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

+1

The nDAC is a truly amazing product and must be considered phenomenal VFM, given its relatively modest price point. I wouldn't call it source agnostic though. A long time ago I demoed the CD5XS vs. CDX2.2 into the nDAC and could clearly tell the two apart, with the CDX2.2 sounding noticeably better all around, not night and day but better for sure. I have to say though that even a 320kbps Spotify stream from my MacBook Pro via a cheap €12 optical cable sounds pretty damn good and it's not far behind the same album ripped to AIFF and played back via UPnP and ND5XS/DC1 into the same nDAC/PS555/SN/HC2 system. I am loving it and what a great way to discover new music.

I, too, have one final thought on this topic and, IMO, it is the elephant in the room.

Actually, it is our rooms, or more specifically, the speaker/room interaction which dominates everything else in our audio chains.

I was just reading an article in the current issue of TAS about KEF, the current Blade speaker, the company and its history.  The part that caught my attention was how they had spent $150k USD on a quasi-anechoic chamber to facilitate listening.  Not talking about the anechoic chamber used for test and measurement -- this one was built strictly to allow critical listeners their best chance at hearing subtle differences, and comparing notes!

I think that the one given we can all agree on is that our listening rooms are vastly different. On a scale from hard bare surfaces to quasi-anechoic, I suspect that no two of our rooms would be rated the same.   And my point is that, if I claim to hear a difference between x and y in my room, it is quite possible that on hearing the same setup in someone else's room, those differences could all but disappear...or become even larger, who knows.

So, I have a suggestion that you guys may want to consider (or not, it really is no big deal either way).  What do you think would happen if Andy made the trek to James's home, HTPC in tow?   Would the results be identical?  Even better, what if you guys took all your gear up to Kent and....

Just a thought.  And thanks again guys.

Hook

Hey Hook,

I'd kind of agree with you - apart from the fact that all other tests we did provided large differences. Easily identifiable and easy to turn into adjectives describing what we heard. There was no difference in opinion - the differences were there and could be communicated.

If this were an issue, don't you think it strange that the only phenomena masked is the differences between transports? Also, don't you think it weird that I can hear the same differences between components James hears but that all of a sudden the room is the reason I can't hear differences in sources?

Darn it.. I wasn't going to reply 

You had no choice Andy...I forced you to do so through my expert use of an evil mind trick.  

I do think it is interesting that James was able to hear the differences and you were not.  It tells me one thing, and makes me want to speculate on some others.   The one thing that seems for sure is that the differences between the NDX and your HTPC were relatively small.  I know this because, if they were large, and you still said you could not hear them, then James would have said your hearing is poor....and he didn't.   So, onto the speculation.  James has listened to the NDX a lot and you haven't, so perhaps he has picked up on some very subtle things it does, and has learned to recognize certain cues.  Or, maybe expectation bias played a role.  You did not expect to hear a difference, and you succeeded.  Another evil mind trick! 

But even if these difference are absolutely miniscule, the fact remains that James was able to reliably distinguish the NDX from your HTPC, and you were there to see it.   So, unless you think he somehow cheated, you must now concede that these differences do, in fact, exist.  That, in itself, would be a big change from two years ago.  Recall, we started all of this with the assumption that these differences did not exist, and could not possibly exist, because of the DAC's architecture.  Well, fast forward, and we have now have moved past that assumption, and on to a completely different discussion.  And that is simply whether or not these differences are significant enough to matter.

Clearly they do not to you, and that is ok.  But they do matter to others.  You just saw Totemphile's post.  If you do a search, you can read about Richard Dane's CD player experiments (I seem to recall he prefered some old Meridian with a digital output).  Others, like myself, prefer the NDX.  And the list goes on...

You may think that many of us are very foolish for spending a lot of money to improve on an area that you believe offers only limited returns.  Heck, you may even be right.  I happen to think there are even better examples though.  Every listened to the difference between a Dynavector XV-1S and an XV-1T?   If yes, can you hear a $3800 USD difference?   I can't.  I mean, they are both ridiculously good, but my mind shuts down at the thought of spending $9250 USD on a freaking cart! It just does, and there is nothing I can do about it.

