How to build a computer transport for XX ?

[xp9433]

Last night I had the first opportunity to listen via XXHighEnd (engine #1 0.7j) and evaluate how it sounds:

1.   Compared to Foobar
2.   With computer’s Windows components and setting adjusted for low noise low power etc. operation.

I have been experimenting using my business laptop (not a very upmarket approach but it allows me to listen in bed) as preparation for investing in a computer based music server and transport. My very basic system:

Dell Inspiron 6400 notebook > Trends Audio USB (NOS) UB-10 DAC > AKG 701 headphones

Immediate Reaction:
Right away XXHighEnd was better than Foobar - as all of you already know. To a worthwhile degree, it was more clean, clear, detailed, transparent, and dynamic than Foobar. Thank you Peter!!

However I was also interested in what impact recommendations in “The art of building Computer Transports” would have (low noise, low heat, low power (Intel’s lower spec Core 2 Duo 1.6GHz E2140; dual 2.5” notebook SATA drives, etc.. with all unnecessary Windows and operating functions disabled), covered best here:

http://www.audioasylum.com/forums/pcaudio/messages/1/19242.html

As this directly affects my proposed computer configuration and purchasing requirements, especially considering XXHighEnd recommends Vista whereas XP is recommended for the other. I also wanted to use the Computer for DVD/TV etc., which is not part of the “cics” philosophy.     So I followed many of the recommendations in his article:

After Changing Computer Component Settings:

The configuration changes changed the sound for both players for the better. XXHighEnd was marginally improved – very slightly more of everything described above. The sound through Foobar improved proportionately much more, with the final result that Foobar with the reconfigured computer was now a lot closer to the sound achieved by XXHighEnd. XXHighEnd was still better but by a smaller margin. However, I don’t know what changes to the computer settings resulted in the small improvements from XXHighEnd.

I realise on a better system the differences could be more pronounced, so I am looking for input from members on this forum.

Has anybody here experimented with building a computer transport with special components (and setup) that make a difference to sound quality, and can make recommendations on what sort of computer might get the best out of XXHighEnd?

Thanks
Frank
New Zealand

[xp9433]

The art of building Computer Transports creating CD transports with a view to dramatically reduce its cost whilst improve performance.
Windows XP SP 2 is recommended and efforts are made to optimize hardware, power supply, PC components, and important BIOS settings to drastically reduce or bypass electrical noise, minimize non-audio motherboard traffic & overheads, and reduce mechanical vibrations.
Optimise Power Supply, PC Hardware, Soundcard. (Every item recommended to achieve low power usage, noise, & vibration)

Case:        Something like Zalman's TNN-300 for good heat noise performance.
PSU:        Enermax Noisetaker II 485 watts with fan speed set to minimum
Motherboard:       Biostar P4M890 M7 PCI-E
CPU:        Intel Pentium Dual Core processor 1.6GHz E2140, underclocked to 1.2GHz
             (faster CPUs with higher clock speeds not recommended as more RF radiation occurs)
Memory:         1 x 1GB DDR2 533MHz
Harddisk:          2 x 2.5” Laptop SATA drives 100+GB each, 8MB+ cache, using RAID 0 (avoid IDE).
          Use 3.5" drive bay adapter.
          RAID setup can be challenging but recommended with audible improvements as IDE overheads are eliminated.

Windows XP & Soundcard Software Optimisation
Remove all unused programs
Remove most Windows software components.
Reduce video interference and traffic.
Disable many system devices
Optimise soundcard buffers and set playback buffer for lower latencies
Limit Windows services.

[soundcheck]

Has anybody here experimented with building a computer transport with special components (and setup) that make a difference

Thanks
Frank
New Zealand

Hi Frank.

Interesting findings.

The tweaks you're mentioning I applied almost a year ago with quite some success.
We (especially Peter, Adrian, me and some others) were discussing the issues at the BD-Design forum.

Some issues to mention:

Processor-Clock:

The best sound you'll get at highest clock rate. The drawback is heat and powerconsumption.
Especially on a battery driven notebook that's not manageble.
However I found a nice compromise. I am running now 1.67GHz on a 2.16GHz Thinkpad P60.

Two solutions to overcome above:

1. a rocksolid overdimensioned powersupply,400-500W and/or
2. a fanless PC

with a notebook your options are unfortunately rather limited.


OS:

If you really want to go for a Microsoft based system, it has been proven that Windows 2000 is soundwise a much better choice over XP. It is a known fact that latencies are much lower under Windows 2000.

Vista needs in any case  to be compared to Windows 2000, instead of comparing Vista with XP.

