Does the transport to an nDAC make a difference? Oh - and other musings...

Oh...  go on Andy, put us all out of our agony...

Teehee...

In brief...

NDX bare vs streamer: streamer had brighter clearer presentation, bigger soundstage and was more controlled. NDX was more boxed in by comparison, but you could tell they are from the same stable.

upsampled audio vs straight: Upsampled seemed to push highs and lows more forward with a recessed mid compared to the straight through. My comment was it sounds processed.

Naim interconnects: Hi-line is more controlled, more relaxed and bigger sound. Doesn't push the sound into your face quite as much. Nice.

Now the biggie: computer->nDAC vs NDX->nDAC. I couldn't hear a difference - any time I thought I did it wasn't repeatable for me. James could hear the slightest of differences (perhaps a little clearer was how he put it) Having said that, (and without wishing to put words into James' mouth) neither of us believed anyone would spot the differences consistently if they weren't doing what we were doing (which was playing the track on both machines at the same time and swapping inputs on the DAC). Either were more than adequate as a transport and all the other equipment differences made a good order of magnitude greater difference (including the hi-line). 

I'd say we actually got bored of trying to hear any differences, settled on my streamer and spent an hour thrashing the DBLs with some music with the hi-line between the nDAC and 52. Have to say, despite the ridiculous price of the hi-line - it makes the system gel very nicely. I might very well be interested in a second hand one of these......

My summary then: yes and no (there's the get out of jail card right there ). It seems some are able to tell a difference but the difference is minuscule compared to all the other differences we heard today. If James were in an audition today, I think he'd have bought the streamer and pocketed the £3000 difference (in fact he said as much). 

I'm sure James will comment - not just about the tests but of his thoughts of the DBLs and xbmc as an integrated video/audio system.

Just a PS to the above.

The bare NDX vs streamer/nDAC, upsampled audio and interconnect tests only took a single track to compare. We both agreed on the description the other gave to the differences. They were there, easy to hear and we just moved on.

In the back to back of the two units as transports to the nDAC we must have listened for a good hour and when James expressed a preference he chose the NDX as his preferred source more often than not (I was switching inputs without him knowing which was which) which is how I've come to the conclusion that some can tell a difference. When he switched between sources for me, I couldn't tell a difference between the sources.

Edit: the reason we used switching to compare was that when we played a track in its entirety, it was too difficult to be even remotely consistent about what we were describing.

Hi Andy - it was great to finally hook up and get to hear the DBL's in a domestic setup.

Firstly the DBL. I can understand why Andy loves them. They deliver music with scale and a real punch. If the track demands bass, the DBLs deliver it but it's very clean and controlled - no overhang and it goes low with ease. Rage Against the Machine was a very visceral experience ! The other thing for such a large speaker, is that they are very discrete and with Andy's setup compliment the room and equipment perfectly without dominating the room. Impressive stuff. 

As Andy has mentioned above we were able to compare a number of combinations. NDX and nDAC differences were quite clear. The nDAC(+XPS) sounded more open and detailed than the NDX which sounded softer in comparison but i can see why the sonic signature of the final output stages in these two units polarises opinions somewhat. I thought the nDAC with Lavender cable could be a bit too 'in yer face'. With the Hi-Line in place, the sound became more relaxed without losing the detail. We both agreed this was a worthwhile improvement. 

Comparing the xbmc media server and the NDX was quite enlightening We compared using a USB stick and UPnP - the same tracks ripped by Andy using EAC. I found the NDX driving the nDAC to be slightly cleaner in the bass, but to be honest couldn't reliably pick one over the other. We both agreed it wasn't night and day and not worth worrying about. Comparing native to upsampled using the NDX as source , came down to us both preferring the native setting. 

One more thing worth mentioning is the xbmc setup Andy uses. What a great interface for music and movies. It's the first time i've come across it and it's a very neat and intuitive way of aggregating a number of formats into one easy to use interface. 

One thing that stood out for me, was the nDAC in Andys system was very source agnostic. The differences between the NDX and xbmc source were minimal which did surprise me but i heard it with my own ears and that's good enough for me.

As with all of these things. mileage will vary in different setups (that's my get out of jail card !) but i can concur with Andy's findings in his original thread. 

Overall a very enjoyable day and good to listen to some familiar tracks through a classic 52/135/DBL system. Adding the Hi-Line to the nDAC was (for me) the icing on the cake. 