Thing is, I do agree with most of what you say.  The DAC is amazing and underpriced.  And with a relatively small spend, you can buy and/or build a front-end that sounds great.   Spending more may not be prudent, but diminishing returns is obviously not enough to stop many of us from seeking an optimal solution in this particular area of our audio setups.  Call it source-first-mania!  

I also think overspending is simply one of those things about this hobby that we all have all been guilty of at one time or another.  My own personal belief is that so long as we are frugal In other areas of our lives so as to compensate, and so long as we are truly talking about discretionary spend (and not the rent money), then there really isn't a whole lot of harm that can come from splurging on a sexy front-end for the DAC.

ATB.

Hook

Hook,

I doubt James was listening for cues that he was used to - he doesn't normally play the NDX through the nDAC - it's the nDAC that converts the digits to analogue so it's that whose sonic signature you are listening to. All the transport is doing is providing the bits and the only difference that can be there is in the digital out. 

We sat there for an hour trying to rationalise what we were hearing (or weren't which ever way you look at it). Much of the time was shoulder shrugging and head shaking as the differences James heard really were minor. I've also tried in the past playing direct from USB memory stick and I was unable to tell a difference there either and that's supposed to be "the best" method of feeding data into the DAC. My conclusion - based on a sample point of 4 (I've also tested a £60 Sony CD player) is that the difference between "best" transport and "worst" transport is fairly minor as I can hear changes such as different cables in the system but I can't detect the difference between a cheap CD player and the best transport (USB stick - or is that now not considered the best).

As for testing CD players. Nope. Totally uninterested in doing that - the silver discs are in boxes in the garage now - and that is where they will stay. Convenient sound quality is where it's at these days. 

I also think the thing that people are doing is equating expense with better SPDIF out. How can a £200 PC do what my £5k box is doing - it MUST be better as it is more expensive. Lets take this another way. There are a number of SPDIF addons for PCs that are designed for audiophile output. They will do the same job as a better designed streamer output - they will reduce the jitter in the outgoing stream - they will sound different. And if I'd seen a tangible benefit from the NDX - I'd be ordering something along that line myself.

Just to throw another variable into the mix - I suppose it could be my PC that's not providing bit-perfect. I haven't actually physically checked that. Or that optical inputs sound slightly different to electrical ones... 

BTW: one other thing that I haven't commented on - usability. I don't think the Naim streamers have a particularly user friendly interface. For example, I don't have a smartphone nor an iPad so control would be limited and I would baulk at paying another £400 for a remote control! Also, from watching James control his NDX, the interface looks limited. TBH, that alone for me nullifies any sonic differences. 

As to a change of position - no, not really. If you look back over the posts towards the end of when I posted in 2010, you will see a softening of my position. I have actually believed that people can hear a difference for quite some time. You can't prove absolutes unless you have the right measuring equipment and can mathematically prove something doesn't exist. There will always be something new around the corner that changes things. I do think James' phrase source agnostic is a good one. Agnostic implies an ability to believe in something. The nDAC basically levels the playing field, but can be influenced a little one way or the other.

To say feeding the nDAC with a cheap player is a waste is nonsense - it makes music as was evidenced when James was here. To say that the only way of getting good sound is by spending thousands on a streaming product is also nonsense. The advice given here really should bear that in mind for people coming into streaming audio. I'd personally rather spend £2.3k on a PC streamer and nDAC than £3k on an NDX for example. 

Hi Andy -

We are on the same page, except the use of the term "source agnostic".

According to the online Free Dictionary (the first one I hit via a Google search), agnostic is...

"...One who believes that it is impossible to know whether there is a God.

So, the translation from religion to our discussion would be:

1) Source theist - one who believes that digital sources for the Naim DAC sound different.

2) Source atheist - one who believes that digital sources for the Naim DAC cannot sound different (because, after all, that's what the white paper implies...).

