcard,and makes it possible to hear differences between OSs that you may not have previously noticed. You will also be able to hear SQ differences between  music files saved on HDD and SSD for example, with SSDs using improved power sounding better.

It is also possible, as I have done, to use an  internal SMPS and then further improve the power to the other areas mentioned.

For my 2 internal SSDs I regulate the +12V supply down to 2 separate isolated +5V supplies. 

This prevents interaction between them, as well as preventing their sudden current draws getting into other sensitive areas of the PC. It also helps to ensure that you have only one earth reference to prevent earth loops.

This is more difficult to achieve when you use separate Linear PSUs.

Some members use SATA filters to do similar, but they aren't quite as effective as the other methods.

I tried that originally.

 Even my internal LG BR writer uses very clean +12V and +5V via an  internal  dual JLH PSU add-on which sits in the bay below it. ( A variety of Shunt Regulator)

It's also possible to get fantastic SQ from a USB memory stick using a Regen and a regulated battery derived PSU for it.

No USB lead though , just a modified USB-A to USB-B adaptor without +5Vand shield connected through to plug in the USB memory stick and Regen.

My apologies if this reply is a bit disjointed as I have yet to have a wakeup coffee !

 

Alex

[sandyk]

8 minutes ago, zackthedog said:

I recently built a Windows 10 all-in-one server/ripper with the SOtM Usb card and an HDPlex multi-rail LPS.

 

 The HDPlex 300W multi rail LPS greatly benefits from the use of a JLH PSU add-on after the rail supplying Regens, USB cards , SSDs etc.  It just isn't quiet enough by itself. 

I have recently made a dual JLH version for a Sydney friend using this HDPlex PSU.

[sandyk]

32 minutes ago, ElviaCaprice said:

If you added a sCLK-EX to that mobo and tXUSBexp PCIe card, you wouldn't need Server 2016.  Your SQ would benefit greatly above and beyond just an OS change.

 

 You would also be around $700 poorer ?

[sandyk]

1 hour ago, zackthedog said:

 

The external  200W supply isn't quiet enough?  I'm open to that.  It's better than a single SMPS powering the whole computer, for now.  Where do I find out more about these boards, and how would I implement one in my setup?  The SSD, USB card and motherboard are currently powered from the HDPlex.

 

Unfortunately, they are DIY only, and there are no more PCBs currently available.

I can  however provide a link to the original artwork if somebody wished to get a PCB maker to make a batch of at least 10  PCBs as well as provide full details including component types etc.

You would however need to be a fairly experienced DIY person.

It does however show the advantages of using lower noise PSUs than the HDPlex.

 

Yes, the HDPlex would still be preferable to most internal SMPS.

What voltage rail are you using in the HDPlex to power the SSD, is it the +5V rail or the Adjustable Output rail ? 

Alex

 

http://users.tpg.com.au/gerskine/greg/default.htm 

Click on Power Supplies and you will find full details including a copy of the original article in Electronics Today International by John Linsley Hood. 

[sandyk]

2 hours ago, ElviaCaprice said:

I suspect the clean power with the better clocking negates any negative SQ issues with software processing bloat.

 

 Much better power often negates much of the advantages of aftermarket Xtal Oscillators !

[sandyk]

32 minutes ago, zackthedog said:

Well, I'm a pretty experienced DIYer, though tiny solid state bits aren't my strong point.   Big honking tubes are more my line but I have a friend who's expert at the smaller stuff and is always game for a project.  I'll take a look at the link, thanks.

 

If you decide to go ahead with this, send me a PM and I will supply full details, including more recent improvements, and also how to convert the Current Limiter section of the PCB to take a LM317T set for +5V out into the Capacitance multiplier section.

This means it can convert from +12V to +5V for powering 2 SSDs separately.

The copy of the original article came from me, as well as all of the changes to the original published article which was only capable of use to 100mA  (as published) .The modified version  with the Current Limiter section bypassed can be used at currents of >2A at > +12V. This dual PCB can be configured for use with + and -  supply rails, as well as 2 +VE outputs.

 Incidentally, if your friend is able to photo etch a PCB, the artwork is suitable for this. It is a single layer PCB.

[sandyk]

9 minutes ago, esldude said:

As Dante works over ethernet, how would the OS effect the sound?

