New NDS Streamer

Am I glad I am booked into the Grand for the Show.  Can't wait to see what you will be revealing to us. 

Unfortunately I won't be there (some of us have to do some work here you know)...

Phil

Originally Posted by AMA:I just wish someone would have had a little word in my ear before I bought the XP5 XS 

Allen, we have to be tolerant to this.

People who buy Naim today supply the funds for the new Naim tomorrow.

If every one stops buying the old gear there will be NO future gears.

I invested in Naim because I like the team and I want them to develop new products.

So I'm not going to be surprised to see that tomorrow they release the product which will make my system outdated -- this is what I was investing for.

Naim can not stop the sales for one year because they know the new gear is coming.

It will ruin the business -- who's gonna win then?

In fact I bet your old Mac/INT202 combo would match the digital output quality of NDX/XP5 at far less price. But in case you don't feel complete without a dedicated streamer  -- when did you say you scheduled KDS/1 audition? 

Hmmm -- I'm puzzled... Possibly, Naim has developed a better S/PDIF generator 

a.... and a better DAC

Patrick, INT202 which Allen used to own before is essentially a Firewire->S/PDIF converter -- so no DAC inside. It features the built-in transparency check which secures the bit-perfect reading of data from HDD to the S/PDIF generator. But the quality of S/PDIF generator is the main factor -- it's hard (if possible at all) to make it flawless and it's the major source of output jitter. Most probably Naim made it a bit better than off-the-shelf solutions.

I'm not so sure Allen...

You should have checked the led short-circuiting then  

If the data is properly copied to INT202 buffer there is no way it may sound different depending on the player which copied the data.

I would rather distrust the transparency checking from Weiss rather than believe that a bit-perfect copy can remember its origin. I'm sorry. 

Allen, I trust your ears and that you hear a difference. The question is why does it happen.

Before we go further may I ask you question: if I take a bit-perfect copy of CD (for example using EAC) and put a clone CD in CDS3 -- will it sound any different from the original?

If you admit it may sound different then I have no arguments to continue and I suggest to stop here.

If you believe the sound will be the same then how is it different from copying the data into INT202 buffer (unless the transparency check in INT202 is not true implemented or was not ON during the tests).

Hi Allen - I must admit since in my opinion I have effectrively perfected the playback performance (until I hear something better of course ) I have given up further tweaking, leave it all alone and enjoy the music and drama plays. Its the first time in my 'hifi' pursuit - which started in '83 as a kid - that I have not wanted to upgrade. Yes curious about other machines, but no irritating itch to upgrade...

AMA / Allen, I  suspect you are hearing different voicing between the Linn and the Naim line. As you might have seen me write earlier, DACs are subjective, so it could be one manufacturers implementation and compromise decisions suites you as opposed to another. There is no right and wrong just shades of differentness.

Simon

So nobody expects the NDS to be similar with the DSM series?? Pre-amp DAC

Fred

Allen, I'm sorry if I was not accurate in my replies.

You asked me on why three SW sound different and I actually answered this question (because the data transfer from HDD to INT202 was not bit-perfect despite what Weiss is indicating with LED).

You also asked me on KDS vs nDAC/555PS -- but I didn't get this point clearly. I mean if you ask me why two different streamers sound different than it's obviously an easy answer -- mainly because they have different DAC sections.

Also, there was nothing wrong about Noogle -- he asked a question and got answers from several people, including me. He didn't like any of them -- but he's got a full right for this. 

I don't expect it will feature a built-in preamp. I think NDS is rival to KDS. Successful? We'll see.

Allen, I'm afraid there is a kind of mess in terminology. Let me classify the problems.

Problem 1. Different SW sound different through INT202. The reason is that different SW players and INT202 firewire driver do not copy the HDD data bit-perfect to the INT202 memory. At least they can not do it in run-time. I believe this is down to the drivers and OS performance. In other words some (or all) of your SW driver setups were not reading the FLACs in a bit-perfect way or possibly OS was interfering with transfer process.  I'm sure the dedicated ethernet streamers don't have this problem and always output a bit-perfect (but jittery!!!) bitstream.

Problem 2. Different streamers sound different through re-clocking DAC, for example nDAC (that was the question by Noogle btw). I think that high quality streamers like Linn ADS or NDX always output a  bit-perfect content -- an exact replica of the HDD file. Most probably Foobar/WASAPI/INT202 is also a bit-perfect streamer. The reason why they still sound different through re-clocking DAC is that quality of S/PDIF generator is different for different players and output bitstream is not good enough for nDAC to decode all the bits so nDAC start flipping the bits at the entrance and this creates a certain sonic footprint on each and every streamer. Besides the data transmission through the cable can be further contaminated by the ground and RF issues or multiple barrier reflections which further complicates decoding the datas at nDAC entry.

Problem 3. Different streamers sound differently through analogue output (say NDX vs KDS). The reason is mostly down to different DAC sections.  Even if the bits are perfectly delivered to D/A converter they still can be converted to analogue in a very different way -- which is what all D/A conversion buzz is all about.

All three problems have different origin and different ground -- and I published my thoughts on this matter. Some may not like this explanation, the others 100% agree.

Most of people usually don't use technical terms for objection. For example, Noogle said: "This is a theory without an evidence". Fair enough -- but at the same time he can NOT bring any evidence that a chunk of bits recorded by nDAC and stored in the memory is the EXACT copy of the original FLAC data. So that everyone stays with his own opinion.

It's pretty straightforward to set up a PC or Mac for bit-perfect operation.  There is a website aimed at Computer Audiophiles which has detailed instructions.