Even though Peter is stating that he runs extrem low latencies with XX under any (Vista-!?!?) conditions, Flecko is reporting quite some audible differences between  Windows 2000 and XP.

Harddisk:

Highly recommened are professional 500GB drives. They are ten times more reliable than the cr*ppy consumer stuff.

2* 100GB in a notebook are not recommended. The 2 disks  generating a lot of heat and a lot of other unwanted effects especially in a notebook. If you go for .wav track playback -- 100GB is not sufficiant.

Any good SW-Player plays the tracks out of RAM. I doubt that RAID or IDE/SATA or anything makes a difference from
its technology perspective.
I'd rather suspect RAID generating more activity on the processor, thus it would be my less favourable solution.

That's of course different if you talk about a network based storage. 
My next invest will be most probably a 2* 500GB NAS connected to the audio-machine via GB-ethernet.

Good luck

[xp9433]

Soundcheck

Thanks for your input. I should make clear that the transport I want to build is a PC (I am just experimenting with the laptop). So all the comments I make about building a transport including the use of 2.5" SATA drive etc is in relation to a PC!

"The best sound you'll get at highest clock rate. The drawback is heat and powerconsumption.
Especially on a battery driven notebook that's not manageble.
However I found a nice compromise. I am running now 1.67GHz on a 2.16GHz Thinkpad P60."

Comment from "cics" about the E2140 duo Core compared to the 6300 > "Sound improves across the board - the level of sonic purity will shock you!" This sounds really good, but the processor has to deal with other than audio playback.


"If you really want to go for a Microsoft based system, it has been proven that Windows 2000 is soundwise a much better choice over XP. It is a known fact that latencies are much lower under Windows 2000"

I would likely go with a new Vista based system if Peter develops XX so I can configure with active XOs etc.

"Harddisk: Highly recommened are professional 500GB drives. They are ten times more reliable than the cr*ppy consumer stuff.
2* 100GB in a notebook are not recommended."

I was looking at putting 2*160GB SATA 7200 drives in the PC. With RAID it avoided the supposedly undesirable (sonically) IDE.

I guess I am trying to find out if anybody else on the forum has experimented with PC configuration that drag extra performance out of XX?

Cheers
Frank

[soundcheck]

Regarding my HD recommendation read the other thread (Western Digital WD RE2 - WD5000YS) . Don't waste your money on 160GB drives. We're talking Tbytes these days.

Core Duo over single Core Machines are an obvious improvement.

The introduction of PCI-E is IMO as great as the dual-core intro in terms of audio quality. On classic
PCI busses the devices had to share the bus. This was causing severe latencies on the bus, which had a major impact on e.g. USB sound.
With PCI-e the devices don't have to share the bus any longer. It's now rather a point to point connection.
You won't have any Latencies on a PCI-e bus.

Cheers

[PeterSt]

Hi Frank,

Nice effort you made here. And thanks for sharing the way you did.

I waited a few days whether someone would jump in, but I don't think there's anybody yet who took the effort. Also, there's of course my advise on this in this board : Don't be neurotic on shutting down services and all .

If it were for myself : yes, it should differ, but I expect it not to be audible, and strangely enough I don't feel the need to have it better. Of course I have this second agenda because XXHighEnd was deliberately made for you to avoid all this fuzz, so how could I myself listen to something different from what I advise to you ...

There are a few other things though :

Maybe the most important is that other players won't get stable. With this I mean that what one day is very okay, the other day just is not, and where one CD may sound good, the other just does not. For that matter, my own testing for "a" situation always lasts severeal days without switchting back or AB-ing. In these several days I listen to, say, 20 CDs (rather random) and if they all sound ok, it *is* ok. Also, if they *all* have this flair of whatever it is, this flair is incurred by the player (or my equipment for that matter). Now, with e.g. Foobar there is no consistency. The music-data itself creates the quality of that player, or whatever it is that changes it each time.
I too tweaked my PC(s) to the bone back then.
So keep in mind that "being consistent" is a quality (issue) by itself for PC playback. With the other players I know I get crazy of just that (found your best tweak, but tomorrow it counteracts ) and with XX I just never noticed that.

Oh, and do not use your best CDs to test things, but take the worst CDs to find out quality differences. Okay, maybe not the very worst, but those you were disappointed about so far. On that matter, I provoce the "theory" that you should own 1% or less bad CDs, instead of 50% or more. The latter is just not true, but your player can make them sound like that ...