James

Very enlightening and interesting write up from you both - many thanks, Guy 

Sorry, the post is lost on me:

when you say streamer, are you referring to the ndx+nDac combo?

No - a PC. I have an HTPC running xbmc that I use to stream audio to the nDAC. We delivered the same files to the nDAC using the PC and the NDX and couldn't reliably tell the difference between the two "transports".

Hi James...

Have to say I really enjoyed the day. I'm listening to stuff now I haven't listened to in ages wondering if the hi-line would make a difference!! Really glad you enjoyed the day too - and thank goodness that we shared pretty much the same conclusions. If we hadn't can you imagine the bun fight..... 

Just to say thanks - I appreciated you spending the 3 hours on the road and coming over and bringing the NDX with you. If there was one transport I wanted to compare against the PC it was this one. Job now done. 

Anyway, enough of the brown-nosing - I have a complaint! You forgot to mention the mesmerising visualisations xbmc puts out whilst playing music .

Great thread, great times....  Im glad you guys hooked up.

Andy,

Would you mind describing your HTPC a bit?  How are you getting digi out of there?  What are you using for mobo and proc etc...

For video are you using on-board (sandy bridge) video or do you use a card?  etc...

Thanks for taking the time to meet and chime in here.

Patrick

Hi Patrick,

I have 3 HTPCs here. 2 of which (including the one used today) are Intel Atom/Nvidia ION systems. Built for low power but that can still do 1080p decode. I'm all for the hardware doing the decode if possible.

The 2 Ion systems cost less than £250 each - and they both have 30G SSDs in to keep them quiet (ask James if he thought there was a PC in the room )

Hi Patrick.

Realised I hadn't answered your question... Audio out is through the onboard optical port.

Congrats Andy and James!   Has it really been two years?   Wow....

Am glad you guys finally made the time to get together, try some tests, and listen to some music.

Hook

Are these DIY HTPC's?  I am interested either way, but would prefer to look at a supported product.  My personal experience with DIY computers has yielded relatively low VFM.  

Thanks for taking the time to share with us.

Nick

is it to say that ndx is redundant?

to ndx it must have a nas/sever/pc/mac as transporter,maybe have to nadc with ndx

and ndac also have nas/pc/mac/cdp as transporter,ndac does not have a ndx to convert

people hardly can tell the difference betwen ndx and ndac as a converter.

maybe ndx is convenient to enjioy the music.

Ah yes - like being on acid 

James

Note - only caffeine was used in these experiments !

Ok8888, if you can't hear much of a difference netween NDX and NDAC, the chances are the rest of your system (amplification, speakers, ears) might be the limiting function. In which case if you want to develop your system performance, upgrade these first before the ndac.. By the way I am not sure how you upgrade ears...

Yes. I build all of my own PCs (except laptops) and run various O/S depending on the purpose of the machine, but I freely admit I'm not a "normal" user - there are 9 PCs in this household of 3 people - 5 of which I built myself. Having said that, the only requirement of the PC in use was some form of digital out - I didn't choose the components specifically because they had measured as excellent digital outputs and you could quite easily use any pre-built PC to run the software if it had a digital out.

What the DIY PC route did allow me to do was choose the components I wanted. The main processor is a low power processor, the disk is a silent SSD and I power it with a silent power supply in a specialist HTPC case. To stand next to it - you wouldn't know it's a PC as it doesn't look like one (it looks like a hi-fi component), it doesn't sound like one (it is essentially silent as it doesn't need a lot of cooling which was a high priority for me for something that is situated in the living room) and when you interact with it you don't know it's a PC (the user interface is driven by a normal remote control and it boots into a fully integrated media centre environment - you never see a traditional computer UI at all). 

I'm not 100% sure what you are saying here. 

If you want a single integrated streaming box with high quality analog outputs, the NDX delivers. Without having heard it, I would expect the UnitiQute to deliver high quality audio too. There was certainly an easy to spot difference between the NDX acting as a converter and the nDAC acting as a converter though. 

The primary reason for the visit can be traced back 2 years to a thread on here when I had just bought my nDAC and was blown away by the sound of it being driven by my media centre PC. It was a significant step up from a CDS1 that I had before and that was before I added the external power supply on the nDAC. In that (very long) thread, my argument was that transports didn't matter when using the nDAC as the way the nDAC is designed it effectively removes the effects of bad design in the transport. The thread polarised opinions - some argued that the quality of the transport matters whilst others sided with me. James offered to come over and see if it were my ears that were at fault or whether what I was saying held true.