 

Yeah, we know, "bits are bits" and that's all that matters according to some.

It shouldn't matter with Coax SPDIF or USB either after making changes to PSUs ,RAM, Clocks, installing SATÂ Filters  etc. provided that there have been no changes made after the port that it leaves from, but a large number of members will now tell you that it does matter !

[sandyk]

10 hours ago, esldude said:

So if the OS in use effects sound quality upon playback, I would assume you'll think it effects sound quality during recording/mastering.  If that is the case, then it seems likely there could be interactions where playing back a Windows mastered digital file might respond to Windows playback differently than Mac playback.  So with all the trillions of possible permutations how likely is it that one OS simply has better sound than another?  And that is ignoring how OS activity varies tremendously depending upon the thousands of possible permutations of hardware it runs upon?  

 

Things that make you go.................................hmmmmmmmmmmmmmmmmmm!

Dennis

 In fact, I do believe that it could affect Recording and Mastering as well to a small degree.

 Given that Mac Minis sound better with a JS2 Linear PSU, I have often wondered if Barry Diament's recording could also improve slightly if his setup used a Linear PSU instead. I understand that his clocking has recently been improved by an order of magnitude , so I wonder if this will further improve his next recording session.

You may remember too, that some years back I reported about Silverlight from NYC and myself ripping a track from a Hybrid Norah Jones SACD,(Come Away With Me)  and him reporting that my rip sounded better than his ripped using a Mac despite our .md5 checksums being identical.

 It could very well be that the current Mac OS now outperforms W10 for audio performance though.

Alex

 

 P.S.

 If the SQ of Barry's recordings further improve after the clocking upgrade, do you REALLY believe that the 1s and 0s of the final result will be any different than they would have been without the upgrade ?

[sandyk]

9 hours ago, esldude said:

But if bits ain't bits because filtering isn't perfect, how can filtering (which is by its nature imperfect) make bits perfectly bits??

 

MQA-Many Questions Abound

 

  That may be a good question to ask Peter St. given that he manipulates bits using his XXHE software and various "engines",  and his software, as well as his DAC is highly regarded.

[sandyk]

1 hour ago, One and a half said:

What does work though is a Xeon platform, ECC RAM, consumer OS and files on large HDD PCIe fast pipe.

 

There was a report in Hi Fi Critic Vol.6 No. 2  that claimed the faster the better, including RAM too..

[sandyk]

17 minutes ago, bodiebill said:

This has the beauty (and I think Peter would agree in that department) that I can detach LAN, and also detach the PC's SSD (by switching off the external power supply), and even detach the stick with the wtfplay OS as this is loaded in RAM. And with longer tracks I even detach the usb stick or -drive that contains the music files and the HDMI cable, i.e. monitor. Can it be more minimal? And it sounds great as well (with only very slightly different sound signature compared to Windows which is hard to describe).

 

 

 However, the SQ of what you are hearing is governed mainly by what was done BEFORE loading the music into RAM.

[sandyk]

1 minute ago, WuNgUn said:

Are you referring to upsampling and filtering, etc...?

 

 I am referring to everything, including the OS used before the results are played from RAM.

There will be a small improvement resulting from playing after doing all this, but I have found that the SQ is governed mainly by what was done before it was saved to RAM for playback.

[sandyk]

10 minutes ago, plissken said:

Did audio setups across AU suddenly improve when the Tesla battery bank switched over in 140ms when the generating station went down?

 

 What a stupid thing to say. The Tesla set up only took over the supply of part of the South Australian end of the grid from the Eastern States. As for the quality of the power generated under those conditions, who knows ?

 The Tesla power may even have more higher order harmonics than the normal supply due to the Inverters needed to step up the voltages ? Someone else with expertise in  that area would need to answer that question.

[sandyk]

20 minutes ago, WuNgUn said:

Oh shit...we're gonna talk about $4000 USB cables now, and how they're sonically better than a standard USB cable....

 

 I'm not. But others who have heard the differences between USB cables with better isolation between D- and D+, and +5V VBUS and 0 volts may.

 As far as I am concerned the best sounding USB cable is NO USB  cable, just a modified USB-A to USB -B converter.

[sandyk]

12 minutes ago, WuNgUn said:

Where ever there is digital transfer of data, there is always going to be data error, whether be over WAN or LAN or SSD to RAM or RAM to cache...but there are also algorithmic error correction constantly going on as well, as far as I believe...

