DSD playback on superuniti future?

No replies so far. May be I should explain a bit more;

DSD is the format used on a SACD (Super audio CD). Sound quality is as high (or higher depending on who you talk to) then 24/192. In the old days SACD's were not rip-able but currently there are methods to rip a SACD and , better, downloads of SACD\DSD material is gaining momentum. Files sizes are comparable with 24/192 files.

All in all perfect for streaming. A dedicated DAC is needed (for now?) to play the files (for example Mytek Stereo 192 ) to your (and mine ) Naim gear.

My questions are therefore:

Is anyone in the naim community intrested is playing DSD files via their Naim gear?

Does anyone know whether the feature is in the pipeline of Naim?

Regards

Frits

Not particularly familiar with the format, but then again most formats don't interest me. Is there something special about DSD which renders it incapable of being rendered as a WAV? Compression? Post recording processing maybe? I can't see the need for yet more propitiatory formats for playback, but if HiRes material is becoming available form SACD which can be accurately rendered as a 24/192 WAV I don't see what's not to like. Need to hear one though.

Hi Frits -

I agree that DSD holds a lot of promise for streaming "master recordings".

I am currently using a Korg MR-1000 to record my vinyl to DSD files (one file per side), but am also (very slowly!) converting those DSD files to 24/88.1 FLAC files (one per song) for streaming purposes. 

I can currently use my recorder's analog output to listen to the DSD files, and they sound amazing.  But it is a cumbersome setup, so I do not do this very often.  I hope that someday DSD will be supported in Naim network players, but I doubt this will happen anytime soon.

So, I am afraid the only option for today is to look at external 3rd-party DAC's that support DSD via a USB connection to a PC or Mac.   Meitner, DCS and Playback come to mind, but there are others as well.

Good luck!

Hook

Hi Frits,

Not an exact answer to your question but I rip the SACD image to 24-176 flac files (via foobar) and put them on my NAS, which transcodes them to wav as they are streamed to my NDS.

they sound great 

All the best

Stefano

Now that sounds promising.

Hi Stefano,

Actually how do you rip? As far as I have learnt SACD-players dont have a digital out (may be the latest Denon's do btw). Or are you using a PS3?

Can imagine the sound q :-)

Hi Harry,

Adding links does not work on this forum so I have to do it in writing :-)

http://en.wikipedia.org/wiki/Super_Audio_CD (google wiki sacd) and you have a first glance on the SACD format. And take it from there depending on your apatite.

My view: DSD is far superior over WAV. But dont let us start a debate on it.

Hi Hook

Thanks for the inside. U are using some Pro equipment :-)

I was hoping Naim would jump on the opportunity. I hope they have a outside view and see the increasing amount of available (mainly classical) DSD recording (on top of home made ones) and are crushing as we speak on soft/hardware for their customers.

Thanks all.

If Naim personal is willing to comment, please do.

What is Naim's view on DSD streaming?

well, that's the point... I don't rip the SACD, but its image (iso file) i.e. somebody has ripped it off the SACD for me.

You can find them around

Yet another interesting format, why it is technically better I have not really seen explained. But forget it because we can buy at least 20 titles, so Naim you must really added  into all streamers and servers now!!

Of course High res is important, but the software availability should come before Naim puts a lot of resources into such a project. I see a lot more sense in   working on Airplay or much better a Naim implementation of the songcast feature ffrom Linn, so all music can be played and syncronized. Of course it is not highres to stream youtube from the pc, but if I want to here something available only from my computer I would much prefer to be able to listen to it on my hifi, instead of always going back and listen to the 20 outstanding titles in 10 different high res formats.

Claus

DSD is allegedly superior to PCM and DACs from Chord Electronics and dCS support direct DSD streaming [Chord has USB interface so you can just send it DSD from your Mac] ... it is possible to convert to PCM, but then you lose the advantages (whatever they are). If you agree SACDs sounded better than DVD-A then DSD is superior. 

Another forum member explained to me how to rip a SACD, but you need a SACD universal player to do it and you end up PCM out of its HDMI which you send through a splitter to separate the audio channels from the video ones. 

However, if you download DSD then ripping is unnecessary, it is nice to have something that can play it. The Creedence Clearwater Revival catalogue is available as DSD so there is plenty of other very good stuff in the format. It is shame to downgrade to WAV (a bit like turning a SACD into a DVD-A).

In a comparative test, the subjects couldn't tell which was which.  

With nearly 50,000 PCM tracks in my digital juke box, I'm not 100% sure I need DSD, but it would be nice feature on the DAC555. 

This is where it gets difficult for me. What's the bit depth and sample rate of DSD? Superior to 24/192 WAV? Or yet another proprietary format trying to paint users into a corner? If the numbers are comparable but DSD sounds better, then we're just into another different versus better discussion with no correct answer. If the resolution is higher then that's another matter - bring it on.

