Best input for MSB Reference DAC and K50

Hi,

I’m about to pull the trigger on K50 (no option to try before I buy but I have tried a K21 USB) and would like to know which is best way to connect to a MSB Reference DAC?

I have had good results with via USB (with ProUSB optical module) as well as with the Renderer (Ethernet), with a slight edge with USB. SPDIF/Coaxial has a really good analog flavor to it but lacks some resoolution compared to USB and doesnt allow for DSD files.

My current setup is NUC → USB → MSB DAC (with Roon) as well as DCS Network Bridge → Coaxial RCA → MSB Dac.

In my current setup AES (via the DCS) is my least favorite connection sound wise. I have not tried I2S but have good things so curious if that is the way to go with the K50? It’s the only MSB module I dont have so would need to purchase that one to try.

Thanks in advance for the advice!

1 Like

Hi there, TBH it is not an easy thing to say, as it depends :slight_smile:
On your ears, your room, your equipment, your source material.
For us here, I2S, AES and the standard non Pro-ISL USB input are all awesome.
Im listening right now to an OLADRA G4 via standard non Pro-ISL USB to the Reference and it sounds fantastic.
Honestly, you will need to test it to see which you prefer.

1 Like

Thanks a lot for your reply! Will do some testing once I received it.

Any technical reason why you choose to use non Pro-ISL USB input instead of the ISL?

Cheers!

Generally not a fan of Fibre Optic anything, it just adds another layer of conversion.

Is that Mark’s opinion as well? (the other Mark for those not knowing)

1 Like

Yes, that’s the Antipodes view.
It’s actually two extra layers of conversion.

We use an MSB Reference DAC and MSB S500 Amp into Marten Mingus Quintet 2’s as one of our references; it is an outstanding system that enables us to hear exactly what we have built and the changes that we make.
The Reference is a beautiful DAC, and any input is going to sound fabulous, tho if pressed, we would put USB, I2S, and AES ahead of PRO-ISL inputs on a Reference DAC. Pro-ISL is a great idea for a noisy source.
On a K50 G4, USB would be the preferred server output; on an OLADRA G4, I2S and USB are both options, followed by AES, as the digital outputs on an OLADRA G4 get an extra level of re-clocking compared to the digital outputs of a K50 G4.

We would contend that we build music servers that sound “right” and natural, while getting you closer to the music without the need for anything further.
That’s our view; obviously, there are a myriad of options available.
If you like what you hear, then it’s right for you.

5 Likes

I second Mark’s opinion here. A-B between pro-ISL and I2S with Oladra G4 and settled with I2S. For me, the I2S is more fluid / natural compared with pro-USB, I can’t say anything about normal USB as I don’t have this module. But again, trust your ear :slight_smile:

*Through Mark’s comment, I can see why Antipodes hasn’t come up with the native Pro-ISL solution like others do :stuck_out_tongue:

2 Likes

I use an Oladra into a MSB Reference.
I have tried USB, S/PDIF, a fibre-optic option.
But once I tried a ZenSati Silenzio AES/EBU cable, I never wanted to change it.
It sounds amazing.
I don’t know why it’s such a suck-it-and-see approach, but I am very happy with what it sounds like. Stunning.
I did not try this solution when I had the K50.

3 Likes

Very clear thank you! Will have to get myself an I2S module looks like…

Thanks for the input sirs

Do you have any recommendations on a good I2S cable please?

Also forgot to mention but the K50 I purchased is the G3 version (was ex-demo from a dealer)

yet MSB imposes it to connect Digital Director and Dac…

For me with nordost valhalla 2 cable the aes is really good

1 Like

Mark,

I am curious about the ‘extra’ clocking you mention the Oladra has versus the K50. My recent understanding is that the DAC provides the clocking as it receives the data from the server. If that’s the case, how does extra clocking in the server make an impact?

A bit long winded but this is the input I refer to.

Craig

Digital audio binary data, represented by ones and zeros, records the amplitude of sound samples. However, the integrity of the digital signal is not solely dependent on the accuracy of these amplitudes. Timing plays an equally vital role and the timing of the signal is created by the music server used. Imagine a turntable with high frequency wow and flutter distorting the signal’s timing and you begin to realise that the ones and zeros are only half of the picture.

In digital audio, unlike its analogue counterpart, timing inaccuracies can be reduced during playback. This is achieved through buffering and regenerating the digital signal, and synchronising it with a low phase noise clock, in a circuit that is well designed to maintain its stability. This process enhances the quality of the digital signal in the time dimension, and therefore it enhances the sound quality of the resulting analogue audio.

