Mac Pro Quad Boot Reference Music Server

I'll be very interested in knowing the results with openSUSE 11. What software are you going to use on openSUSE for playback? , one of the most popular in the Linux world is Amarok.<br />

Hey Chris, can you please explain (or point me in the right direction if you have previously) how the Berkeley Audio Alpha DAC enables you to test for bit perfect output? I'm sure it's quite simple, but I can't imagine how it does this. Is it a matter of comparing the original signal to what is output? If so, how can one be sure the original is bitperfect?<br />

<br />

Would love some more info... nice quad boot BTW, although I'm glad it's not my machine

Hi Chris,<br />

<br />

I'd be interested in how it verifies 'bit perfect' too.<br />

<br />

In the past, I have done this in a rather convoluted way, by importing the two files to be compared into some signal processing software (Agilent SystemVue), then performing a mathematical subtraction. Only if the result is zero over the entire file length, is the comparison 'bit perfect'.<br />

<br />

Interesting that the files sound different. This possibly points to a short term sample rate anomaly. Maybe one of the OS's doesn't perfectly re-time the samples it gets from memory ?<br />

<br />

Chris.

Chris,<br />

<br />

Looks great, all those operating systems on the same hardware.<br />

<br />

You tease us by saying they sound different, and are both bit perfect ... but which sounded the better ?

Hi pichonCalavera - I'm going to try as many as I can that appear like viable solutions. I will likely start with Amarok.

Hey poo and Chris - The Berkeley Audio Design Alpha DAC verifies bit perfection when playing an HDCD encoded track and illuminating the HDCD indicator when all the bits are passing through perfectly. Don't quote me on this, but I believe the HDCD flag is in the 16th bit. Any bit variation and this flag will not be passed to the DAC. Thus the HDCD indicator will not illuminate. All of the common things like using a digital software volume control and using the wrong output plug-in in MediaMonkey will not allow the HDCD indicator to illuminate.<br />

<br />

I think comparing two files with software like Agilent SystenVue is a good thing to guarantee your files are identical. But, it doesn't help determine if the playback application like MediaMonkey or iTunes is passing the track bit perfect to your DAC.<br />

<br />

The sound differences between bit perfect applications are definitely interesting. Like I said earlier, I don't have the answers, but I am continuing the research. Hopefully some other readers can test a few applications and share their opinions.

Hey BEEMB - I'm not ready to answer that question yet. I know which one i like better right now, but it would be a disservice to pick a winner without many more extensive listening sessions. I know that's kind of a lame answer, but it's an honest answer.

Hi Chris,<br />

<br />

OK, thanks.<br />

<br />

In John Atkinson's review of the AE, he indirectly proves that iTunes is bit-perfect :-<br />

<br />

http://stereophile.com/digitalprocessors/505apple/<br />

"Some audiophiles have dissed the AirPort Express on the grounds that its digital output is not bit-accurate. However, I found that this was not the case, that the data appearing on the AE's digital output were identical in the original file. To check this, I compared a WAV file with a duplicate that I had captured on my PC from the AirPort Express's S/PDIF output. I used iTunes on my PowerBook playing a version of the file encoded with Apple Lossless Compression to feed data to the AE. The files were bit-for-bit identical, proving that the AirPort Express is transparent to the music data (as is ALC, for that matter). "<br />

<br />

So, if the bits are right, the only other thing to get wrong is the sample timing. Presumably, OS X sounded best ? It is de facto in recording studios after all.<br />

<br />

Chris.

Hi Chris - I think we should clarify a little bit more what he proved. In my opinion he proved the combination of iTunes and the Airport Express are bit perfect. I am pretty sure that iTunes by itself on a PC is not bit perfect.<br />

<br />

Good link though!<br />

<br />

I am withholding my opinion, for now, about which systems sounds better :-)

Hi Chris,<br />

<br />

Yes I think you are right.<br />

<br />

John Atkinson's test used iTunes on his Powerbook (OS X) to stream to the AE, whereas the captured file on his PC was pre-stored from the SPDIF capture.<br />

<br />

All my tests on file comparison were done using iTunes on my iMac. I only use Windoze when I absolutely have to, and run it as a VM using VMWare Fusion.<br />

<br />

Chris.

