What we have learnt from the Naim DAC...

Posted by: PureHifi on 11 November 2009

I thought it was time that we posted a few tid-bits of our experience with the Ripping, storage and playback of music after our resent promotional event that formed a part of the New Zealand Naim roadshow.

As a Naim retailer we learned a lot from the New Zealand Naim distributor, Chris Murphy, and are finding the whole move to digital storage and playback very interesting...and a lot more involved than a lot of people might initially think.

Our roadshow kit was as follows:

CDX2 (latest with digital out)
HDX (used with digital out)
DAC (with and without a 555PS)
282/supercap, 250.2, Fraim/ Ovator S-600's

Apart from the obvious new speaker exploration we settled down to get to grips with the new CDX2 + DAC and experiment with the HDX a bit more. In particular we wanted to get a solid handle on the performance variations with ripping CD media and the storage and playback.

Ripping the software used does affect the final playback, Chris had files from the same CD Track ripped to WAV with several software packages (WMP, iTunes, EAC, DBpowerAmp, HDX, etc) and the outright winner was the HDX rip - iTunes sounded horrible - ragged & sibilant would best describe it.

Storing It became apparent that the storage medium also influences the audio performance. The simplest example was writing the same HDX ripped music file to different USB memory sticks and then playing them back via the USB input on the DAC - cheap memory sticks generally played back poorly compared to a higher priced stick from LaCie. Perhaps this reflects on Naim's choice of Hard Disk drives used in the HDX.

Playback Several things in the playback domain have cropped up recently for us, in both Naim's solutions and other brands that we stock. We have found, in terms of digital playback from PC laptops and MacBook, that the media player software has a huge impact on the music quality - WMP was poor but Winamp and Foobar were great by comparison (never got around to iTunes). The USB cable between a Cambridge DacMagic and our laptop was also a hugely variable item, we tried a range of normal looking USB cables (some with superior shielding and RF stoppers) and ran them against a Wireworld Ultraviolet USB1 cable, results were outstandingly in favour of the Wireworld item over all others tried. We could not try our Wireworld USB cable on the WD external HDD because of it's use of a mini USB connector but I am looking forward to testing it on a LaCie HDD that has the right USB connector.

The Western Digital 1TB external HDD feeding the HDX also benefited from a power supply upgrade, it might sound over the top but it made for a better result in the music (and was a good use for a spare NAPSC).

I can't explain in technical terms why a lot of these findings are the way they were, we judged everything on its sound merit and I am reporting our results.

What it highlighted most of all was the wonderful way that Naim's R & D has worked to make all those variables disappear for a customer in the HDX product - by providing Superior Ripping, storage and playback in one box.

The DAC is a wonderful device and will be a very successful product.

Posted on: 15 November 2009 by AMA

quote:
Only one of them at most is the correct one!

And this is the one which is selected by receiver.
What's the problem?

Posted on: 15 November 2009 by likesmusic

quote:
Originally posted by AMA:
quote:
Only one of them at most is the correct one!

And this is the one which is selected by receiver.
What's the problem?

Acording to NAIM here The rate at which the memory fills and empties is controlled by the DAC automatically selecting the oscillator that matches the average incoming clock frequency

So, let's say in the case of redbook CD, if the clock in the sender is slightly fast, the NAIM DAC will have to pick a clock that is slightly fast, and if the clock in the sender is slightly slow, the NAIM DAC will have to pick a clock that is slightly slow. It has no alternative if buffer over- or under-flow are to be avoided. But, since it is red-book cd we are talking about, we know what the correct clock frequency should be - only one clock can be correct for the data. (The one appropriate to 44.1kHz). But the NAIM DAC still 'sees' the clock in the sender, which may not be accurate, it is just not susceptible to jitter from it. I presume that the DAC must also change it's clock if buffer over- or under- flow is impending, which is a form of very slow jitter.