That was easy.  I think we both agree those two terms are now very clearly defined, right?

So, can we say the same for "source agnostic"?

First, there is:  "...One who believes that it is impossible to know whether or not a digital source for Naim DAC can sound different."  Nope, does not translate.  We know they sound different.

Second, there is: "...One who is skeptical about the existence of differences among sources for the Naim DAC, but does not profess true source atheism."  Or, more simply, "One who is doubtful or noncommittal about whether sources for the DAC can sound different."   Again, neither definition works because both cast doubt on whether the differences exist.  They do, we concur that they do.

This quote from the agnostic Wikipedia page is even more applicable here:

"...In the strict sense...agnosticism is the view that human reason is incapable of providing sufficient rational grounds to justify the belief that deities either do or do not exist...".

Again, not buying it.  The "unknowable" connotation totally bugs me.  We have plenty of attributable causes.  We know about bit imperfect and jitter and RFI and so on.   And not only do we have all of the experiential evidence, but we also have Simon-in-Suffolk putting the NDX and other digital sources on a Spectrum Analyzer, and seeing the differences on the screen.

So, if you were you ask me what some some better terms would be, I would suggest:

1) source "perfectionist"  -- one who believes that optimizing the digital source for the Naim DAC is something worth pursuing, and

2) source "rationalist"  -- one who believes that it is better to seek out the "sweet spot" for the DAC's digital source. 

According to the Free Dictionary, a rationalist believes...

"...that the exercise of reason, rather than experience, authority, or spiritual revelation, provides the primary basis for knowledge."

I think that definition translates to here very well.  You are a "source rationalist" and I am a "source perfectionist" when it comes to the most under priced product Naim has ever made....the DAC.

Other than that, I think I pretty much agree with everything else you said in your post.

Hook

PS - Except...when did I ever say that "feeding the nDAC with a cheap player is a waste"?  I never said it, nor did I even imply it.  In fact, I have gone out of way to say that it makes sense for most.

PS - Get an iPad, join the cult.  I never thought I would want one or like one, but now I am totally hooked.  It is particularly ideal if you listen to music in a darkened room (as I do).   I do agree that N-Stream has plenty of room to grow.  Right now, on my network and with my new NAS, it is way more than adequate.  N-Stream is fast, functional and bullet-proof.  And we have been told that Naim is developing some serious enhancements.  So I am betting that, in the end, N-Stream will be more than competitive with anything else out there.  And, finally, I suspect it will continue to be the only thing out there that offers system automation -- volume control and input switching.  Dude, the only criticism I have of your setup is all those outdated remote controls on the table! 

I'll just chip in here...

I'd expected to go to Andy's, have a listen to his media centre driving the nDAC, we'd plug in the NDX and marvel at the difference in sound quality -  but we didn't and i must admit to be more than surprised. Although Andy's system was new to me, i was happy that we could both hear differences between the analogue outputs of the NDX and nDAC and the difference between the Hi-Line and Lavender cable between the nDAC and 52 (and differences between NDX digital native and upsampled) If we'd heard no differences here then i think we'd have called it a day there and gone down the pub !

Whilst i could 'pick out' the NDX from the media centre, it wasn't a big difference and as Andy said most of the time it was difficult to tell each digital source apart. I still think (as i hypothesised in the previous thread) that the nDAC is not totally source immune. Given a bit perfect source, and a large ram buffer which could hold a track at a time to enable the input clock to be turned off, then the nDAC could be come totally isolated from any influence of the digital source. In our testing difference between the sources were very small. We could have chucked a CD transport into the mix but this introduces another variable - read errors vary each time the CD is read which could influence the final result. 

The thing with all of this is there are so many variables with different systems, different environments (from an EMI point of view) and different system layouts and wiring that i'm sure mileage varies as the other posters on this thread confirm. Certainly we didn't find any advantage on using the NDX as a source with nDAC which was a good result for Andy. His particular setup with xbmc is a rather nifty way of managing music and movies and via the nDAC and DBL's the end result is rather good 

James

PS : Andy - you didn't see NDX control at its best. Via nStream and UPnP on the iPad its pretty good You're always welcome up to Surrey too. 