The other thing I wanted to mention, is that one really should not test this with headphones.
Okay, I was very surprised when the first person here reported that he could hear that XX even sounded better through headphones (which you, say, confirmed), but I did not much expect that. The only thing it tells me that it "works" better even in areas I just did not put my attention to (and nor do I own headphones ).
They key to all is found in the room reflections. I state (and have defined) that when a player is good, room reflections don't bother you. And with XX they just don't (with the kind of prerequisite that your equipment must not destroy all, and with the main cause of impedance mismatches when it does). Not even in the worst rooms (I must say that I never tried an undecorated concrete hall though).
Obviously with headphones the phenomenon just isn't there, let alone how it can make a difference. Maybe it does have its influence ... I just don't know (what and how).

You could read Standing waves and XX on this subject.
To emphasize the matter I can tell you that since XX, my room correction equipment for my two subwoofers could go, and there is no way anymore even they can incur for annoying places in the room. It sounds good everywhere.

All 'n all, when you test XX for improvements or compare with other players, please take this all into account. This is where the differences really are and where the difference is made.

Frank, I too like to hear from others wether they could further improve the soundquality of XX and what they did for that. That's what this board is for.
But I do now want everybody to go out and tweak his/her PC in order to achieve that. Just because it was one of the explicit reasons XXHighEnd was created for : to not have to do just that stupidness. But for those who take the effort like Frank did, rather compare a fully tweaked Foobar et al, with an untweaked XX. I would even be more fair, just because I say tweaking isn't necessary. But have the soundcard out of the PC !! (which would be about SPDIF/passthrough in your (audiophile) cases *or* you use an USB connection and the soundcard is not involved).

Peter

[xp9433]

Peter

Thanks for your well thought out response. There are probably two issues for me in building a PC and using XX:

1.   Will a Music PC built for low noise, low power, low heat, low vibration, and low radiation improve overall sound quality when using XX?

2.   Will all the tweaks to reduce/cut overhead traffic and unnecessary services be worth doing with XX?

My guess/gut feeling is that the items (1.) would be good in any high end digital audio device, whereas the items in (2.) can be worked around, and are not so critical when using XX.

Some here recommend Windows 2000 with XX. You could interpret that to mean: “They already have proved that a lower noise, power, heat, and radiation environment plus low service overheads actually works!”

Windows 2000 uses only 64MB RAM and starts up only a few services, whereas XP needs 128MB and starts 33 services automatically, and Vista needs 512MB and starts 56 services automatically.

It is not surprising a non technical person like me has trouble trying to select the best PC building option.

I guess my best option is to work out the minimum processors/RAM/speed/HDD I need to run DVDs and store a music collection and go from there.

As for testing with other than heaphones. I am an old audiophile that until this exercise hasn’t worried much about digital. CDs were for the car and a portable player! I don’t have a decent CD player. And my current system is really built around TV/Movies, not music.

My music source was 2000 vinyl records and my modified Well Tempered Classic turntable (still have these). In this last year or so I sold off my expensive Plinius amps and Swan speakers (didn’t fit in the lounge of our new home – a long story) and I went for convenience. Only recently I decided to rebuild a quality all digital system.

So when you consider that, listening through good headphones and an outside USB DAC + headphone amp doesn’t seem quite so bad!

I have read your comments on standing wave and look forward to be able to test it.

Cheers
Frank

[PeterSt]

Hi Frank,

Let me first say that, of course, everybody is "allowed" to tweak the PC as liked, and that in my previous post I only tried to emphasize that it should not be necessary. Otoh we actually don't know what that can additonally bring. All in other words, your thread is about what to do to this respect and it is perfectly legit ... as long as people don't interpret it as a must to have very good playback with XX.

As a sidenote there's something I forgot to tell in my previous post : do note that you actually are comparing apples and oranges when you compare XXHighEnd on XP with a random (OS) version of Foobar, because for 99% chance with XXHighEnd under XP you were not playing bit perfect, while with Foobar you surely will have (KS or ASIO). This already tells us the first tweaks which *are* important for XXHighEnd : let run all bit perfect under XP or use Vista with Engine#3. The first (XP) ought to be unreacheable because it needs special drivers hence a special soundcard, or (and) with USB you won't be able to play tracks larger than 64MB. The path to outlay how to achieve this starts with "use Engine#1" (with #2 it is impossible and #3 is for Vista only) and ends with "have a Fireface and use the MME drivers". I mean, there's probably much more which will do it, but it is really up to yourselves to test it and prove the bit perfectness.

For USB I just don't know whether it is bit perfect, just because I never had the means to check that, which comes down to a DAC with a digital output so it can be looped back for checking. According to the structure of all, I think USB is not bit perfect, but half of the world just does not agree with that. In the end I just don't know.

