Mac Pro with Lynx s/card or Mac Mini with Weiss INT202

[ianmacc]

Hi

 

I have a high quality Naim hi fi system including the Naim DAC.

 

I want the best possible sound quality so I have three possible choices of source:

 

Option 1/ SSD Mac Mini with Boulder PSU running Amarra, firewire to Weiss INT202, then on to Naim DAC.

 

Option 2/ Mac Pro Quad Core running Amarra, firewire to Weiss INT202, coaxial cable on to Naim DAC.

 

Option 3/ Mac Pro Quad Core running Amarra with Lynx AES16e sound card, coaxial cable on to Naim DAC. No Weiss INT202.

 

Could anyone say if there is a clear winner in these three options?

 

Looking forward to your help and advice

 

Ian (Manchester, UK)

 

[Mahesh]

Hi,

 

I think the real question here is whether you want Weiss INT202 or Lynx AES16e. The firewire Weiss will work with any modern Mac. The Lynx card requires a Mac with a free PCI express slot, which rules out Mac Mini.

 

Changing Mac model alone will not affect sound quality, as long as everything is properly configured. So, choosing Weiss leaves you free to select the Mac you prefer, purely based on preferences not related to sound quality.

 

The same goes for SSD vs. rotating hard drive. Both provide data transfer rates high above what is required for playback of 192/24 material, so the choice will not influence SQ. As for operational noise, a Mac Mini with SSD surely is quieter than a Mac Pro. But moving the Mac into another room eliminates noise problems.

 

I don't have any experience with either card, so maybe someone else will shed some light on which is the best choice.

 

[johniboy]

One opinion on a comparison RME AES32 (comparable to the Lynx) with Weiss INT202 can be found here: http://www.computeraudiophile.com/content/Highest-quality-AES-EBU

 

At the end it is probably a question of taste...

 

[Mr.C]

I would say that the Mac Pro w/ Weiss would be the best. The Weiss has bested the Lynx according to some users, and the Mac Pro's are accepted as sounding better than the Mini for whatever reason. Another option is a G5 with Weiss which may result in a better (and less expensive) option than the current Mac Pro.

 

[Mahesh]

Hi Mr.C,

 

I would like you to explain why you would hear a difference between playback on a Mac Mini vs. a Mac Pro, considering both are correctly configured, and you use the same sound card (a Weiss, for instance) and playback software.

 

Mahesh

 

[Audio_ELF]

Some people have noted differences in SQ when using high end equipment and comparing MacMini to older G5 machines. I am unsure of similar comparisons between MacMini and Intel MacPro machines. There have been theories postulated that the different power supplies can provide differing signal quality, or that the G5 chip somehow provides as better processing path.

 

For most systems the differences are unlikely to be noticed.

 

For Ian... if you choose the Lynx AES16e card and a MacPro, you'll need some form of transformer to convert the balanced AES signal into an SPDIF signal for the Naim DAC. A better choice may be the RME HDSPe AIO card which has a dedicated SPDIF output.

 

Eloise

 

[johniboy]

i am using the pci version of the RME HDSPe AIO card and it is really great! Especially the BNC connection is rock stable and delivers wonderful sound. Unfortunately I did not have the chance to compare it to the weiss devise but would love to!

 

[ianmacc]

Thank you for all your replies.

 

I think I am going to stick with the Mac Mini, Boulder PSU, Weiss INT 202 into Naim DAC.

 

I have already auditioned this set up and it sounded fantastic. For me it was close, if not equal to the sound of Naim's CDS3 in my system which is approaching £8000.00.

 

Kind regards

 

Ian

 

[AliRowe]

Hi

I am new to the forum so please be patient.

 

1- I also have a Naim pre/power amp combination. Their whole philosophy is the pure power supply is king.

To upgrade just introduce a dedicated power supply to the dac, cd player or pre amp and instant quality improvement.

I have done it and can testify to it.

