HOW I´VE BUILT MY NEW PC FOR XXHIGHEND AND WHY I´VE CHOSEN THE COMPONENTS

[boleary]

Thanks Juan, now that is some sage advice. I'm afraid I'm a lot less technically experienced than most in this forum. I think I've settled on beginning this project with fan cooling, but having a case that can evolve into water cooling...on the other hand I can be a risk taker who jumps right into difficult waters (results vary, somtimes foolish sometimes brave!). However, given that my funds are a bit limited, your words really hit the mark.

Praphan, I very much appreciate your input too. As I'm all about function and don't care much about form, the open air test bench may be the alternative, if I do decide to run over the cliff.

Will be working through the weekiend in Nashville, so I won't have much time to put into the project, though I am excited about checking out the live music scene!

[boleary]

Oh the voices in my head....thanks to all who have posted in this thread.

I went ahead and water cooled (Zalman Reserator 1 v2)and used a cheap (Boy....do I mean cheap, $39!) case with a Corsair AX650.

MB is an ASUS P9-X79WS witn 16 gigs of 1600 DDR3--3930kcpu.

Let me just say that my old Dell Core 2 Duo, 32 bit,  did a fine job, but  I now have MORE of every range. The bass, in particular, has gone to another league entirely.

However the best....I now have on the fly volume control, if nothing else that was my primary goal!

Again, thanks to all.

[boleary]

Generally faster was better.

I am not so sure Stefano said that ?
And actually I'd think it is the other way around (theory only !)

Peter

Edit : FWIW I have mine running at 833Mhz (can do 2100).

Last night I remembered the above exchange from earlier in this thread. I went into the ASUS bios and, in the Ai Tweaker (overclocking) section I played around with the memory frequency settings. I've always used the EZ Automatic overclocking function of the bios and have wondered why it automatically set the memory frequency to 1373 MHz when I have GSkill-1600, CL9-9-9-24. Anyway I set the Memory frequency to 1648, 1373, and 1098 and gave each a listen. There was a very significant change in sound for the better the lower the frequency. There is a 824 Mhz setting I've yet to try, but it seemed that the sound became so delicate and smooth at 1098 that it I really need to listen "here" for a while before changing this setting again.

Just an FYI regarding things I don't understand but which my button pushing nature and ears regard as worthwhile!

[Nick]

Brian hi,

I will take a look at the memory speed as well, in the past I have always gone for faster settings but this sounds very promising.

Post coming up soon on the Adnaco fibre interface, there is more evidanve now of its effect, Paul has been borrowing my setup to try out whilst I sort out my amp. I think it has made quite a posertive impression.

Paul,

Please post your thoughts on the Adnaco if you would like to.

All the best,

Nick.

[listening]

Hi Brian,

maybe buildung a Rolls-Royce and driving it with 50 miles/hour could be a solution. Additionally I got better results by reducing the processor speed and switching the AMD from unganged to ganged mode for memory transfers. Ganged meanns that every core is using the 128 bit memory bus exclusively - the other cores must wait.

Georg

[Scroobius]

Hi Nick - I have a couple more things to try and will then post my experiences in a new thread I suspect it will be a biggish thread.

Best

P

[juanpmar]

Most of us are using now the USB 3.0 as the interface to connect the music pc to a) external hdds, b) to connect the music pc to the NOS1.
The advantage to improve the USB 3.0 speed for transferring data is obvious but I wonder if the speed of the USB 3.0 has any influence in the sound quality when transferring data to the NOS1.

I´ve found this interesting review: Faster USB 3.0 Performance: Examining UASP And Turbo Mode http://www.tomshardware.com/reviews/usb-3-uas-turbo,3215.html

Turbo mode can be found in the bios (at least in the Asus boards) and can help to burst the USB 3.0 performance. http://event.asus.com/mb/2010/The_Best_USB3_Experience/The_UASP_For_USB3.0.htm

Manually enabling Turbo mode on a board from a vendor that doesn't offer its own utility is possible by modifying the HKEY_LOCAL_MACHINES\SYSTEM\CurrentControlSet\Control\usbstor\054C00C1 registry key. Add a 32-bit DWORD value named "MaximumTransferLength" and specify the maximum transfer size value between 65535 (64 KB) and 2097120 (2 MB) in the data field.

I enabled Turbo mode in both ways, in the bios and in in the registry and it seems to me that now I have a better sound.
Maybe others could test it and let us know about it.

Best regards,
Juan

[boleary]

Hi Brian,

maybe buildung a Rolls-Royce and driving it with 50 miles/hour could be a solution. Additionally I got better results by reducing the processor speed and switching the AMD from unganged to ganged mode for memory transfers. Ganged meanns that every core is using the 128 bit memory bus exclusively - the other cores must wait.

Georg

Hi Georg, Per Uncle Google, it looks like "ganged" is an AMD only tweak. Don't know if there is an Intell equivlent. I guess one buys a Rolls Royce to be as insulated from road vibrations as possible. Of course never exceeding 50 mph keeps those vibrations to a minimum. 

Brian

[listening]

Hi Brian,

tried the memory tweak too, but setting the lowest speed possible. Very smooth and relaxed sound. Fits optimal to Oscar Peterson / Fred Astaire Complete Norman Granz Sessions 

Best regards,
Georg

[boleary]

Glad to hear it worked for you too, Georg. I tried the lowest setting 848 MHz and the sound was a bit too relaxed or too "woody" sounding so I went back to the 1048 (I think that's it) setting. Gotta be on the look out for those little statments from Peter; like gold nuggets, they are. 