Hi Harry - 

As you know, PCM samples have a bit depth (e.g., 16, 24) and a sample rate (e.g., 44.1kHz, 96kHz, etc.).  It is the same for DSD, except the sample rates are much higher (2.8mHz, 5.6mHz), and the bit depth Is only 1.  

In their white paper entitled "The 1-bit Advantage - Future Proof Recording", Korg explains:

"...It’s easy to follow the benefits of higher, and super-high sampling rates. But what might at first confuse the reader is the benefit of moving from high bit rates down to 1-bit. Surely the increased resolution of higher bit rates must be more accurate, right?

The basic concept here is that once the sampling rate is increased to such high levels, each step doesn’t need to be defined with such detail. With such frequent readings of the current state of the audio waveform, each step need only be defined in the simplest of terms – has the signal increased since the last step, decreased, or remained the same. 1-bit offers only two values, a 1 or a 0. Either up from the previous sample or down. And at these super-sampling speeds a steady state can be represented by alternating 1’s and 0’s. The chance for error in such a system is much less than in multi- bit approaches. Consider this: 

In a 1-bit system the possible values for each measurement are simple, either it's a 1 or a 0 - there's little chance to get a wrong value. In a 24 bit system there are 16,777,216 possible values. So which system is more likely to be accurate for each reading?...".

As a proof point, Korg shows a square wave sampled at various bit depths and sample rates for both PCM and DSD.  As expected, the squarest looking wave was produced by 5.6mhz DSD.

Generally speaking, SACD (1/2.8mHz) is comparable to DVD-A (24/96), and double-rate DSD (1/5.6mHz) is comparable to hi-resolution downloads (24/192).  Of course, each format has their proponents...

ATB.

Hook

Thanks Hook. It will be interesting to see where this format goes. 

Interesting how many different discussion papers there are on the benefits or not of DSD vs PCM.

I am only copying here what some experts have to say about this so please don't shoot the messenger 

First a bit of history and argument:

" DSD as a format dates back to a time when 1-bit sigma-delta converters were state-of-the-art and digital processing algorithms were limited by available processing power. The original arguments for DSD included better fidelity due to less processing (no downsampling or upsampling), and lower distortion due to the inherent linearity of 1-bit converters. Things have moved on considerably since then. Modern multi-bit sigma-delta converters achieve excellent linearity without resorting to the aggressive noise shaping of DSD, and modern desktop processing power is such that DSD files can be downsampled to high bit-rate PCM with incredible accuracy.

The only remaining 'argument' for DSD is that it has a wider bandwidth than any available PCM format, but when the whole audio path is considered, this becomes meaningless. Bandwidth limiting occurs at every other stage in the chain: microphone, ADC input, DAC output, amplifier, loudspeaker, ear(!), so when seen in this context, a digital bandwidth of over 1MHz is irrelevant, especially when most of it is taken up with quantization noise."

Then a bit about issues with making native DSD processing possible in typical PCM DACs

" There are a number of problems involved with supporting DSD files. Firstly, the architecture of typical DAC designs is heavily tied to PCM, so some fairly major architectural changes would need to be made. Secondly, only certain DAC's can support DSD (even then, because they are fundamentally multi-bit DACs they cannot truly operate in 'pure' DSD mode). 

Also, it could be argued that DSD is actually inferior to correctly handled PCM. DSD uses aggressive noise-shaping to push the 1-bit quantization noise out of the audio band. This in turn requires an aggressive analog filter to remove the noise. Either that or hope that the following audio stages are not affected by it.

A further drawback of DSD is that it cannot be processed in any way without first converting it to PCM and, realistically, also downsampling it to a more reasonable sample rate. This pretty much defeats ths purpose of adding DSD support in the first place (although its exactly what happens to most DSD recordings). Even the most basic of operations such as ramping down during network dropouts would involve a conversion.

Probably the best solution, both for DACs and for DSD playback in general, is to convert DSD to high bitrate PCM (e.g. 88.2kHz 24-bit) using an offline software downsampling process. The downsampling process can then filter out the shaped quantization noise in the digital domain, and the resulting file can be played on almost any device."

In summary, it is very difficult to justify adding native DSD support to the a PCM DAC, especially when a solution already exists to downsample offline to high bit-rate PCM. By using an offline process, a hugh amount of processing power (i.e. any modern PC) can be used to achieve highly accurate results. An added bonus of downsampling is that it can also be configured to remove most of the high frequency noise from the DSD stream, so its quite likely that the converted file will actually sound better that the DSD original."

Geoff

Harry

I wasn't saying DSD is better than PCM or vice versa. As Hook and Geoff explain different parties put forward different views often with lots of maths to support their arguments. 

I put downgrade in italics to question it rather than emphasise it. 

For me the only problem would be if some music I wanted to listen to was DSD only and I couldn't play it ... my hope is that either my player will cope irrespective or all music will be available in a format I can play. 

A bit like a few years when I could have bought a SACD player only to find DVD-A became de facto - as it happened neither did - fortunately I settled on a CDX2, which has done me well, but is going be upgraded to a streamer soon - audition pending.  

All the best, Guy