 

2 - Therefore apply the same theory to the mac mini/mac pro and the power supply could be why some notice the improvement on the Mac Pro or G5. Bigger more stable supply. Thumbs up to the Boulder supply for Mac mini.

 

3 - Another card to consider would be the Asus STX pci as that is also reviewed elsewhere on this site.

I am sure others would be better placed to review the merits of the Lynx vs Asus cards.

 

4 - considering you have a Naim Dac and the Weiss Int202 and the care taken to develop them, the quietest " electrical environment" computer should be used to supply them. Along the lines of the CAPS server. Fanless case, SSD's, quality Psu etc.

 

What is it we are trying to achieve from the computer ? Bit perfect data stream with low jitter? I'll throw that one out there.

 

I would be interested to know what an toslink optical connection would sound like to the dac from MacMini as this would decouple the electrical interferrence.

 

Alistair

 

 

[Mahesh]

With all due respect, I think there are some misconceptions of what a computer is and does.

 

According to Encyclopædia Britannica a computer is a "programmable machine that can store, retrieve, and process data". (http://www.britannica.com/EBchecked/topic/130429/computer)

 

In order to do so, a computer consists of a CPU (central processing unit), and different data storage media (ram/cache/disk), and last but not least, data buses for transporting data back and forth between internal and external components. Data flow happens at almost unimaginable speeds, and with very high precision. In fact, if only a single bit is incorrectly transferred, the whole system can crash. In order to ensure that data is never misinterpreted, various error correction procedures are implemented at different levels, both in hardware and software.

 

If we assume that everything is correctly implemented, a computer does not do any mistakes, at least not on the hardware level. Now, one can argue that since a computer is designed and manufactured by humans, and since humans can and do make mistakes, then there ought to be errors introduced in most computers. We've probably all heard about the famous Pentium bug (http://en.wikipedia.org/wiki/Pentium_FDIV_bug). And yes, one cannot guarantee that a computer is 100 % error free. But today's design and manufacturing techniques involve a high level of quality control and a stringent testing regime, ensuring that by the time a computer reaches the market, the chances for hardware bugs are pretty slim. Also, since the whole system is so heavily dependent on correct bits being transferred, an error will in most cases manifest itself early in product testing, and thus be corrected before release to consumers.

 

Since bits are represented as electrical signals internally, a computer needs a power supply. One might think that one can improve a computer system by tweaking the power supply. This is not the case. As long as the power supply is properly dimensioned for the power requirements of the computer, and the voltage/current delivered is within the operational range of the system, improving on the power supply will not make a computer better at processing data. This is in contrast to the power supply in your amplifier or CD-player, where the quality of supplied power can and will affect the end result--the sound you hear.

 

Amongst other things, a computer can be programmed to transport audio data from a file on disk to a sound card, be it an internal sound chip, a Lynx card on the PCI data bus, a Weiss DAC on the firewire bus, or another sound device. How well the computer manages this task depends on the software it runs. There is a complex stack of different software involved. Of course, the playback application, be it iTunes, Amarra, FooBar2000, or another app, is an important piece in the "software puzzle". But you must also consider the operating system of the computer, with its device drivers, process scheduling, etc. In the case of digital audio playback, it is important that the playback application process is prioritized before other less time critical tasks. With the multicore, high clock frequency (yes, the computer has an internal clock, too!) processors of modern computers, playback of digital audio should not represent a big challenge.

 

Keep in mind that a computer does not treat audio data any different than other bits and bytes it processes; the same error correcting systems are in action, ensuring that the hardware treats the data exactly as it is programmed to. So, if all the software involved works as intended, any fairly modern computer should be able to deliver a digital audio stream to the output device without errors.

 

Now, there are ways a computer may have an impact on sound quality, some of which are:

 

- If the playback software or OS alters bits in the audio stream before it reaches the output device, for instance, if sample rate conversion is applied.

 

- If the audio stream is not delivered to the output device in time for conversion or clocking onto the digital audio connection.

