Superuniti vs Sonos

[davedotco]

Dave, it's good to know that there are others who believe the Transport is an important part of the equation.

I'm afraid i have some pretty 'off the wall' ideas about what does and does not make a difference in a decent system. I have done enough blind testing to know that picking one DAC from another is pretty near impossible when both units are giving of their best, but I have also heard some well regarded DACs sound quite poor on occasion.

My pet theory (today anyway) is that audible differences between 'good' DACS can be caused by the way they handle jitter and out of band noise, I know that modest levels of jitter (and noise) can be shown to be inaudible in themselves, but the effect they have on the DAC and the artifacts this produces may not be.

If different transports, both delivering 'bit perfect' output sound different, could this be a reason?

 

[matt49]

This afternoon I went to a local dealer to demo the Naim Superuniti alongside the Linn Majik DSM. The speakers were PMC twenty 23s, which I suspect may be a little dry and analytical for my taste, but they’re obviously jolly good boxes.

I’m still trying to work out whether to invest in a streaming solution alongside my existing Sonos kit, or whether to pipe Sonos via digital coax into a DAC plus amplification. So just like my home demo of the SU earlier in the week, I streamed music direct through the Naim/Linn boxes and also from Sonos via digital coax into the Naim/Linn.

Three conclusions:

1) the Majik DSM is a super piece of kit. Hugely detailed, wide soundstage through the PMCs, satisfyingly deep and well controlled bass. What particularly impressed me was the control of complex, multi-layered music. (One of my reference tracks is ‘The Great Curve’ off the 2006 remaster of Taking Heads’ Remain in Light: epic!) The Superuniti, by contrast, is more rhythmic but deficient at lower frequencies. Acoustic and classical music sounds less atmospheric.

2) comparing the Majik DSM on its own with Sonos going into the Majik via digital did seem to show up some slight but persistent differences. The Majik on its own seemed smoother, whereas the Sonos into Majik sounded grainy by comparison.

3) both Linn and Naim boxes have their own control apps for iPad. I found the Linn interface a bit easier to use and its functionality greater. It's also worth noting that the Linn can do AV duties whereas the Naim can't.

I’m doing some more listening tonight at home, comparing the study set-up in my signature (where the Sonos goes into the M-DAC via digital coax) with my PC piping the same files into the M-DAC via USB. In theory the latter should be better, as the M-DAC can control the source’s clock via USB, whereas it can’t via digital coax.

 

[bigboss]

The ZP90 (or whatever its called now) has relatively poor jitter performance. If I recall correctly, north of 250pS RMS. In context, the Squeezebox Transporter (which has one of the best jitter numbers of any source) is 15pS RMS.

This doesn't matter one jot if your DAC uses some form of jitter mitigation, but it MAY matter if your DAC just slaves itself to the streamer clock.

The rule is "DAC with jitter mitigation - any old streamer will do, DAC without jitter mitigation - take care about your choice of streamer".

To be honest, the clock doesn't belong in the streamer at all, but we have Sony and Philips to thank for that.

But didn't some scientific tests show that a jitter of 400-600ps is needed to cross the audible threshold? In other words, you can't perceive jitter if it's less than 400ps?

If but life were that simple. Jitter is a random process and isn't really a single number but a distribution. The shape of that distribution apparently matters. There are a number of papers on the web that have established that purely random jitter is less offensive than jitter that is correlated in some way to the programme material. It is tough to find a recent real world study on the threshold of jitter detection.

Digital audio (and particularly the S/PDIF interface) is encoded in a way that leads it open to correlated jitter - which listeners detect more readily.

Looking at the maths (not mine, so I hope its right...) it is possible to turn jitter into equivalent 'bits' of resolution.

324pS of jitter on a 0dB 15KHz 16bit digital audio signal has the effect of reducing the resolution from 16bits to 15bits - I dont know if that is detectable, but it certainly it feels like it should be.

If you can come across a recent paper that uses real world correlated jitter rather than purely random jitter and establishes a threshold of detection, I would be very grateful if you could post a link.

Anyway, its all a moot point if the DAC has a topology that reduces input jitter, the quality of the audio output should be independent of the streamer. So as far as the O/P's comparison goes, if the DAC he uses has a design that mitigates jitter, then any old streamer should be fine.

https://www.jstage.jst.go.jp/article/ast/26/1/26_1_50/_pdf

Note that 1000ps = 1ns

 

[CnoEvil]

I think we hear things in much the same way.

As I mentioned earlier, I have found that Linn amps need a speaker with a touch of warmth, and the PMC Twenty Series are very clean, neutral and a little lean, which could come off as dry unless paired with a richer sounding amp (eg. Class A / Valves etc). Linn I believe uses a form of Class D.

I think the Vertere would work nicely with the DSM, but that's a guess

I have also found that the Linn DS sounds better using it's own DAC/Transport (probably due its upsampling and Dynamik p/s.... and is hard to beat at its price point.

Linn has been around longer, which means more time to get the control software right.

I look forward to see if we have got another addition to the Linn DS thread.

 

[matt49]

I look forward to see if we have got another addition to the Linn DS thread.

Yes, another possibility would be to go for the Majik DS with separate amplification, which will give me many more options, including class A. But in the meantime I'm going to get one of my Sonos ZP90s modded, to see if that gives a significant improvement over the out-of-the-box version. Even if I decide to go the DS route in the end, the modded ZP90 can do service in my study set-up.

I'm currently listening on my Grados to the ZP90 via digital coax into M-DAC vs PC via USB into M-DAC. (It's such an easy comparison to do because both control apps are on my PC so I can run the music exactly in parallel and then just switch source with one click on the M-DAC remote.) The latter definitely has the edge, a slightly less aggressive and granular presentation. Suggests that modding the ZP90 to obviate the jitter is at least worth trying.

 

[davedotco]

I look forward to see if we have got another addition to the Linn DS thread.

Yes, another possibility would be to go for the Majik DS with separate amplification, which will give me many more options, including class A. But in the meantime I'm going to get one of my Sonos ZP90s modded, to see if that gives a significant improvement over the out-of-the-box version. Even if I decide to go the DS route in the end, the modded ZP90 can do service in my study set-up.

I'm currently listening on my Grados to the ZP90 via digital coax into M-DAC vs PC via USB into M-DAC. (It's such an easy comparison to do because both control apps are on my PC so I can run the music exactly in parallel and then just switch source with one click on the M-DAC remote.) The latter definitely has the edge, a slightly less aggressive and granular presentation. Suggests that modding the ZP90 to obviate the jitter is at least worth trying.

I have some modest experience with the Cullen/Wyred4Sound mods that seem to work well, in the early versions you could get a proper 75Ω connector on the SPDIF output, not sure if that remains the case.

FWIW I find that while I can not consistently hear differences in rca digital connectors though I usually find that a 75Ω BNC terminated cable to be superior on most counts, it is DAC dependant though.

Ref your setup, connecting a computer to a DAC electrically can allow RF noise to leak into the DAC, this often manifests itself as a sort of hashy, grain. Try disconnecting all connections to the computer and try an optical cable only, see if that helps