Melco N1Z

Posted by: Graham Clarke on 06 June 2015

This product doesn't work.  Or does it?  Or should it? 

 

If you fall into the "it's all 1s and 0s and changing network components won't make a difference" camp then stop reading here as you'll disagree with my conclusions...

 

OK.  Rewind.  In a couple of recent posts, Norman @ UHES has hinted at a new product that has impressed them.  Forum rules prevent him from starting threads promoting or discussing non-Naim products.  In reality, this product isn't a direct competitor to a Naim product, it actually complements their existing range.

 

Having discussed the Melco N1Z with Norman and Ian I thought I should mosey down to Stirling House for a listen.  If you've not heard of the Melco a reasonable description would be to call it a high end audio grade NAS drive.  That does sell it a little short though as simplicity through ease of set up was also a key design consideration and it's far better finished than the standard plasticky NAS drive offerings available.

 

I won't go into huge detail about the specs of the unit (check out the web site), but there are two models, a hard disk based unit; N1A; (4TB total storage) at under £2K and an SSD unit; N1Z; (1TB total storage) with a host of other enhancements such as separate power supplies for LAN and other components.  N1Z is just over £6K and it was this model we auditioned.

 

N1Z has two Ethernet connectors (ditto N1A), one connecting to a switch, or presumably directly to a separate NAS drive and another which connects to the streamer.  We used NDS/555.  Amplification was Statement preamp and power amps into Ovator S600s.  Not exactly a shabby system

 

The SSD unit has two 512GB SSDs, so you can either mirror them for fault tolerance or span the volume across both drives for the full 1TB.  Existing streamer users will already have NAS storage meaning that fault tolerance on the Melco isn't a strong requirement: if a drive goes down then replace and restore data from your existing infrastructure.  For instance I use a 4 bay Synology unit with 12TB available storage via RAID (4x 6TB drives) which is also backed up to two separate 2 bay Synology units.

 

New users could use the Melco without any other NAS although in that set up they should use the RAID mirroring which will half the total capacity.  512GB does seem a little low to me.  Melco are owned by Buffalo Technology who make storage products, so I wouldn't be surprised if in 1-2 years' time the capacity increases.  We'll have to wait and see whether upgrade options will be provided to existing customers.  Also, with one leg in the IT industry I have slight concerns about warranties and repairs.  If a six year old Naim product breaks, it is repairable.  I have a niggling feeling that the answer to a broken six year old Melco will be "buy a new one" as that is what you would expect to hear about a laptop, desktop or even server product.  That's expensive at £6K...

 

The firmware of the SSDs has been modified from standard.  Due to the different way SSDs store data compared to HDDs they perform housekeeping tasks in the background due to write/read/delete cycles.  Apparently this creates noise in the system and the firmware changes on the Melco SSDs minimise the tasks given that the majority of accesses will be read, not write.  Hence they are labelled "Audio grade SSDs".  I couldn't find any evidence that Buffalo make HDDs or SSDs, so I expect they use OEM drives with the aforementioned firmware changes.

 

I have more than 1TB of ripped CDs so what is nice is that the Melco units can be used as a pass through, meaning that music stored on an existing NAS drive can still be accessed.  In this situation, consider the Melco more like a switch.  Apparently when running in this manner the SQ is still better than direct from a switch due to the noise isolation techniques used within it, think of it passing through a quiet room.  So in use I would put my most frequently used music directly on the Melco (another copy still on NAS) and the remainder would be NAS only.

 

We did an A/B comparison between playback direct from the original NAS and then directly from the Melco.  We didn't test the pass through option.

 

I won't mislead by saying this was a long, intensive comparison, likely it was 45-60 minutes total play time.  However that was far more than was needed.  After switching from NAS to Melco, playing the same track of course, the improvement was apparent after about five seconds!  We mainly used two tracks, Crystallize by Lindsey Stirling and Strong by London Grammar.  On both there was a lot of additional detail in the extreme LF bass.  I thought this already sounded pretty darn good given the Statement set up but the Melco moved it forward once more.  Notes and nuances were audible which previously just blurred together.  We did several A/B/A comparisons and the second "A" seemed slightly flat compared to the Melco.  Norman also noticed that there was less smearing of the attack phase of notes with a more instant response.