Above "tweaks" are far far more important than anything else. Mind you, with XXHighEnd.
Btw, to avoid the rather confusing of the above, one should just use Vista and Engine#3 ... and keep in mind that Engine#1 and #2 in there just should NOT be used because of Vista explicitly resampling those, that being even worse than "not bit perfect" in general. Also see here for a small proof of it, which you can perform yourselves : Test track with distortion which isn't there... .

Back to your post Frank ...

1.   Will a Music PC built for low noise, low power, low heat, low vibration, and low radiation improve overall sound quality when using XX?

No, no, no and no again. However ... why not make the decision on this yourselve(s);
First of all, keep in mind that the context of (answers and) questions is XXHighEnd; I know what I do and do not (avoid) in there and from that base I say No. Now,
When you *know* that any bit perfect means, might it be JRiver, Foobar, XMPlay name them, are bit perfect because you just can measure it (and I mean loop back, not "DTS tests"), how would *you* explain the sound can differ by means of heat, vibrations, radiation and all. The only thing you possibly can say is that those things incur for jitter ... Well, be my guest, but I just don't know.
(the latter is exactly the point ... I don't know what the other players do to inject jitter, which they obviously do).
BUT :
There is one thing I found prone to producing jitter, and that's (generally spoken) everything which puts out the data to the soundcard. Remember, the soundcard should be outside the PC as a base rule, but it has to be fed. I think I can fairly say (meaning : without real proof) that any on (mo)board chip producing that data stream, injects jitter. Whether this is caused by radiation or the signal path to the chip accordingly I don't know, but to my findings a PCI card doing that job gives far better results. Sadly enough these cards amongst eachother differ also, but that I personally blame to power ratings (I mean kind of : when there's enough power to control the stream decently, the variance in between the bytes is less, hence less jitter). Here too : no real proof of that (not by me anyway).
This would or could imply that a sufficient (overrated) PSU is a good thing to have.

2.   Will all the tweaks to reduce/cut overhead traffic and unnecessary services be worth doing with XX?

I really can't think of how. However, this is kind of dangerous because I did not even try it out (and you did, and heard differences). BUT :
What you tried for XP, I tried the other way around for Vista/Engine#3 and I tried to blow up the PC as much as I could and I couldn't hear a difference, let alone notice glitches etc. For that matter, please note that XXHighEnd was actually born (created) for Vista/Engine#3, and that the possibilities (hence Engine#1 and #2) for XP were created as a surrogate solution to my so many months working on #3 never seeing the light at the end of the tunnel. This more or less implies that what happens in XP (better #1 and #2) is a kind of rough implementation of what #3 (back then) intended. It works okay I think, but in theory much much less than Vista/#3.

From this, obviously is implied that you really should checkout Vista/#3 before you take the troubles in the XP environment of which I say that it really cannot be the best.
Besides that, Vista/#3 sounds TOTALLY DIFFERENT from XP#1/#2 (even with XP in "bit perfect mode" which I just can have).

Because of the latter, the big fun should be that when you find differences between XXHighEnd and another player, your next level of listening has been born then, because now you are going to be surpised with the differences in between XXHighEnd settings itself. From that you learn better what to listen to, and in the end you really won't be able to even begin understanding how it came that you liked PC playback before (hence with the other players). So keep in mind : this is the process of knowing what to listen for / look for, like the standing waves thing. On that matter I often have another good example for everyone to checkout very easily : put your hand or body on a rock solid element like a heavy table, and *know* within really one second that #3 is playing, just because you feel how it impacts on very low frequency vibes which are so straight that they make move anything. And ... no stories here, you just can check it out ! Uhhmm ... not with headphones.

All 'n all you see that there so much to explore which you don't know about yet, and of which you don't know yet how it sounds i.e. how to perceive it, that you really should start tweaking the PC lateron. FIRST you should get rid of standing waves if they are there, just because I say that if you have them, something is wrong with the equipment (hehe, just keep this in mind for some later point in time, when you fully agree with it). Only then you can start judging XX' settings, and only after that you can start the finetuning (which it really would be for XX as per your own experience, I think).


On Windows 2000 ... I am kind of more or less about (etc.  ) sure that this is about the KMixer of XP what this is compared to.
Those who say that W2K sounds (way) better than XP ... I fully believe because of the logic with KMixer. Otoh, they should really compare with Vista/#3 and if *then* W2K sounds different for the better, I am going to find out why and no matter what, copy that "behaviour".
This thread caused me to drag a W2K PC to my house, which I very sadly still didn't listen to because of so many other things to do currently. I really should do it soon ...

Peter

[xp9433]

Peter

Thanks for perservering with my queries. Your clarification has helped.

I will order a new PC with Vista and configure with some of the recommendations that soundcheck made.

Of course then I will bug everyone about DACs and Amps .....

Cheers
Frank