You wouldn't be able to download TB worth of data without error correction and have a 100% perfect replication, correct?

 We aren't talking about errors here. We are talking about resulting files with identical checksums.

However, I am not going to keep repeating everything that I have written on this subject for every new Freshman member who comes along.

You could try using the Search Facility for example.

[sandyk]

28 minutes ago, bodiebill said:

 

OK, I see what you mean, but I do not fully agree. Detaching the LAN, usb's and HDMI after playback has started, in my wtfplay setup, makes a very noticeable improvement SQ-wise.

 

 I don't doubt that it does to the person who made the changes.

However, you would be looking for differences that a normal listener may either not notice,  or even be aware of unless pointed out to them.

[sandyk]

1 minute ago, Ralf11 said:

 

 

Best xmas gift ever!

 

WuNgUn, you can just ignore this claim as it is just a silly game played by a prime candidate for a Darwin Award ;]

 

Your analysis is correct - there is no technical basis for the claim - put your $$ and effort into something worthwhile (spekers, room treatments, better source material masterings, galvanic isolation, etc.)

 

He doesn't need to. All he needs to do is peruse many of the threads in the Uptone area of the forum, for example, where he will find a wealth of information written by a well respected E.E. and designer, on the various possible areas of SQ improvement including Linear PSUs. USB Regens. DC  cables etc. etc.

In other areas of the forum he will find numerous threads and reports about improvements when using better USB cables, including the very recent  Lush USB cable from the designer of XXHE player software for Windows, as well as his highly acclaimed Phasure NOS DAC designs.

 

If this poster is GENUINELY interested, as distinct from Ralf11 who had already made his mind up , he can  PM me for further information.

[sandyk]

12 minutes ago, plissken said:

 

Prove it. You pm'd me a link for two files that sounded different, but they also had different hashes. Go figure.

 

 

 The attached demonstrates exactly why I do not trust you.

You deliberately attempt to twist what others say to your advantage.

Of course they had different checksums, as one file was processed using SeeDeClip Duo Pro to reduce clipping and compression.

 

"On ‎14‎/‎10‎/‎2017 at 2:41 PM, sandyk said:

Hi

 You should be able to hear differences between the original and the SeeDeclipped version of Better Be Home Soon.

A couple of other members are able to. These are the same 2 uploaded files discussed in the thread.

No, the differences aren't night and day.

They have been sent as Uncompressed Zips.

 

Regards

Alex"

[sandyk]

11 minutes ago, plissken said:

 

Ok, Can you resend me some files please.

 

Read my previous reply and you will see why I will decline your request.

 

"... The attached demonstrates exactly why I do not trust you.

You deliberately attempt to twist what others say to your advantage.

Of course they had different checksums, as one file was processed using SeeDeClip Duo Pro to reduce clipping and compression"

[sandyk]

1 hour ago, Ralf11 said:

 

c'mon - man up

 

send him the files like you said

 

or are you scared of something??

 

 I sent him the files as described, and he deliberately misrepresented them in this thread.

They were as discussed in the thread of the time,. which was about trying to improve over compressed music files. Yes, he did correctly identify the differences between them, but even you should have been able to hear the differences between the original file and the Declipped version.

 

I have no intention of further involving Plissken in any comparison tests.

 

 

[sandyk]

7 hours ago, WuNgUn said:

I'm not trying to troll you, nor am I going on a goose chase searching for such "claims" on an open forum...

If you don't want to make your point here and now, that's fine too

 If you bothered to search the forum you will find numerous posts by me about the positive results of 6 separate DBT sessions performed by E.E. and well respected technical writer Martin Colloms of Hi Fi Critic Forum ,on comparison .wav files that I sent him. The results were also published in Hi Fi Critic vol.6 No.1.

 Famous Recording and Mastering Engineer Barry Diament has also confirmed my reports by way of a comparison CD that I sent him.

[sandyk]

7 hours ago, PeterSt said:

 

Merely peaky current draw. An SSD is the worst at that.

 

 That can be catered for as I have done, by regulating the internal +12V down to +5V, followed by a "shunt"  regulator which in my case has a simulated capacitance of around 2 FARADS. It also has 2 separate , isolated outputs for powering the OS and a Music SSD.