2 years passed and James had coincidentally acquired an NDX which could be used as a high quality transport into the nDAC. I know that makes the assumption that high cost = high quality but that's another discussion and I only have low cost transports here. The result was that we couldn't really tell the difference between transports and - as James put it - the nDAC is source agnostic. I know James qualified it with "in my system" and yes, you could construct a transport that injected noise into the nDAC, but for a correctly setup transport I would expect the same results to be repeated in other peoples systems - there is nothing special about the transport driving my nDAC. There is also nothing special about the interconnect either - it's a 3 metre length of optical cable bought off e-bay for £3. 

In my mind (and I suspect James' too), the question has been resolved to my satisfaction. You can feed an nDAC with almost anything and, assuming the transport is bit perfect and you haven't royally messed it up electrically, it doesn't matter - it will sound the same. And that sound is pretty darned good! 

Which brings me back to my original question - posted those 2 years ago... Why is the nDAC so cheap? 

For completeness: the spec of my HTPC is as follows:

• Motherboard: Zotac IONITX mini ITX board with 1 PCI-e slot. Processor: Intel Atom 330, graphics: Nvidia ION, Memory: 2 GBytes with 512Mbytes shared memory between the processor and graphics.
• Power supply: Silverstone Nightjar ST40NF silent power supply (complete overkill, but I had it lying around). If I were building from scratch, I'd use something like a 120W pico-psu which would be much cheaper.
• Case: Antec Fusion Remote Black Media Center Case. Bought it as it has an LCD display but - to be honest - it isn't really that good.
• HDD: generic 30Gbyte SSD.
• Remote control: Gyration 2.4GHz RF controller (you don't need line of sight to use this). I have other htpc's using £10 generic MCE remotes though.
• Operating system: Linux (although you can use windows if you want)
• Software: xbmc

Although not relevant for the testing, this box also has a TeVii S480 dual-tuner satellite card in the PCI-e slot. I use these to record broadcast signals here using MythTV as the PVR software. These recordings then get added to the library so that xbmc can "see" them so can be played anywhere in the house that has a media centre (3 htpcs under TVs and on the laptops).

Hi Andy -

With no wish to draw this endless debate out any further, I mean after all, it has been over two years (!!), but I now feel compelled to point out that others, including me, have tried similar experiments, and have gotten very different results...

As you know, I used to be firmly in the source agnostic camp.  But then, a little more than 10 months ago, an NDX and a pair of Ovator 400's were delivered for me to demo long-term.  After burning in the speakers for a couple of weeks, I set up a blind test for Mrs. Hook, and then repeated it later on for one of her girlfriends.  I cut open, and set up some large cardboard boxes as screens (not that either of these ladies could tell a SuperCap from a SuperLine), and used a remote to switch inputs.   Both the NDX and my DIY PC (with its RME 9632 card) played the same song delivered via UPnP, and they were closely synced in time so that the switching was pretty seemless.  I also used a Radio Shack SPL meter to set the levels as close as humanly possible for the two inputs.  In other words, I made an effort to level the playing field.   And oh yeah, he rest of my system was a 252/SC/300.

Fast forward to the end of the tests, and Mrs. Hook picked the NDX *every* time.  Her friend picked the NDX 8 out of 10 times the first pass through, and 10 out of 10 times the second time through.   This was enough to convince me that, regardless of how subtle or personally insignificant these differences in sound quality may be, they do, in fact, exist.   My subsequent long-term sighted tests (yeah, I know, expectation bias) had the same results.   After using the NDX for a few days, I went back to the PC for an evening of listening, and it was less engaging, and less enjoyable.   So, after repeating this a few times, I bought the NDX.

Thinking back, there are only a couple of other differences in our setups worth pointing out.   First, we used a DAC with a 555PS.  IME, this was a very nice step up from the XPS-2.   And second, both my NDX and my PC were connected via BNC.   The NDX used a DC1, and the PC used a custom-built DB9-to-BNC conversion cable.

I think that there a lot of people who do not have the time or interest in configuring and technically supporting a DIY setup.  I think this aspect is significant, because one of the reasons why I was open at all to the NDX as a source was because of availability.  There were too many times when I would sit down to listen (almost always late at night, and after a long, stressful day at work), and instead would spend my night repairing what should have been a plug-and-play component.   By stark contrast, in the last 10 months, between the NDX up front, and the re-purposed PC in the closet running Asset, I have not had one single night of downtime (except for the hour or so when I upgraded the NDX's firmware and installed the 24/192 board).  The NDX, IME, is a 100% bullet-proof solution for streaming.  Expensive?  Sure.  But in my situation, and assuming this continues for a number of years, it has already been worth it.