 

I'm not going to make any comments on the technical reasons as to how this product works as frankly I couldn't confidently back up such comments but I know what my ears told me.  By comparison, I only recently learnt how the rear differential of a car works, allowing the wheels to rotate at different speeds when cornering, but it never stopped me using a car before...

 

My only real gripe is that at present it is only available in silver.  Hardly a good match to Naim!  The N1Z is narrower than the larger N1A and it looks like the styling is designed to complement Linn systems.  The N1A is now also available in black (UHES had one, finish looked good), with N1Z shipping in black in three months' or so.

 

A short demo but I was impressed. UHES have kindly offered to lend me one for a week or so, possibly next month, so I'll get the chance to hear it in the context of my own system and room.  I'll report back on that as it happens.

Posted on: 07 June 2015 by Graham Clarke
Originally Posted by Simon-in-Suffolk:

Graham do you have more info on this 'isolation' port. Physical isolation is of some benefit, certainly with RFI etc, but in my experience the greatest tangible benefit would be provided by creating a controlled bridge, or a routed subnet with certain broadcast forwarders such that the LAN traffic on that LAN segment would be minimised and so related TCP/IP data noise in the receiving or network player  stack would be minimised. I think this becomes more of a sensitive issue with combined streamer DACs like the ND series, perhaps less so with a seperate DAC.

Simon

Sorry, I don't have further details.

 

I do wonder whether adding complexity (more cables, more "IT" hardware etc) will actually help though as you have less control over the components between the data stored on disk/SSD and the device turning it back to an analogue signal.

 

Simplistically, I'm assuming that having a single Ethernet cable between data storage device and streamer would give better results as everything is then a known quantity.  That seems to be the approach taken by Melco.

 

Posted on: 07 June 2015 by impy

Afternoon everyone

 

I have just read a review of this in Hifi plus - Alan Sircom really rates it for sound quality.

 

I am not into computer audio at present, so I can't give you any technical details.

 

Regards

 

Trevor

Posted on: 07 June 2015 by Simon-in-Suffolk

Graham, thanks, yes much professional hardware equipment is often depreciated over 5 years, but my accountant colleagues tell me that's really more from an accounting asset value management perspective rather than the equipment is expected to fail or be replaced in its 5 year. As you suggest I think Alan is really referring to consumer 'IT' which appears high churn short life.

It sometimes suprises me how many organisations have 10 year old equipment and older and its maintenance support is good business.. as the equipment is so reliable , and there are also a sizeable few that don't appear to have any maintenance support and run at 'risk' but because the kit is 'rock solid' and there are some 'spares' there seems little concern... Clearly some of us try and point out this is not always the best course of action.. 

Its funny how expected longevity between professional software and hardware seem to have significantly diverged. 

Simon

 

PS the single Ethernet lead's traffic will be the same whether it's on a multiple port switch or a single port on a device on the same LAN (subnet). One would need to seperate the broadcast and multicast broadcast frames out to have a data traffic impact.. And that will require some thought and needy complexity.. But would be effective. I have experimented at SinS towers a little while back in this regard, and it seemed worthwhile. However I have found separating and decoupling the DAC from the network streamer a more practical and effective proposition.

Simon

 

Posted on: 08 June 2015 by analogmusic
+1
 
I'm also keenly waiting to hear the new Chord DAC. Chord has set up huge massive expectations, and I hope it delivers. 
 
 
Originally Posted by Steve J:

 the new MOAD from Chord. I know which I'd prefer to buy, and which will sound better.

 

Posted on: 08 June 2015 by Foxman50

Oh the anticipation 

Posted on: 09 June 2015 by nbpf

Thanks for the report Graham! I checked the Melco webpages after the units were discussed in this forum a few weeks (perhaps months) ago and I have to say that, both in that occasion and after reading your report I felt both intrigued and disappointed.

 

I certainly would expect devices like the Melcos (and the SSS Rendus, the Auralic Aries, etc.) to have a positive impact on the sound quality of the replay chains downstreams.

 

But I see such impacts as obvious consequences of poor designs of those replay chains.