 

- The requirements of a computer power supply are quite different than those of an audio power supply. The switching computer supply may generate noise back onto the mains wiring, and/or electromechanical fields, etc. If this noise/radiation reaches the audio circuits, it can interfere with sound quality.

 

- If the sound device is dependent on the computer power supply, and does not implement its own voltage regulation and/or noise filtering, the computer power supply may affect the quality of the output signal.

 

However, if every other factor is kept constant, simply switching computers will not sound different, assuming that both computers are powerful enough for the playback task.

 

Statements like "a Mac Pro sounds better than a Mac Mini" are simply meaningless, without any closer explanation of the two machines, how they were configured, how the software was set up, etc. There are too many factors involved, every one of which may or may not contribute to the final result.

 

I have also seen statements claiming that PPC Macs sound better/worse than Intel Macs, allegedly because the PPC is a better/worse processing unit for audio data than Intel. Again, a computer does not care about what data is being processed, it simply does as it is told by the software, no matter which processor it is equipped with. However, Mac OS is compiled specifically for the CPU it runs on. Hence, the Mac OS binaries look different for the two platforms, even though the OS version number is the same. It might be that the PPC compiler Apple uses generates more efficient code than their compilers for the Intel platform, or vice versa. But I highly doubt that it would influence either platforms ability to playback audio files.

 

Finally, I would like to comment on the claimed differences between bit perfect software players. As far as I can see, if you achieve bit perfect digital output from the computer, that is, the bits received by the DAC are exactly the same as the ones in the file on the hard drive, how can there be any differences between different playback software?

 

I am not arguing that different sound output devices, analog or digital, doesn't sound different, and I am certainly not arguing that different DACs doesn't sound different. But if your computer is able to output bit perfect audio to the sound device, there is nothing more to improve on in that end of the playback chain.

 

Mahesh

 

[The Computer Audiophile]

Hi Mahesh - Thanks for sharing your comments with sources to support your points. Very cool. I disagree with a few things and agree with a few others, but so be it. It's no fun to go back and forth over some of this stuff and causing a scene. This hobby is supposed to be fun and a great music server can increase the fun greatly.

 

That said, I think a good discussion could be had about whether or not two bit perfect digital signals can sound different.

 

"... if you achieve bit perfect digital output from the computer, that is, the bits received by the DAC are exactly the same as the ones in the file on the hard drive, how can there be any differences between different playback software?"

 

I think it would be interesting to run the following test.

 

1. Playback any song on a good system with low jitter and measure the output to guarantee bit transparency.

 

2. Change nothing except the amount of jitter. Increase jitter to 999 nanoseconds, or some outrageous number so the effects are clearly audible while at the same time maintaining bit transparency.

 

The results should be contrary to the belief that all things bit identical sound equal. I could be wrong, but it's something I've thought about for a while and would really like to test.

 

 

 

[Mahesh]

Chris,

 

I certainly agree with you. Changing the amount of jitter on an SPDIF signal reaching the DAC certainly will make things sound different.

 

But jitter is not an issue internally in the computer, not until you clock the audio stream out on an SPDIF interface, or directly into a DAC with fw/usb. Up until that point, bit perfect is bit perfect, regardless of jitter. Maybe I was not clear enough on this in my writings.

 

Mahesh

 

PS. I find this article on jitter in digital audio quite informative: http://www.tnt-audio.com/clinica/diginterf1_e.html

 

[AliRowe]

Hi Mahesh

 

Thanks for such detailed comment, this is what makes this site worthwhile.

 

Chris, an aside, the look, layout and quality of the topics are a credit to you.

It cerainly made my enjoyment of music and tech more worthwhile.

 

I digress, Mahesh, you have said much that backs up what I have always believed. Information is information, bits are bits.

No matter what computer the bits delivered to the dac from the hard disc SHOULD be the same. So following that through, would the method used to connect to the dac, cable or toslink, make any difference to the audio stream?

 

As you point out, the software and implimentation have more to do with changing things than the hardware and once the signal has left the digital domain things get much more sensitive, and are open to "influences" of equipment.