SPDIF is a simple and robust digital interface.  I doubt Naim would use it so liberally if it were as fragile as you suggest.

Agreed.  Different analogue circuitry, different digital filters, etc. etc.

Imho -- this is not true at all. All S/PDIF generators are different quality. The clocks feeding them are different quality. The PSs supplying a power for clock and generator are all different. The performance of various S/PDIF output modules are very different. Hi-end manufacturers (like ANUK, dCS, Linn, McIntosh, Mark Levinson, Waida etc.) spent YEARs in research trying to establish a flawless S/PDIF connection between transport and DAC -- and mostly fail. The general concept today is to use I2S or ethernet streaming to achieve almost flawless data transmission towards the DAC section.

As far as I know Naim is using a very specific custom-designed S/PDIF generator in CDX2.2 and CD5XS -- which is still not perfect to supply nDAC with 100% guarantee decodable btitstream. The off-the-shelf S/PDIF generators are even worse.

Mario - How did you come to the conclusion its vibration. It could just be your system is susceptible to noise generated by the switch and moving the switch around gives you this variation in system performance ?

AMA - Different SW sound different through INT202. The reason is that different SW players and INT202 firewire driver do not copy the HDD data bit-perfect to the INT202 memory. At least they can not do it in run-time. I believe this is down to the drivers and OS performance. In other words some (or all) of your SW driver setups were not reading the FLACs in a bit-perfect way or possibly OS was interfering with transfer process.

How did you prove this. Did you discuss your findings with Weiss. I'm sure they would be interested to hear that you've found a bug in their bit transparency checking (although it worked ok for me and was easy to test - just reduce the volume in iTunes) ?

James

Sorry to hijack the discussion.

I have it on good authority that:

The NDS will not be "launched" at Bristol. There will follow a dealer launch event or roadshow of dealer pre launch events. So looks like Spring/Summer.

It incorporates a DAC which is musically superior (it is claimed) to the nDAC.

It will have a dig out although for the present there is no Naim component worth plugging it into.

It will (if an extra Burndy is purchased) use both rails of the 555PS.

It will also be compatible with the 5SX PS  (or whatever it's called - I've lost count, you know, the new one) and the XPS2.

Should this information pan out - as I believe it will because I would not waste my and your time posting wild speculation - and it sounds good- as I believe it should - this little piggy is heading for a shit fest.

In US vernacular a "shit fest" is not good at all.  Maybe UK vernacular is 180 degrees different?

I believe the etymology of the phrase relates to sayings such as "Happy as a pig in sh*t" as opposed to the opposing phrases such as "brewing up a sh*tstorm"...

I'm taking it as meaning he'd be deliriously happy to the point of not knowing which way to turn first...

Phil

Indeed. Although I suspect I may be facing the speakers.

Guys, engineers at Weiss are knowledgeable people and I'm sure they know what I'm talking about.

If bitstream is delivered to INT202 box in a perfect way there is no chance it would sound different through its S/PDIF output. This is the same as to clone a CD and then say that a copy sounds different from original. Not that I'm too arrogant with my Ph.D. -- but being an open-minded person I did these experiments in the past (many times actually -- and with different people) and I could not find a difference -- which is in-line with my knowledge of digital audio.

If two SW sound different through INT202->S/PDIF-> nDAC then the only reason is that the bit sequence is not securely reaching the INT202 output -- most probably because of the driver/OS issues and heavy workload on the device driver (could be Windows drivers do unnecessary pre-processing or PC Firewire output flips the bits at high bitrates or INT202 Firewire input flips the bits at high bitrates or INT202 memory buffer flip the bits at high bitrates -- could be a sum of all of these).  

To understand why the device driver (Firewire or USB) may flip the bits  I give you an easy example. When  you copy a file from one folder on your PC to another folder on the same PC you always get a bit-perfect copy of the original file. Here comes a very important question: do you think that each and every bit was copied at the first run? No single mistake during copy? No second trials to get copy done to match the checksum? No third trials? If you believe that all millions of bits are copied just in one go -- you are WAY from the reality of what is exactly going on inside the modern computers. If you do understand that copying the bits is substantially non-stable procedure than you can easily explain why audio bitstream can randomly fail during the run-time transmission.

To make things worse I can tell you that bit-flip is not only an HDD copying issue. Modern CPU make millions mistakes every second on the simplest procedure of data summation -- and built-in controllers and excessive-coding secure the operation is done right -- if not at the first trial, than at second, or at third etc... Not many people understand that because they know that Windows Calculator will always give you 2 x 2 = 4. But nobody knows how many EXACT nanoseconds this will take. This is very different from so-called "real-time operational systems" which can not do the check-ups and second trials and are scheduled to perform each and every bit-operation at a fixed time slot.

Allen, the transmission was NOT bit-perfect and this is WHY you heard different SW sound different.

Try to re-evaluate the leads on the "bit-perfect" transmission -- you will most probably find they are speculative.

I only want to share my view on this -- though I see that you are not comfortable with this  

I'm absolutely OK if you disagree. This is a good thing about forum.

Possibly I'm wrong and later on I shall find the way to understand it and get in piece with my ego 

I promise -- if I find I'm wrong I will surely publish this on the forum. 

I won't be oinking until I'be heard it at home. However, assuming my source isn't messing with my head (unlikely), Naim seems about to deliver something that could have been designed and built just for me. Or rather us. Or lots of people!

In nine days time I'll know if the physical spec is in order. Then comes the fun part.

Anyway, back to the real discussion....