Posted on: 15 November 2009 by Hook

quote:
Originally posted by js:
quote:
Originally posted by glevethan:
JS

I assume when you say MM you are referring to Mac Mini. Are you implying that, so far, you have been unable to get satisfactory results with Apple computers?

Gregg
Sorry again. Not at all. Media Monkey. Haven't tried anything MAC other than an Ipod touch with great success. Thanks for that. Certainly don't want a misunderstanding. I suspect that Macbook/Amarra will be outstanding as I did finally manage very good results from a PC with a free player program.

And away it goes.

JS-

Thanks for posting your computer audio findings. One thing I like about computer audio is that auditioning and tweaking is relatively easy and almost always free. Like you say, with a little effort, some very good results can be found.

While I await my Naim DAC delivery, am using a M-Audio Transit to feed the 24/96 Toslink input on an MF V-DAC. I like Mediamonkey for organizing the library, and EAC for ripping problem CDs. MM's sound quality is not ideal under Vista 64 SP1. Waveout seems better than DS or ASIO, but still not up to a mid-fi player's quality (mine is an Oppo BD83).

Could not get foobar2000 wasapi plug-in to work for anything beyond 16/44.1 flac files, but am having very good luck with xmplay. Its wasapi output plug-in works great in my setup for both redbook and 24/96 flac files, and the sound quality is well above that of my player. As you said, MM V4 will support wasapi, at which point I'll give it another listen. Until then, it taking me multiple apps to get a good result.

Regards,

Hook

Posted on: 15 November 2009 by BobF

quote:
Originally posted by DHT:
Should so many 'trade' members be allowed on the forum, they have their shops to peddle their wares, I for one am becoming increasingly tired of this twaddle.
We all know what is 'best' for you JS.

JS and PureHiFi have posted some interesting comments on this thread, They both strike me as enthusiasts and are clearly labeled as trade members. If your "We" includes others on this thread count me out, but it may be the Royal We or perhaps you have a mouse in your pocket

Bob

Posted on: 15 November 2009 by DHT

For the attention of the forum users, let it be known 'BOBF' is happy to be fed mush.
Happy now?

Posted on: 15 November 2009 by AMA

quote:

I presume that the DAC must also change it's clock if buffer over- or under- flow is impending, which is a form of very slow jitter.

likesmusic,
I think your presumption is not OK.
As I got it from this Naim DAC S/PDIF receiver evaluates the input bitstream clock and then selects a FIXED clock from one of 10 standard clocks. These fixed clocks are quite far away from each other.
The regular digital source with decent jitter will not take your clock far from the one of standard presets.
At least that far that your buffer will stack over/under.

But if your clock is so strange and/or moving far away from standard clocks then a SHARC will implement ASRC to retrieve a master clock -- which is your presumed scenario.
Which is -- I agree -- not the best option as it will not reject all the jitter.
As well as PLL is not the best option as it will not reject all the jitter.
As well as I have no idea what is the other DAC on the market which can handle such a crazy signal successfully?

As well as feeding that crazy jittery transport into a Naim DAC is not a good idea (but Naim DAC will try the best to reject a jitter through ASRC).
Ten standard clock presets bridge the range of all popular sample rates.
If you feed a decent transport into a Naim DAC you can be close to almost 0 jitter for almost any popular PCM bitstreams.

I really don't see any problem with multiple clocks + ASRC backup.

Posted on: 15 November 2009 by js

quote:
Originally posted by ghook2020:
quote:
Originally posted by js:
quote:
Originally posted by glevethan:
JS

I assume when you say MM you are referring to Mac Mini. Are you implying that, so far, you have been unable to get satisfactory results with Apple computers?

Gregg
Sorry again. Not at all. Media Monkey. Haven't tried anything MAC other than an Ipod touch with great success. Thanks for that. Certainly don't want a misunderstanding. I suspect that Macbook/Amarra will be outstanding as I did finally manage very good results from a PC with a free player program.

And away it goes.