 

My computer at present is a PC with XP64 running on it. I am ripping cd's to FLAC and AIFF using db poweramp.

I am looking to output the files to a dac into my Naim system.

If the software is of "issue" for playback, I am looking at media monkey and foobar, suggestions as to which would be appreciated. I really want to integrate with iTunes for convenience if possible.

I am thinking of installing the asus STX in one of my free PCI slots and taking the data out to my dac via the S/PDIF output on the sound card.

Any comment in the viability of this solution would be appreciated.

Hopefully this will give me a quality data stream.

I do not wish the soundcard to handle the D/A conversion, I just want a replica of the data to come out of the computer environment accurately and faultlessly. Careful dac selection will then give me the sound I prefer.

 

Mahesh, thanks again for confirming things for me.

 

Another option would be to go down the Mac/amarra route as Ian has chosen to, but if I understand correctly the results should be the same, up to the point of data entering the DAC, as a PC set up. Given that the respective OS and playing software is set up correctly to deliver what it should, in the case of cd's - 16/44 bitstream.

Another point Mahesh, would adding a boulder power supply to the Mac Mini as Ian indicated be worthwhile? How can that improve the quality of the digital output?

 

Thanks

Alistair

 

 

 

 

[Mr.C]

"Statements like "a Mac Pro sounds better than a Mac Mini" are simply meaningless, without any closer explanation of the two machines, how they were configured, how the software was set up, etc. There are too many factors involved, every one of which may or may not contribute to the final result."

 

The Mac Pro is inherently configured differently from the Mini being as they are different form factors. I can think of a few reasons to explain why it might sound better. The Power supply as has been discussed

 

"The switching computer supply may generate noise back onto the mains wiring, and/or electromechanical fields, etc. If this noise/radiation reaches the audio circuits, it can interfere with sound quality."

 

or other components which generate RFI/EMI.

 

It has been noted at a high profile event with many industry professionals in attendance that SSD sound better than spinning disks. Why? There is little explanation, but it still makes a difference.

 

Personally, I listen to music, not theory.

 

 

"Of course, the playback application, be it iTunes, Amarra, FooBar2000, or another app, is an important piece in the "software puzzle". But you must also consider the operating system of the computer, with its device drivers, process scheduling, etc. In the case of digital audio playback, it is important that the playback application process is prioritized before other less time critical tasks. With the multicore, high clock frequency (yes, the computer has an internal clock, too!) processors of modern computers, playback of digital audio should not represent a big challenge."

 

So why do the software sound different then?

 

 

 

[Mahesh]

Hi Alistair,

 

Thanks for enjoying my efforts on explaining computers.

 

Chris, an aside, the look, layout and quality of the topics are a credit to you.

It cerainly made my enjoyment of music and tech more worthwhile.

 

Thumbs up for that! It is amazing how this site is almost completely free of person attacks and flaming found on most other audiophile forums. And there are A LOT of knowledgeable members, willing to share their knowledge with others. (Chris, do you moderate the postings a lot, or are people here as friendly as they seem?)

 

I digress, Mahesh, you have said much that backs up what I have always believed. Information is information, bits are bits.

 

Yes, as long as the bits are inside the computer. I don't think that anybody would accept that their Word documents suddenly contain different characters than the day before. The computer has plenty of error correction mechanisms, both in hardware and software, and on every level, to prevent bit flipping from happening. But when an audio stream leaves the computer, things are not that simple...

 

The method you use to connect the DAC to the computer can and will influence sound quality. An inherent problem with the SPDIF standard [1] is that the clock signal is interleaved with the data stream, with something called biphase mark code [2]. So the data stream carries both clock signal and audio data. On a coaxial cable, the signal is transmitted with electrical signals, and on a toslink cable, optical signals are used, as you probably know already. Both have potential problems as transfer lines for digital audio signals. I'm no electronics engineer, so I will not try to go into details on that. For a general discussion of jitter in digital audio, see [3].