 

I am not claiming one can do better at the present but I very much hope some smart company will soon produce a DAC that just loads a selected list of files (I do not care whether via USB from attached storage, TCP/IP from file servers or via other protocols from other sources) stores them in a local dedicated memory, cuts any communication with (and power supply to) the outside world (included, upon user setup, the control points used to pause and resume playing) and starts replay.

 

Such a design would be source agnostic by construction. It would make the impact of upstream devices (servers, streamers, cable interconnects, etc.) on sound quality nil.

 

I understand that, until we have such a DAC, the Melcos (Rendus, Aries, etc.) do matter.

 

But I do not want to invest (significant amounts of) money in sophisticated devices built to alleviate the consequences of essentially flawed DAC/streamer designs.

 

On the other hand, I would be  happy to support companies that invest in overcoming such flaws and promote short data paths and straightforward designs. 

Posted on: 09 June 2015 by SongStream
Originally Posted by nbpf:

Thanks for the report Graham! I checked the Melco webpages after the units were discussed in this forum a few weeks (perhaps months) ago and I have to say that, both in that occasion and after reading your report I felt both intrigued and disappointed.

 

I certainly would expect devices like the Melcos (and the SSS Rendus, the Auralic Aries, etc.) to have a positive impact on the sound quality of the replay chains downstreams.

 

But I see such impacts as obvious consequences of poor designs of those replay chains.

 

I am not claiming one can do better at the present but I very much hope some smart company will soon produce a DAC that just loads a selected list of files (I do not care whether via USB from attached storage, TCP/IP from file servers or via other protocols from other sources) stores them in a local dedicated memory, cuts any communication with (and power supply to) the outside world (included, upon user setup, the control points used to pause and resume playing) and starts replay.

 

Such a design would be source agnostic by construction. It would make the impact of upstream devices (servers, streamers, cable interconnects, etc.) on sound quality nil.

 

I understand that, until we have such a DAC, the Melcos (Rendus, Aries, etc.) do matter.

 

But I do not want to invest (significant amounts of) money in sophisticated devices built to alleviate the consequences of essentially flawed DAC/streamer designs.

 

On the other hand, I would be  happy to support companies that invest in overcoming such flaws and promote short data paths and straightforward designs. 

Is that not what a streamer with a built-in DAC already is?  NDX/S etc?

Posted on: 09 June 2015 by nbpf
Originally Posted by SongStream:
Originally Posted by nbpf:

Thanks for the report Graham! I checked the Melco webpages after the units were discussed in this forum a few weeks (perhaps months) ago and I have to say that, both in that occasion and after reading your report I felt both intrigued and disappointed.

 

I certainly would expect devices like the Melcos (and the SSS Rendus, the Auralic Aries, etc.) to have a positive impact on the sound quality of the replay chains downstreams.

 

But I see such impacts as obvious consequences of poor designs of those replay chains.

 

I am not claiming one can do better at the present but I very much hope some smart company will soon produce a DAC that just loads a selected list of files (I do not care whether via USB from attached storage, TCP/IP from file servers or via other protocols from other sources) stores them in a local dedicated memory, cuts any communication with (and power supply to) the outside world (included, upon user setup, the control points used to pause and resume playing) and starts replay.

 

Such a design would be source agnostic by construction. It would make the impact of upstream devices (servers, streamers, cable interconnects, etc.) on sound quality nil.

 

I understand that, until we have such a DAC, the Melcos (Rendus, Aries, etc.) do matter.

 

But I do not want to invest (significant amounts of) money in sophisticated devices built to alleviate the consequences of essentially flawed DAC/streamer designs.

 

On the other hand, I would be  happy to support companies that invest in overcoming such flaws and promote short data paths and straightforward designs. 

Is that not what a streamer with a built-in DAC already is?  NDX/S etc?

Not really as NDX/S appear to profit from dedicated devices like the Melcos. Also NDX/S do buffer incoming data but do not store local copies of entire playlists. This implies that communication and synchronization with sources (and, dependently on the data formats, uncompressing) and control software is going on at replay time. Even worse, NDX/S need a wired Ethernet connection at reply time for high resolution playback. This is a bit insane in my view and implies that these devices are by no means source agnostic.