JS-

Thanks for posting your computer audio findings. One thing I like about computer audio is that auditioning and tweaking is relatively easy and almost always free. Like you say, with a little effort, some very good results can be found.

While I await my Naim DAC delivery, am using a M-Audio Transit to feed the 24/96 Toslink input on an MF V-DAC. I like Mediamonkey for organizing the library, and EAC for ripping problem CDs. MM's sound quality is not ideal under Vista 64 SP1. Waveout seems better than DS or ASIO, but still not up to a mid-fi player's quality (mine is an Oppo BD83).

Could not get foobar2000 wasapi plug-in to work for anything beyond 16/44.1 flac files, but am having very good luck with xmplay. Its wasapi output plug-in works great in my setup for both redbook and 24/96 flac files, and the sound quality is well above that of my player. As you said, MM V4 will support wasapi, at which point I'll give it another listen. Until then, it taking me multiple apps to get a good result.

Regards,

Hook

I'm still at XP in the laptop so no WASAPI there. Try Foobar with the ASIO plugin. I didn't have time to play with the buffer settings but the defaults seemed fine. I now have 7 in my home PC but don't use that one for tunes so no real experience with WASAPI though I like the concept. I also suspect the the big 2 can write good codecs so it's possible that the most common players may still have more to offer via WASAPI. In your position, that is if you were looking for one player for everything and didn't want Foobar/ASIO, I'd consider retrying those in that environment and zero out all the fade and leveling options. Let us know if if you do and the result. Sounds like you're doing fine on your own though.

There's also a J River player that I thought less good in the past but need to try again. I believe there's a native ASIO option in that one.

Posted on: 15 November 2009 by likesmusic

AMA .. I read the white paper too and absolutely understand that the DAC selects a FIXED clock.

But I'm not sure I understand that there is only one such fixed clock for each possible input sample frequency.

If there was only one clock for, say, 44.1kHz red book then at some point the buffer MUST overflow or underflow, unless the clock in the sender was perfectly matched to the clock in the DAC, which isn't likely.

Maybe I'm wrong. I don't know how consistently accurate such clocks can be. To some extent it depends how big the buffer is. But if it is big, then the DAC will be useless for DVD etc. as the delays will destroy lip-sync.

Posted on: 15 November 2009 by Hook

quote:
Originally posted by js:
quote:
Originally posted by ghook2020:
quote:
Originally posted by js:
quote:
Originally posted by glevethan:
JS

I assume when you say MM you are referring to Mac Mini. Are you implying that, so far, you have been unable to get satisfactory results with Apple computers?

Gregg
Sorry again. Not at all. Media Monkey. Haven't tried anything MAC other than an Ipod touch with great success. Thanks for that. Certainly don't want a misunderstanding. I suspect that Macbook/Amarra will be outstanding as I did finally manage very good results from a PC with a free player program.

And away it goes.

JS-

Thanks for posting your computer audio findings. One thing I like about computer audio is that auditioning and tweaking is relatively easy and almost always free. Like you say, with a little effort, some very good results can be found.

While I await my Naim DAC delivery, am using a M-Audio Transit to feed the 24/96 Toslink input on an MF V-DAC. I like Mediamonkey for organizing the library, and EAC for ripping problem CDs. MM's sound quality is not ideal under Vista 64 SP1. Waveout seems better than DS or ASIO, but still not up to a mid-fi player's quality (mine is an Oppo BD83).

Could not get foobar2000 wasapi plug-in to work for anything beyond 16/44.1 flac files, but am having very good luck with xmplay. Its wasapi output plug-in works great in my setup for both redbook and 24/96 flac files, and the sound quality is well above that of my player. As you said, MM V4 will support wasapi, at which point I'll give it another listen. Until then, it taking me multiple apps to get a good result.