 

The length, quality and construction of a coaxial SPDIF cable may affect the signal it carries. Nevertheless, in most situations the bits carried through the cable will reach the DAC at the other end without errors, but this is just half the information. The clock signal embedded in the electrical signal will be more or less jittery, depending on many factors. It is a non-trivial task to remove the jitter from the signal, and, what is more important, why would you expose the audio data in your computer to a potentially jittery link in the first place?

 

I think that keeping the audio data in the asynchronous domain (which the computer is) as long as possible before converting it to a synchronous signal (which is the inherent nature of a sound wave) makes a lot of sense. That's why there's so much talk about async firewire and usb DACs in this forum. In such a device, data is transfered asynchronous from the computer to the DAC, where a receiver chip clocks the data into the converter chip with a (hopefully) high-precision clock signal. The receiver, clock and converter chip can be kept close to each other, reducing the chance of introducing jitter to the audio data before conversion.

 

There are tons of ways to actually implement this concept, and I have tested none of the products on todays market. So I am in no position to advice you on which device to buy. (If you look at my signature, I use a Squeezebox as digital source, a device reported to have a fair amount of jitter on its digital outputs.) But from my theoretic understanding of the subject, going for an async fw/usb DAC is very wise. Try to audition a few devices and see which one you like the best.

 

When it comes to playback software, I firmly believe that any program being able to output bit perfect data to your sound card, SPDIF interface or async DAC will do the trick. My belief is based on my understanding of how computer hardware and software works. But quite a few people claim to hear differences between different playback software, but to my knowledge, bit perfect is bit perfect as long as you are in the asynchronous domain of the computer. Again, I advice you to test the different software available, and see which one you like the best. For some reason, the free or cheap options are almost only available for Windows. On Mac, you can use iTunes, if you can live with manually switching sampling rate. Or you can spend $$$ on Amarra. You chose. But it is possible to achieve bit perfect playback on Windows. Chris has published a very nice article on how to achieve this with Windows 7 elsewhere on the site.

 

Whether upgrading the power supply of a Mac Mini may improve things, IMO depends on what device you use to output digital audio from the Mac. If you use the integrated SPDIF out, then, in theory, upgrading the power supply may improve things. The reason being that the internal SPDIF output depends on being fed power from the supply via the motherboard. But this depends on the internal hardware design of the Mac, so there are many factors influencing the end result. If you connect the Mac to an external async fw/usb DAC with its own power supply, I cannot see how upgrading the power supply of the Mac can make a difference. But again, some claim to hear a difference, so go ahead and experiment to find out what you think.

 

Good luck on your way to computer audiophile nirvana! :-)

 

Mahesh

 

[1] http://en.wikipedia.org/wiki/S/PDIF

[2] http://en.wikipedia.org/wiki/Biphase_mark_code

[3] http://www.tnt-audio.com/clinica/diginterf1_e.html

 

[Mahesh]

Hi Mr.C,

 

It has been noted at a high profile event with many industry professionals in attendance that SSD sound better than spinning disks. Why? There is little explanation, but it still makes a difference.

 

Let me make it clear that I have great respect for other people's views and opinions on these subjects. I am in no position to question the experiences of other people. A persons perceived difference in sound quality between two components cannot be questioned. What can be discussed are the reasons for perceiving the differences. And it is not possible to ignore that psychoacoustics and placebo effect may be a part of the explanation. I am not saying that this is the explanation for the event you refer to, I was not there and cannot tell.

 

So why do the software sound different then?

 

I don't know. But let's say that audiophiles are divided down the middle on this question. Then, half of all audiophiles can hear a difference between different bit perfect software. Compared to the total number of inhabitants of this planet, I think it is not possible to claim universally that there is a difference.

 

Note: I'm not pretending to have the final answer, or to be closer to the truth than anyone else.

 

Mahesh

 

PS. I, too, like listening to music a lot. But I like theory, too! :-)