[PeterSt]

Gotta be on the look out for those little statments from Peter; like gold nuggets, they are. 

Maybe I am not sure myself why I'm sometimes thinking what I think. But okay, it works out a lot.

But by now I really wonder what's they actual key factor when we talk about "just PC's". I mean, I now have that hexa core running myself in optima forma, but how to relate the sound to what "happened" ?
Sometimes (if not always) I think it is too difficult to proceed on things. This may be one of them. Look :

What this PC shows as the very first is a snap unheard. Bass has a resolution - the same. You even hear the stings vibrate of a cello (this was my aim ever back : make the somewhat higher frequencies "vibe" as well, and the instrument to do that was a cello in my view). But why ?
IIRC the memory is at 1333 which was the default. CPU runs at 4GHz (which is not the default, that's 3.2GHz). Ok, fine.

An important variable in the equation is that super-duper 0.9z-7 feature (yea yea, it will come to you, I'm sure ), and on my normal audio PC day before yesterday I shut it off to compare again. Unlistenable. Remember, this is what you listen to today; all is relative and for me too it was the best. Today ? unlistenable. Sheer distortion. And I am the most serious. So ...

tried the memory tweak too, but setting the lowest speed possible. Very smooth and relaxed sound.

Thus, smoothened things a bit and now the distortion is less audible ? should be something like that IMO. But when so, it should mean that the higher setting is the better, if only first that other distortion is out of the way. Thus imagine that you tweaked it the exact other way around. Instead of towards moody, it's going towards metal. And *then* you are not disturbed by anything. Something like that I perceive from that hexa core PC.

If it is correct that the higher settings (yea what ? higher current draw or higher freuency current draw ! -> smoothening !) create the better resolution, I should be able to proove that. Thus, it looks like I can have it because of less inherent distortion, and then there should be more snap without it being disturbing.

By now I am close to being as far as the "snap level" being a measure by itself. Thus, the less snap the worse it should be, and the other way around. However, it is not allowed to have yourself disturbed by harshness jumping in at the more snap levels. When the most snap is there, it is time to work on the harshness. Of course this is not to be done by eliminating the harshness, but by eliminating the distortion which comes from elsewere. And I know, this is undoable ... it would come down to getting another ampo etc., right ?

No, not right. Not right because you will be having 0.9z-7.

By the time you have that, you will start to be a theoretical nerd, like me. So, what I noticed with my normal audio PC is that this most pure sound is accompanied with less snap. Think like the stick hitting the cymbal being less audible, while the cymbal itself is okay. But it can't be, because pure = pure and it's main cause is higher resolution (I'm sure). This is up to the attack of an electric guitar being not there at flagiolettes, while another situation shows you the attack is there for sure. And I said it more often, flagiolettes are (somehow) the most difficult to bring forward (possibly they contain all the harmonics in the world - I didn't look it up.

So, we have that other PC, and suddenly all comes together. A bit of the USB3 solution, which *also* was just necessary to let my super-tweak work out.

Guys, when we operate at this really high level of squeezing out the best, and we see how much difference these non-sense appliances make - especially when we don't know the real theory behind it all (which actualy shouldn't even exist), we must not forget to re-judge, for example !, Vista and the lot. Not that I am telling you that we really should do that, but what I do say is that no element operates on its own. And again, if I now listen to that "distorted" pre-0.9z-7 sound, how could I even have been judging something like Vista. Or an SSD. Or RAMdisc. Whatever. Everything.

It is going to get too compleeeexx ...
(but in the mean time I am trying to see the difference between my speakers and the real woman singing. Okay, right. I see the difference. But I can't hear it anymore ... )

[Scroobius]

I was sitting in my favourite armchair last night sending myself nuts (OK so I am already) listening to fibre or perfluorinated polymer plastic (POF) versus Copper (Cu) USB interlinks (chemical names for Peter ha ha).

Once I settled on my preferred link (more later) I sat back and began to think surely now, surely this time the SQ is as good as it gets. So what I was listening to is muddled and distorted compared with what is coming in 0.9z-7  -  Peter you are such a tease!!!

P

[boleary]

Might SFS be one of those areas we revisit? To me the change in sound caused by changing the memory speed was very, very similar to changing SFS, but in the opposite direction: increasing SFS smoothes the sound and reduces harshness the same way that reducing memory speed does.

Just a thought....

[manisandher]

An important variable in the equation is that super-duper 0.9z-7 feature (yea yea, it will come to you, I'm sure ), and on my normal audio PC day before yesterday I shut it off to compare again. Unlistenable. Remember, this is what you listen to today; all is relative and for me too it was the best. Today ? unlistenable. Sheer distortion. And I am the most serious.

Hi Guys, greetings from Rio (I'm in Brazil for a few days business)! Far be it for me to preach to anyone, but as soon as Peter mentioned his Noble-winning feature, I pretty much stopped tweaking all together, with a view that it was pointless. But I have a few weeks off in July and hope to be able to join the fun then.

Thanks to everyone who's shared their experiences to date - there will be so much to play around with... can't wait...

Mani.

[juanpmar]

Hi Guys, greetings from Rio
Mani.

Tall and tan and young and lonely
The girl from ipanema goes walking
And when she passes
Each one she passes goes haaa...
 

Enjoy it, Mani!!

Best regards,
Juan