Regards,

Hook
I'm still at XP in the laptop so no WASAPI there. Try Foobar with the ASIO plugin. I didn't have time to play with the buffer settings but the defaults seemed fine. I now have 7 in my home PC but don't use that one for tunes so no real experience with WASAPI though I like the concept. I also suspect the the big 2 can write good codecs so it's possible that the most common players may still have more to offer via WASAPI. In your position, that is if you were looking for one player for everything and didn't want Foobar/ASIO, I'd consider retrying those in that environment and zero out all the fade and leveling options. Let us know if if you do and the result. Sounds like you're doing fine on your own though. There's also a J River player that I thought less good in the past but need to try again. I believe there's a native ASIO option in that one.

JS -

Did try foobar2000/ASIO. Had the same problem as WASAPI -- no luck with 24/96 flac files.

Just did a little a/b with foobar2000/ASIO and xmplay/WASAPI using 16/44.1 flac. Could not hear any difference. Both sounded very good. Given both are supposed to produce bit-perfect playback, this is not a surprise. (Just a side note: the Transit supports ASIO, so did not have to use ASIO4ALL).

The only minor issue I have with xmplay/WASAPI is with gapless support. Am hearing a very short cutout between tracks on live albums. Gapless support in foobar2000 was perfect.

Hook

Posted on: 15 November 2009 by AMA

quote:
Originally posted by likesmusic:
If there was only one clock for, say, 44.1kHz red book then at some point the buffer MUST overflow or underflow, unless the clock in the sender was perfectly matched to the clock in the DAC, which isn't likely.

likesmusic, if input clock is perfectly matched with DAC master clock then why do we need a buffer ???
Let the input go straight to converter chip.
Buffer was thought of exactly to REJECT the small clock mismatch which is caused by small jitter.
Otherwise DAC goes into ASRC clock recovery which is different from multi-clock approach (more close to PLL concept). Try to avoid using exotic sample rate recordings and high jitter transports and Naim DAC will do a job just right.

quote:
then the DAC will be useless for DVD etc. as the delays will destroy lip-sync.

Your concerns relate to all re-clocking DACs -- no matter how do the input S/PDIF receiver retrieves the source clock. It's not an issue of a Naim DAC only.

1. Many players (for example VLC) provide an option for customizing delay between audio and video streams.

2. If you use high quality digital transport like HDX I think the re-clocking delay will not be that big.

3. I agree that the main application of Naim DAC is to listen for music rather than watch DVD

likesmusic, Naim DAC design is very advanced. Let's see how it's implemented because I did not yet listen for one.

Posted on: 15 November 2009 by PureHifi

I have been reading the posts with interest and have a couple of things to add:

The tests of USB memory sticks were with the DAC, not the HDX (see OP). Memory sticks used were a 1GB OTi, 4GB LaCie iamaStick, 2GB Toshiba, and a 2GB Adata. I could grade them but the important thing was that there was a very noticable difference, not just to us but a room full of customers.

The iTunes VS HDX rip was demonstrated from USB memory stick into the DAC, from USB attached HDD storage into the HDX and also on a customers MacBook where he used iTunes to play back both WAV files from the Mac HDD (in this case we used a Chord Company Optichord into the DAC from the MacBooks mini toslink optical out).

We have no doubt that very good results can be generated from a variety of products, software and dedication of the end user. What this subject throws up for us is a widening of the way in which people purchase, store and playback music. The huge variation in what can be achieved and what individuals find acceptable is always up to the end user, we are just trying to explore it and share our experience...in the hope that someone will find it useful information when considering their options.

Posted on: 15 November 2009 by likesmusic

Purehifi; you specifically mentioned in your original post that HDX rips sounded better than dbPoweramp rips.

Can you clarify the circumstances under which you found this to be the case?

Posted on: 15 November 2009 by Guido Fawkes

quote:
Originally posted by scottyhammer:
Yawn

Pay attention - there may be a quiz to win a new DAC at the end of the thread.