I've always stayed away from religion on forums but...

I kinda like your terms, but (and there's always a but) it doesn't imply the scale of the experience. If people hear differences, all well and good, but in context, they are minor differences compared to moving on up the amp chain (for example). Here, each step is clear, the benefit doesn't come from extracting the last amount of detail out of the system. Let me quantify that in terms I can get my head around on a product line I know well (having auditioned much of it 10-15 years ago) - the Olive NAP line. I have owned a 180, 250 and now 2x135s. I have also heard the 140 extensively. I'd rate them (in terms of performance with a completely arbitrary idea of 100%)

140: 50%

180: 75%

250: 90%

135: 95%

I'm not interested in what figures anyone else would put on this, just that there are clear and significant benefits going from one to the other. The steps are large.

I've come from CDS1 through streaming via a £200 DAC to the nDAC. There I'd put the figures:

CDS1: 75%

PC+£200 DAC: 85%

PC+nDAC: 92%

PC+nDAC+CDPS: 95%

NDX+nDAC+CDPS: 95%

I suspect that James would slot the last one as 95.2% (or some similarly small percentage - James if you're still reading this, perhaps you could chime in....). Perhaps I can never get past the 95%, but the lift in performance is minor compared to the others. 

BTW, I never suggested you wouldn't seed an nDAC with a cheap player, more the disdain given by others to it. For example in the last few days in the couple of threads I've read I've seen quotes like this one:

 The ndac is really too dear to be used with the sbt. Others might disagree, but i think its wasted on it.

The implication is that the scale I've written above would be more like this:

PC+£200 DAC: 85%

PC+nDAC: 89%

PC+nDAC+CDPS: 90%

NDX+nDAC+CDPS: 95%

That replacing the PC with something costing 10x as much will suddenly give a huge jump in performance. It just isn't the case. It is this very reason I have problems with the word rationalist....

As to iPads. Just don't do it for me. I (nearly) always have the TV on when listening to music as I love the psychedelic patterns it generates so my UI is there. It's a bit like oil projection lamps from the late 60's/early 70's. If I don't the system has an LCD monitor I can use. There is also an app for iphones/ipads or android to control xbmc, but that's a circular argument. Auto-switching - why would you need that - I only have one source! Volume control - I'm still on an olive 52 - I need a remote! Browsing in front of the TV - certainly sir - F12 on the keyboard (I also have an RF keyboard plugged into the back) and the browser pops up (the HTPC is a PC you know ). Much bigger screen and much easier to type. That and the fact everyone is used to a remote control and the UI is exactly the same on the 3 xbmcs I have in the house - it's a simple remote control that has up/down/left/right and the traditional VCR keys. The reason there's a number of remotes is the TV needs one, the HD broadcast box needs one and the HTPC needs one. 

To see the interface I'm using, go to: http://xbmc.org/download/ and pick the one of your choice. You can install it on your PC and it will run as a standard application - all you need to do is point it at your uPnP server or use windows shares (like I do). It integrates ALL media into a single environment - audio, video (we watch way more video than listen to music), has plugins, skins, libraries linked to the internet, local weather forecasts internet radio, plays 99.9% of known formats etc, etc, etc....

The network audio streamer is so 2010... HTPCs is where it's at dude. If you can hear a difference in transports into the nDAC, you should be looking at external USB SPDIF interfaces. They should provide the same benefits to you as an overpriced green box but with the benefits of a true media centre underneath (said in the friendliest of ways )

Thanks James.  Everything you said makes sense to me.  And as far as that "perfect source" goes, at this point, I am guessing that the benchmark is still, probably, a well-made USB flash drive (with no LED).

Again, I appreciate you and Andy getting together, and putting some closure on this topic.  Seems pretty clear that you guys correctly identified, in ascending order, the things that could make the most difference to sound quality in Andy's setup.

Cheers.