Hey Chris - Have you done any listening through VMWare Fusion. I know that sounds painful, but I am just curious. I spent about 5 minutes listening with my XP / Parallels combination and called it a day. It's far too convoluted with the virtual layer between the OS and the hardware.<br />

<br />

I can't see anyone actually doing this, but you never know :-)

Hi Chris,<br />

<br />

No I've not tried that.<br />

<br />

I only really use XP for running scientific / engineering apps which have no OS X equivalent :- Mathcad, SystemVue, and Microwave Office, which is a professional electronic circuit simulation package. Using these apps, I find I get ~ 80% native Windows performance.<br />

<br />

Chris.

Hey Chris I'm assuming you are still using the AES connection on the PCI card to the DAC? What other DAC;s out there use AES do you know just for ref?<br />

<br />

jimi<br />

<br />

PS you are nuts you know! 4 operating systems? you should be the other guy on the mac pc commercials, wonder what they would say to you? LOL

I think the claim of being able to detect bit perfect input is a bit to much.<br />

Inputs like S/PDIF or AES/EBU are unidirectional, no way to establish if they are bit perfect on receiving.<br />

It is a claim not made by Berkely and not by the inventors of the HDCD, Pacific Microtronics (http://www.vt52.com/ftp/semi/pmd100.pdf)<br />

The claim that you are playing a HDCD when the HDCD led is on, it a valid one.<br />

If its on, it probably means that at least the LSB is not altered and maybe this applies to the other bits as well. http://en.wikipedia.org/wiki/High_Definition_Compatible_Digital<br />

<br />

Playing digital audio invokes many steps:<br />

- the Codec to interpreted the contents of the audio file<br />

- DSP by the player software<br />

- the driver to route the audio to the sound card<br />

- the sound card to convert the audio stream to S/PIDF (you are probably using the optical out so S/PDIF over Toslink)<br />

<br />

Using different OS, the codec, the player software and the driver are different.<br />

Let’s assume that in case of simple formats like AIFF or WAV (uncompressed straight PCM audio) the software don’t make a difference.<br />

Let’s assume everything is bit perfect regardless of the OS.<br />

What is left is the jitter.<br />

S/PDIF sends the signal as a bit stream (so digital therefore robust) but the sample rate is derived from the rate at which the bits are pouring in. This is depended on the sender, the clock of the Toslink. A clock is a analogue device. That’s the hybrid character of S/PDIF, digital signal but analogue timing and therefore analogue imperfection.<br />

<br />

It might be that the Toslink produces more jitter as the system load increases (more EMI/RFI and/or dirt on the power). If this is audible, it means that the DAC is sensitive to input jitter to.<br />

It might also be that the claim of bit perfect is false. In absence of any specification of your testing procedure it is almost impossible to establish what might be a cause.<br />

This may be wrong, but I thought iTunes dropped the ability to pass an HDCD signal after Ver. 6.4?<br />

<br />

Four OS on one machine? That's flat out sick! And here I was proud that I can dual boot on my MacBook.....<br />

<br />

<br />

Thanks for the extra info on bitperfect detection Chris, although I'm not sure I understand any better how the Berkeley unit does this in such a foolproof manner. I guess there is also an assumption here that the rip or file itself is bitperfect, though I'm sure few of us ever check; in fact in some situations (buying music on line for example) we generally can't check the accuracy of the file. It is the 'real life' situations like these (though this is a small example) that make me sit back a little and wonder if these 'reference' systems pull more wool over our eyes than anything else at times... "The little green light is on so it is OBVIOUSLY bitperfect!"

Hi Nels - I'm in contact with the people from Pacific Microsonics, or what was PM, regularly and I can say iTunes still passes HDCD.

Hey Jimi - Yeah, I'm still using my Lynx AES16e PCIexpress card and the HD26 to XLR AES cable. There are quite a few DACs around that support AES. I would say more support AES than USB. The Stello DA100 Signature that I recently reviewed has an AES input. It will only handle up to 24/96 though.<br />

<br />

Ok, so I'm a bit nuts with the quad booting music server. I love the comment :-)

sorry should have added upto 192?<br />

<br />

jimi

Hi roseval - Welcome to Computer Audiophile. I also checked out your site and like the information you have there. <br />