Posted on: 15 November 2009 by js

quote:
Originally posted by ghook2020:
JS -

Did try foobar2000/ASIO. Had the same problem as WASAPI -- no luck with 24/96 flac files.

Just did a little a/b with foobar2000/ASIO and xmplay/WASAPI using 16/44.1 flac. Could not hear any difference. Both sounded very good. Given both are supposed to produce bit-perfect playback, this is not a surprise. (Just a side note: the Transit supports ASIO, so did not have to use ASIO4ALL).

The only minor issue I have with xmplay/WASAPI is with gapless support. Am hearing a very short cutout between tracks on live albums. Gapless support in foobar2000 was perfect.

Hook

I'm not personally familiar with the transit but I recall someone mentioning that disabling the input allowed 24/96 operation with their setup. Perhaps it's worth a try. After hearing kernal vs asio on Foobar it's been reaffirmed for me that there's more to this than just bit perfect. They were very close in this instance but ASIO was preferred. Close enough that it probably would have been inconsequential on a lesser DAC. Really, any player via ASIO should be bit perfect but we both didn't especially care for the same one. Confirmation is nice but I'd rather it was all the same.

Of course, I'm not sticking my head in the sand over it.

Posted on: 15 November 2009 by js

quote:
Originally posted by AMA:
quote:
then the DAC will be useless for DVD etc. as the delays will destroy lip-sync.

Your concerns relate to all re-clocking DACs -- no matter how do the input S/PDIF receiver retrieves the source clock. It's not an issue of a Naim DAC only.

1. Many players (for example VLC) provide an option for customizing delay between audio and video streams.

2. If you use high quality digital transport like HDX I think the re-clocking delay will not be that big.

3. I agree that the main application of Naim DAC is to listen for music rather than watch DVD

likesmusic, Naim DAC design is very advanced. Let's see how it's implemented because I did not yet listen for one.

I didn't see the original post but it's in the reply and I can help here. I didn't experience any noticable latency in the DAC. They all have buffers and of course some latency with them but nothing here of note. Reclockers have more of them. With what this DAC is doing, even though it's not your typical reclocker, I thought it may have been possible to notice something but I didn't. In Wavelab, you can see the peak file cursor follow the wave form on the monitor during playback and no disconnect was noticed. Transits are easy to track visually. I think you'll find that with video, it's always the audio signal that needs to be delayed in settings to match the not insignificant lag of video processing anyway as everthing is scaled and usually more than once.

Much higher video bit rates also means an even bigger buffer to do it's thing even without scaling. It's a very poor assumption that any DAC would lag behind video processing. In the USA, surround is a must for retailers so we have plenty of experience here. Should be a complete non issue. I'll verify when I have another opportunity.

Posted on: 16 November 2009 by CPeter

quote:
there may be a quiz to win a new DAC at the end of the thread

Skimming through this thread, I don't think I want a DAC.

Posted on: 16 November 2009 by church warden

CPeter - don't despair, matters are no where near as complicated as this series of postings is suggesting.

Mac Mini - Firewire to Weiss DAC2. Very simple to set up and (though it was not what I wanted to hear at the time of testing) for me this clearly outperforms my CDX2/XPS2 combination by some margin and at just over 1/3rd of the cost.

Posted on: 16 November 2009 by CPeter

CDS2, lovely

Posted on: 16 November 2009 by js

Not trying to overcomplicate things and there are tried and true methods as mentioned. Using an Ipod touch as a streamer is remarkably easy and consistant. No fiddling about with players or interfaces etc. and for me, the DAC did sound better than others tried via the same connections regardless of what they were. The relative level of performance in your system will be personal but I suspect most will be very happy they tried it.

Posted on: 16 November 2009 by Colin Lorenson

quote:
Originally posted by church warden:
CPeter - don't despair, matters are no where near as complicated as this series of postings is suggesting.