Anyway, if there are folks out there looking for good quality sound on a budget, then SB Touch and/or DIY PC's (if you have the time and the skill and the energy) are great solutions that can usually come close to what you will hear with a Naim digital source.   But, IME and IMO, there is greater sound quality to be gained from a better source (and especially in more high-end setups).   I will not argue that the increase in sound quality alone can be justified from a VFM perspective.   But factor in the reliability, availability and long-term serviceability of a Naim front-end, and I believe a long-term TCO argument can be made.

From a TCO perspective, there are also a couple or other considerations worth mentioning:  first, the NDX can always be re-deployed as a very high-quality stand-alone analog source in a second system.  And second (assuming Naim continues to upgrade N-Stream and support the latest version on the NDX), then like most Naim components, the NDX should hold its value very well over time, and resell (or offer trade-in value) at a very good price.

Lastly, since those tests were run, we've had a few other upgrades:  552, Fraim and Audience ar6.  All brought improvements to sound quality.  My guess, should I ever get the energy or interest in re-running these tests, is that I would get the same results.  Actually, my expectation is that these improvements would result in the differences becoming even more magnified.

Anyway, the NDX works perfectly for me in my setup, and I would never consider going backwards to to a DIY solution.   Technical support just isn't as fun for me as it used to be when I was younger.

Hook

PS - Honestly, like I said, I have zero interest in prolonging the debate.  I would simply ask that we all continue to use IMO, IME or "in my setup" (and similar qualifiers) in our posts since your (and, it sounds like, maybe James's) experience was different than mine.  Just wanted to remind (especially new) forum members that you guys weren't the only guys to try this, and that others have gotten different results.

Hook, good write up and my personal expierience matches yours, other than I used Mrs SinS rather than Mrs Hook 

Hook,

If you had to use a SPL to set the levels between the two streamers they aren't producing the same bitstream. That's kind of obvious surely. Or are you not comparing them as pure transports to the nDAC?

TBH, if people are happy paying thousands for their streamers, I don't have a problem - there are FAR more productive ways of upgrading systems than spending there. 

BTW: I haven't had a day of downtime from my HTPC system for the last 2 years. Maintenance is - essentially - 0 unless I upgrade the software which is normally an hour or so every year or two.

Great to see you both got together .

At last.

Good to see you both had a fun day.

I wish you had tried some transports though?

Just to hear the diff.

I can pick holes in your testing/switching? But i wont ,life is to short..

At the end of the day you both had fun.

THATS WHAT ITS ALL ABOUT.IMO.

Regards,

Stu.

I'm going to make one final comment on this.

James was able to tell the difference between the players - whenever he expressed a preference, he correctly identified the NDX as the source (see post 4 in this thread where I mentioned it first). The difference was, however, so minimal that you couldn't put your finger on it. I certainly couldn't tell which was which - any subtle differences were meaningless to me. 

Whatever the differences are, when seen within the context of a complete system, spending money here is definitely NOT good value for money unless you are completely at the peak of everything else. And by peak of everything else I mean fully active with a complete 500 based system...

Whichever way you look at it, the transport delivers 1's and 0's to the DAC. That's its job. It does this over an interface that is serial in nature. If you are happy paying many thousands to eke out the last bit of jitter (as that's ALL it can be affecting the DAC - especially if you are connecting optically - I simply do not buy any interference through the power supply as this would be swamped by the crud that is on household power supplies anyway) from essentially a $0.20 piece of output circuitry, then so be it. And if jitter IS the answer, then there's other ways of getting low jitter SPDIF interface out from PCs. Certainly less than the many thousands people spend on different front ends.

Perhaps people should look at this the other way around. Assuming whole box cost is the overriding factor in output SPDIF design, then the nDAC is a remarkable piece of kit. It enables you to get close enough to perfection with a self-build low cost PC that it is extremely difficult to tell the PC apart from a £3000 piece of hi-fi kit. Surely that should be applauded. 

One final point. If jitter is the overriding cause of differences, then different units of the same product will sound different as the tolerances of the components in the output stage will be different leading to different jitter amounts. Has anyone reported or experienced this?

BTW: James: if you do decide to part with the hi-line - can I have first refusal please - you have my mail address if I'm not posting if you do decide to sell....