This characteristic of digital audio leads to a common misconception that as long as one possesses a high-quality DAC, with a high precision clock, then the quality of the music server is inconsequential. The rationale behind this misconception is that this DAC will be able to perfectly time the data just before the digital-to-analogue conversion.

Of course, there are individuals that are absolutely vehement that it is all just ones and zeroes, so a music server-streamer cannot make a difference. This is so easily dis-proven by listening that it is hard to understand how this fallacy persists. The writings of Neuroscientist Iain McGilchrist on left and right brain thinking give us a clue: “the left hemisphere sees truth as internal coherence of the system, not correspondence with the reality we experience.”; “Meaning emerges from engagement with the world, not from abstract contemplation of it.”; “The only certainty, it seems to me, is that those who believe they are certainly right are certainly wrong.”

Just because perfect digital is theoretically possible, does not mean that real-world circuits can so easily achieve a perfect, or even a good result. In practice, good results require excellent circuit design and multiple stages, amongst other real-world factors. In practice, achieving precise signal timing before the DAC always improves the resulting analogue audio quality.

We have previously used the analogy of cleaning a very dirty car. While it is theoretically possible that a single step might do a good job of cleaning the car, a multi-stage process of pressure cleaning, wiping-down with soap and water, rinsing, waxing and polishing, yields a far superior result in practice.

As an example, a clock stage is only as good as the circuit it is used in and the power supply that powers the circuit. And the re-clocking stage needs to receive a low-noise signal with high-bandwidth, before it can re-clock the signal accurately. If the signals before and after the re-clocking stage are polluted by noise below the bit-rate (such as from employing linear power supplies) then the re-clocking will be largely ineffective. Anyone that claims that digital is ‘just ones and zeroes’ has not applied their knowledge to demanding real-world problems.

The necessity for high-end audio equipment to perform the music server stage stems from at least three factors:

  1. Noise Reduction: Music servers can be engineered with sufficient power, specialised power supplies and circuit designs that isolate the digital signal from electrical noise. Noise, prevalent in general-purpose computers and networks, can significantly degrade audio quality.
  2. Isolation: Isolating processor-intensive digital signal processing from analogue signals is crucial. What will yield the best digital signal will interfere with analogue signals and vice versa.
  3. Dedicated Audio Optimisation: High-end music servers can be optimised for audio playback. This includes the use of high-quality components, optimised operating systems and software applications, and optimised management of services to avoid time-based distortion of the signal.

DACs play a pivotal role in digital audio playback, and so too do Music Servers. It is the quality of both the Music Server and the DAC that delivers the quality of the resulting analogue audio signal.

Referring back to the cleaning car analogy, the OLADRA has two steps, the K50 one step.

6 Likes

Thanks for taking the time to explain that Mark. It’s very helpful, and I appreciate the detail.

I follow the broader point that implementation before the DAC can matter in practice, even with bit-perfect data. Where I’m still trying to reconcile things is separating what’s truly “timing/clock” related versus what’s more about noise, isolation, and coupling effects into the DAC/analog stages.

Either way, I agree the real-world results are often more complex than the textbook “ones and zeros” argument suggests. Thanks again

1 Like

I agree completely with @MarkCole Its is very much system dependent and depends very much on whether you are comparing “apples to apples.” In other words, be careful to choose the best USB within budget with the best AE/EBU within budget. If you simply wish to purchase one choose the best you can afford - be it USB or AES/EBU. I connect my K22 player directly into my Lyngdorf MP-60 as it has a superb on-board DAC with the benefit of RoomPerfect and I require one fewer cable since I sold my 3rd party DAC. Cerious Tech builds amazing interconnects and their Lumniscate line is pretty special . I am using their AES/EBU currently and it is sublime.

my previous cable was AQ firebird; my current cable is RiCable SUPREME MK2

I know the price difference is huge and some might consider it a “downgrade” but I A-B both and prefer Ricable in my system. The price of RiCable makes it so easy to just keep it if I found it not to my liking but in the end I sold the AQ.

In my system it adds more fluidity and emotion to the music, which is a very great combination with Antipodes server and MSB DAC.

1 Like

I have a similar question but it concerns the Renderer when installed on Ref/Select and my Kala 50.
Is there a way to run files directly from K50 to MSB renderer with out going over the network?

Yes, use Direct Stream Output from the K50, lan cable directly from the K50 to the MSB.

Thank you! Happy Monday

1 Like