Mac Mini - Firewire to Weiss DAC2. Very simple to set up and (though it was not what I wanted to hear at the time of testing) for me this clearly outperforms my CDX2/XPS2 combination by some margin and at just over 1/3rd of the cost.

Heard that combination when my brother was trying out the Weiss. On 44/16 I'd describe it as a competent music player but really nowt to write home about. Surprised you prefer it to CDX2/XPS.

Hi-def brought the combo to life though.

Posted on: 16 November 2009 by Joe Bibb

quote:
Originally posted by church warden:

Mac Mini - Firewire to Weiss DAC2. Very simple to set up and (though it was not what I wanted to hear at the time of testing) for me this clearly outperforms my CDX2/XPS2 combination by some margin and at just over 1/3rd of the cost.

Yes I found it quite a lot better too. Nice simple combo. Closer with a CD transport but Mac/Firewire was quite a lot better. Hi def stuff is superb but obviously limited in terms of content.

Looking forward to hearing the Naim DAC now that I know what is possible with similar outlay.

Joe

Posted on: 16 November 2009 by church warden

Colin - so was I and it was not what I wanted to hear. I was testing the DAC to prove to my partner that my CDP was still better than the computer front-end solution she wanted. Everyone who has heard the comparison in my place has chosen the DAC sound - more detail and smoother were the usual comments.

Maybe my Naim pre/power amps were not good enough for the CDX2/XPS2. Though the DAC also sounded better through a Stax energiser and headphones.

Anyway, with a rather heavy heart I sold the CDX2/XPS2, because I had stopped using it, and have a sound I like and am £900 up on the switch.

Posted on: 16 November 2009 by BobF

[QUOTE]Originally posted by DHT:
let it be known 'BOBF' is happy to be fed mush.
QUOTE]

I assume your refering to your posts

Posted on: 16 November 2009 by js

quote:
Originally posted by ghook2020:
JS -

Did try foobar2000/ASIO. Had the same problem as WASAPI -- no luck with 24/96 flac files.

Just did a little a/b with foobar2000/ASIO and xmplay/WASAPI using 16/44.1 flac. Could not hear any difference. Both sounded very good. Given both are supposed to produce bit-perfect playback, this is not a surprise. (Just a side note: the Transit supports ASIO, so did not have to use ASIO4ALL).

The only minor issue I have with xmplay/WASAPI is with gapless support. Am hearing a very short cutout between tracks on live albums. Gapless support in foobar2000 was perfect.

Hook

Funny part is that MM/ASIO is bit perfect also.

beats me.
I forgot to ask. Did you try WMP or Itunes via WASAPI? I suspect they're both pretty good that way. I just configured WMP to play via ASIO in XP. It will also output 24/96 as configured. Don't know the quality yet but I can see that things are working correctly and outputting native on the TC ASIO app. Here's how.

first this: http://sourceforge.net/projects/asiowmpplg/ and read the text file for settings.

Then this for 24/96.

http://ffdshow-tryout.sourceforge.net/

Once it’s installed, launch the audio decoder configuration and under “Codecs” select “Uncompressed” ànd switch the next collumn to 24-bit integer.

Haven't listened yet but it may be worth a go for those still with XP that want a do everything player with these capabilities. I'll listen to it next time the DAC's on hand or maybe before. I've got a few players ready to roll. I'll also try 0 compression setting Flac files to see if there's a difference from higher settings. Speaking of FLAC. There's also a FLAC switch under the same audio configuration settings as mentioned before. I left it off but you could try it to see if it helps in your setup, that is if these even apply in Vista.

Posted on: 17 November 2009 by js

Further along. I now have WMP working with flac via a DCbasssource download and also got a tag app. Files to be added to library for tags but anything can be dragged and played. Plays 24/96 flac files also. Plays at 96K via ASIO and still plays fine through the mixer so in the second case should be down mixing. The last 2 posts' experiments all went without aa hitch so not very time consuming. More lucky than good I'm sure.

Just using other peoples work and my laptop cooperated.