Hook

See... whilst I've been writing War and Peace, James has summarised much more succinctly!

As to control, yes, I did wonder if that was it "at it's best".

Surrey... would love to... shall we plan for sometime in 2014 

...and taken as such. 

Hook (out)

Andy,

the squeezebox touch is not  your HTPC, so you wouldn't really know. With a nicely powered PC, I'm sure you can get to 98% of the NDX + PSU. When I heard the NDX + 555PSU into a XS based system the difference between the NDX with and without the PSU was actually laugable. (specially into a XS series component). So much for source first.

In your terms, here what I was suggesting:

Perfect Transport + nDac = 100%

Perfect Transport + Rega Dac = 80%

SBT + nDac = .8 * 1 = 80%

SBT + rega Dac = 70%

So now the price differential between the rega dac and the naim dac becomes more prohibitive: I don't want to pay that much more to only improve 10%. While really the dacs IMHO are both very close and the difference to my ears is just the house sound.

If I had a better transport, I might be tempted to pay the price difference to get to my '100%'.

Does the transport make a difference: In my view, night and day?

Having said that I think I know what's going on (and why this discussion in my point of view is totally moot).

Your HTPC is a damn good machine.

again,

it's a damn good machine.

There is no difference between the NDX and your HTPC configured just right.

All this proves is that in your system, it doesn't matter whether you use the NDX or the HTPC. I would not generalize this to all the transports in the world:

Here were my findings and totally repeatable:

You take  a squeezebox with no software mods and the stock power supply, and the feed the dac. It sucks big time.

You take the same squeezebox, do a linear psu on it, do software mods (turn wireless off, turn screen off, turn a bunch of useless apps off, give it pcm / wav data) so it's doing nothing but converting lpcm data from the network into sp/dif interface data and it's a cracking machine. For that price point, a no brainer.

My moment of truth, when I realized that the transport makes a big difference (but this was before I had put a linear psu on the sbtouch) was when I hooked the dac to a laptop. I loved what it did. No chance it could get better. Smile on my face etc. etc. 

Coming back to your discussion: My take on this entire thing is that your HTPC is just configured right.

bottom line: your HTPC is configured just right. don't monkey around with it, and don't have this lingering suspicion in your mind that the NDX is better. The NDX (at least it's transport part), is just a network card with a dac stage with some fancy buffering/upscaling thrown in, which as an aside,m isn't very effective, because otherwise transports would really not make any difference.

Now I could spend  a few weekends coming up with this perfect transport, but then my employer would hold the IPs on it. I actualy delete the section above where I had outlined my plan due to IPc concerns.

Going to grab my coffee now. I'm working today.

But the point I was making was that the difference between the best and worst transport on the nDAC is a percentage point or two at most. If it can do bit perfect output, and the nDAC can lock onto it, you're at 99% of what the nDAC can do with the "perfect" transport. At least that's my experience (and I have tried USB sticks into the nDAC too and still not been able to tell the difference).

Because of the way the nDAC works, it makes otherwise lowly sources pretty much the same as high end sources. It really is a game changer..... 

My discussion ONLY relates to the Naim DAC - the way it is designed, it will - and does - pretty much level the playing field. We have proven that. 

All bets are off for other DACs and reading between the lines (and looking at your profile and a few posts posts), you've done your experiments with a Rega DAC. Specifically I am NOT saying all transports sound the same on a given DAC.

BTW, my HTPC is just a computer with no real configuration other than making sure I shove the data directly out of the digital port. When it's doing this, it's CPU and graphics card are going hell for leather displaying visualisations. I get the same results if I feed it data from CDs from a £60 player or directly through the USB port from a memory stick, or, for me, feeding the nDAC with an NDX. 

+1 

It is the particular claim of the nDAC that "The data entering the downstream digital filtering and DAC chips is then completely isolated from the incoming S/PDIF jitter. " and "The Naim DAC suppresses S/PDIF noise through electrical isolation of its DSP front end from the digital to analogue converter and analogue circuits".

So if equivalent bit-correct sources do in fact sound different, it follows that the nDAC has failed to meet its declared design goals.

