Melco N1A

Some interesting observations on here about different people's experiences with Mac Mini specifically, and "Macs 'n' Dacs" generally.

Certainly not all Macs are the same, with, it seems, quite wide variation - unsurprisingly given different model architectures and setups. There seems to be a general consensus (not thinking specifically of this forum) that the Mac Mini is either the best, or at least one of. Some comments here suggest that there may be variation between MMs, which is less expected, though whether that is different model/setup variants is unclear. For serious hifi use my expectation would be that people are only using the MM stripped down to headless, with bluetooth and infrared off, and with the output bypassing the MM's soundcard by using USB output, the USB bus used dedicated to that function, and no other software in use. (Any serious hifi use of any other computer I would expect to be similarly stripped back, and if headless not possible then at least display blanked.) However it is clear that this is not always the case, and in many cases there is no confirmation, so it is not always easy to know how valid comparisons of sound quality may be: it would be helpful if everyone reporting their assessment were to give detail of the MM (or other) setup addressing at least the matters I've just listed, and certainly any comparison is only valid if the setup is known.

Rendering software is not all the same, so assessment of a computer source alone is of little use, instead it needs to be the package of machine and specific software. My consideration here is only with Audirvana as that is what I use, having come to it from the many assessments made by other people on these forums and elsewhere. Audirvana can be set up in different ways, but in particular it can only be expected to give its best in its optimised mode as given in its manual, i.e not using iTunes, its dedicated mode on, and with dedicated usb output etc. I have been surprised at some people not using it this way, and again any comparison is only valid if the setup is known.

Then there is the matter of removal of electrical noise, particularly RF, from the computer's output. All standard computers inevitably will be less than 'clean' in this regard, even substituting fancy things like linear power supplies, so if connected directly to the usb input of a DAC all will depend on the effectiveness of the DAC's isolation and filtering, with some much more effective than others, while galvanic isolation does not mean the same as RF filtering. See* below. So unless the DAC has GI and highly effective RF filtering it is crucial to have an isolator to perform one or both of those functions between the computer and DAC, and of course every such device is likely to be different, some more effective than others - and even assessment of different isolators can only be totally valid for the source and DAC used in the assessment. Therefore details of any isolators used also need to be made clear in any comparison.

Then there is jitter - however assuming the DAC is asynchronous and reclocks the data, as I believe most are nowadays, this comes down primarily to the effectiveness of the DAC in doing that, rather than consideration of the source.

Last, but not least, there is the system used to listen, e.g how revealing it might be of changes at the source end, needs to be known if someone's assessment is to be meaningful (helpful if declared with the assessment, though otherwise at least in the poster's profile if listened at home). It also includes the listener's ears, but that is not something readily assessable and capable of reporting here!

*Take the Chord range: regarding galvanic isolation, the Hugo doesn't have it because, IIUC, it was designed to be able to connect to portable sources with limited USB power, GI requiring too much power from on the source. 2Qute, TT and Dave do have GI. As for RF filtration, the designer of the Chord DACs has made it clear that even his awareness of the susceptibility of DACs to the effects of RF interference superimposed on the data stream has been increasing all the while. As a consequence it would seem that as time has passed each Chord DAC released is likely to have better RF filtering than the last, and so be less susceptible to RF from the source.

+1, you can't expect it to perform properly if it is not properly optimised and set up.  You wouldn't throw your NDS on top of your preamp and expect it to deliver straight off. My 'next step was always going to be an NDX so I was shocked to find myself considering the MM route.  However, after a lot of tweaking there is no way I would go back to a streamer unless Naim brought out one without a built in dac, and even then I'd expect it to play Tidal as well as my MM does.  I have three sources, vinyl with an RP8, Apheta 2 and Naim Stageline, digital with MM, Audirvana +, jitterbug, Regen, Hugo and Tuner with a Nat05 XS and digital easily holds its own and can exceed the others on some recordings.  On some recordings you would do well to seperate the vinyl and the digital if played back to back.  Tidal on Audirvana is superb and very very close to local file play back.  At the end of the day it's all about the music, it doesn't especially matter to me what route others are using to get there as long as I do.

Yes very interesting reading others observations. A shame Simon didn't get on with the Mac as a source but then we all have our own preferences. 

Keeping it simple seems to be the best approach - minimising what the computer is doing seems quite critical to the final result and starting with a clean sheet (like the Micro Rendu, Melco etc etc) with hardware and software just dedicated to the single task of shifting data from disk / network to DAC (via whatever flavour digital output suits the DAC) makes this task easier. Much of the computer audio optimisation is based around stopping the computer doing all the processes it would normally do so having something designed to do the specific task in the first place should give better end results and remove a lot of the variables that come into the computer audio game. 

James

For myself I found a Mac Mini (late 2012/8MB RAM/2x1TB SSDs)  running Audirvana  - both optimised as in my last post - through a Gustard U12 USB/SPDIF isolator/converter into Hugo sounded better as a source than ND5XS. Without the Gustard it was very obviously inferior on USB, and similar or to the ND5XS using the MM's optical output (i.e  not bypassing the MM's sound card, but using optical to remove interference). That was listening through Musical Fidelity P270 amp and IMF RSPM speakers.

For clarification of an earlier post of mine here, the same MM and Audirvana setup is what I compared briefly with Melco N1A recently, through Dave, Bryston amp and PMC Fact 12 speakers. That was not using the Gustard because as a very recent Chord product Dave has very good RF filtering as well as GI. There was no obvious difference to me in that system - however it was only a brief comparison at the end of a session auditioning DACs, so I am not saying there was no difference, just if there were differences they were too small or subtle to be immediiately apparent, and certainly nothing to make me feel I have to rush out and change, but I may well audition fully at home some time, to then decide if it is a worthwhile upgrade or just something to keep in mind for if/when one day the MM dies.

Very quickly off topic, i bet those big IMFs are a lot of fun 

Back on topic...

james n posted:

Very quickly off topic, i bet those big IMFs are a lot of fun 

Yes indeed, and I will never forget their seeming perfection when I once used them outdoors... However I have since replaced with the PMC EB1i. which are better still (indoors - Indon't know about outside).

On the subject of optimising a Mac Mini for audio replay; for those with a pre-2010 model, the ones with the separate power supply, it's well worthwhile converting to a better power supply. I can recommend the Paul Hynes type. It's relatively expensive, but Paul offers a very good money back guarantee if you don't get on with it. I'd be very surprised if, once run in, you decided to send it back!