<br />

Thanks for the detailed post. I value your opinion but disagree with it for the most part. I talk to the co-founders of Pacific Microsonics regularly and I am very confident with my statement that illuminating the HDCD indicator on the Alpha DAC equates to bit perfect output. If the HDCD indicator illuminates on the Alpha DAC, the data is uncorrupted. Theoretically, it is possible to alter HDCD data using specialized software while not touching the LSB, but all of the typical mechanisms that might alter data in a computer environment such as level shifting, dither, SRC’s, etc. will definitely affect the LSB. And, if the LSB is altered the HDCD code is lost. So, as a practical tool, presence of the HDCD light indicates no alteration of the data file. <br />

<br />

In my system I am using a Lynx AES16e digital I/O card that outputs via AES. The Lynx card has around 20 picoseconds of jitter. I also use the clock on the Lynx card. I'm not a big fan of the Toslink output in general. I playback uncompressed AIFF files using iTunes. The Audio Midi sample rate is always set to match the sample rate of the music I am playing. <br />

<br />

<i>"...In absence of any specification of your testing procedure..."</i><br />

<br />

I hope it is clearly evident that I am not running a scientific laboratory test. <br />

<br />

<br />

Anyway, thanks again for your post and I hope to see you share your thorough knowledge of this topic elsewhere on Computer Audiophile.

Thanks for you reply.<br />

I don ‘t expect you to be a DIY scientist and specify a lengthy protocol, but it is not exactly clear if you do hear differences between different OS, it is this with HDCD on.<br />

Indeed you do have a point, if the LSB is not altered the higher ones are probably not altered either.<br />

But if we assume this is the case, all OS deliver bit perfect output so the only other parameter left is the jitter of the AES.<br />

Lynx probably make some of the best cards and 20 ps is around what some do consider the audible threshold.<br />

Maybe you can use your excellent contacts with industry to have the output measured.<br />

Hi Chris,<br />

<br />

Thinking more about this, even the jitter spec isn't the whole story, since it also depends on the spectral distribution on the jitter.<br />

<br />

If the jitter spectrum on the DAC clock is uniform and noise-like, it will be audibly benign, and just raise the output noise floor slightly. Even 1ns of jitter would sound OK in this case. (100ps is equiv to 100dB SNR in a 20kHz bandwidth).<br />

<br />

However, if the jitter is concentrated into a number a discrete tones (sidebands) close-in to the clock frequency, this will be much more objectionable. So, it is not sufficient to simply quote a jitter spec in isolation.<br />

<br />

Chris.

Hey Chris - I agree with you 100%. A simple jitter number in isolation is not the complete story, but it's a start anyway.

Hello Chris,<br />

Thanks for a great site dedicated to (computer) high end audio.<br />

As an obsessed audiophile I've taken great care over the years in putting together the best possible hi-fi system I can afford viz. Jadis JD1 + JS1 DAc + JPS2 preamp + JA200 amp (KT88's) + Genesis 2.5's + Kimber Select all round and Audioquest Optilink between drive and DAC (ST).<br />

Now, I'm very interested in replacing my JD1 with a music server.<br />

I'e read very favourable reviews in TAS and The Stereo Times of the Sooloos server and I'll be able to listen to it in December.<br />

The only snag is the steep price: $13 900/13 900 € (yeah, the importer here applies an extortionate one to one exchange rate!).<br />

Fortunately, I'm a Mac user too so I'm thinking that for much less money and even adding a Troll Touch touch screen I would be saving a lot of money. An sell my JD1!<br />

I know that you're thinking along the same lines except that you recommend changing the soundcard. Why?<br />

Isn't it possible to bypass it (à la iPod with Wadia 170i) and then connect the computer directly via say an optical fiber link (ST) — after installing a ST connector on the Mac — to a top notch DAC (Wadia, Jadis, etc.)?<br />

Now, how about an iPod + Wadia 170i modified with ST output and ethernet connector and then plugged into say a 2To HD?<br />

Thanks,<br />

GM<br />

<br />

Hi GM - Yeah, the Sooloos is very spendy. Great interface though!<br />

<br />

I use the Lynx card to out put a digital signal because it is a proven card, has low jitter, and allow me to play back up to 24/192 material.<br />

<br />

Macs have a built-in optical port to connect to a Jadis DAC, but optical is less than ideal. You can do it any way you want really. The wadia i170 is not a reference component. It is great for certain things, but not as a server.<br />

<br />

Let me know what you're thinking for your server and we'll get it all figured out.