You can't have it both ways. Either it does what it says on the tin, and so is source neutral, or it doesn't perform as intended.

From the discussions and tests I've done, I'd say it nearly reaches its design goals.

The fact that people can reliably pick out transports (I've seen it with my own eyes) means it doesn't quite reach it. The fact it can make a PC sound essentially the same as an NDX means it gets pretty close....

I've done the same experiments with the nDac as with the rega dac.

Did you try the nDac with any 'bad' transports? I could tell there was a big difference between the sbtouch, in all it's modded glory and the CD5XS. The sbtouch on it's end claims to be a very low jitter transport. The CD5XS being a lot better, even into my NaitXS.

I'm just saying that the HTPC is very good transport, and hence you don't hear the difference.

If the "Lock" light comes on on the nDAC for the SBTouch then the jitter performance of the Touch is as good as it needs to be. The nDAC can buffer it just the same as any other transport which it can lock onto. That's the whole point of a buffer. I would argue that there should only be two classes of transport as far as the nDAC is concerned; those which are good enough for it to lock onto, and those that aren't. 

I've tried a £60 Sony CD player, a £200 HTPC, a USB memory stick and NDX. I can't tell them apart (BTW, I forgot to say that my HTPC is also doing the FLAC decoding as everything is in FLAC here as well as running the UI and system).

My mate powers his nDAC/XPS (into 252/250/DBLs - yes, we have very similar systems ) with his PC running Windows. The coax is provided by a cheap USB converter (his mobo doesn't have native out) and linked to the nDAC via cheap coax with a Y splitter so he can send the digital out to two sinks - the nDAC/XPS and a Naim NV2 as he plays films and has a different audio path when doing surround. Yes, you read that right, an RCA Y splitter costing £2 from e-bay.

He back-to-backed this against a CDS3 which he subsequently sold as the PC/nDAC was at least as good but the PC/nDAC is much more convenient. He can't be a...d to do any further comparisons - the sound is as good as he had from the CDS3 and he's happy.

If the source is bit perfect, and as likesmusic said - the nDAC locks onto it, it will fundamentally sound the same - give or take, as has been shown by numerous people hearing differences. Those differences are fairly minor in the grand scheme of things though as I for one can't hear them. If they aren't small something, somewhere, is broken.

If you say so. Let me see if I can get a hold of a nDac and take it for another whirl.

What I would like though, is for someone from NAIM to explain the difference between the CD5XS and the CDX2 used as a transport, as to why it would make a difference. Also the saturn cd player used as a transport.

And if it does make a difference, how that does not contradict the spdif locking mechanism that's in place and the buffering that's done.

I doubt it's related to error correction alone.

Dealers say it makes an audible difference. Are they just saying it to pad the bottom line? (I doubt that).

If you read the tin again you''ll find the Naim claim to eliminate jitter caused by the S/PDIF interface. They are quite specific about this point, it is quite clear they are not claiming to eliminate jitter/noise emanating from the transport.

With regards to electrical isolation of the DAC and analogue circuits, Naim claim to SIGNIFICANTLY REDUCE this. If you significantly reduce a massive amount of noise you end up with more noise than if you significantly reduce a small amount of noise

I wondered when you'd turn up 

You can't distinguish between transport jitter and SPDIF interface jitter. We've had this conversation before - several times - perhaps this time you'll get it. Taking an optical interconnect as it is electrically isolated... All noise from upstream gets translated into jitter at the S/PDIF interface - there is no other means of transmitting this noise. The DAC therefore sees jitter. It doesn't see transport jitter and S/PDIF jitter - it just sees jitter. Much like if it's raining and someone turns a hosepipe on, you don't see hosepipe water and rain water - you just get wet - it's still water! So... jitter is jitter and is (nearly, but not quite completely) eliminated.

Putting it politely, your second paragraph is total nonsense. 

If you'd have kept to your " and one final comment" statement I wouldn't have. you should have left it at that. Remember the size of hole you dug last time.