Review MSB Premier with Director – Silence of the bits

Measurements MSB Premier

We measured the MSB Premier dac with Director. Most of that was done with the Prism dScope III. We also captured impulse behavior via the scope. Now the dScope III is an insanely fine device, but it too has limits. And unfortunately, but simply true: we saw that we reached the limit of the dScope in the linearity measurement.

Whether the noise floor can go any further is also questionable. -165 dB is very low… but perhaps the MSB can go even lower. We don’t know. But we did find ourselves scratching behind the ears at times…. is this going ok?

We wanted to measure the clock internally, but that’s not easy to do. MSB uses a differential clock setup. We have a differential probe, but it is not good enough for this kind of measurement (it is intended for amplifiers). We have – unfortunately! – had to refrain from checking MSB’s claim that the jitter of their clock is in the femtoseconds. Fortunately, we were able to see other things.

Imaging (digital filter)

Frequency response

Impulse

Phase response

Noise & Interference

Jitter – THD – Dynamic range

Okay. Just a little clarification. Let’s start by saying that MSB has seen all the measurements and provided feedback here and there. A device like the MSB is not easy to measure. Everything counts. Really everything…

Let’s start by reporting that the MSB’s noise floor is unprecedentedly low. 10dB lower than the Pasithea even! A true achievement! Anything below 100 Hz is cable influence, we estimate. This because we see a similar pattern with the Pasithea. Also a very quiet dac with -155dB noise floor.

Remarkably – thanks to this low noise floor – we see the 48 kHz clock back in the measurements. That pulse is still below -130dB. Realize that.

Looking at the filter, we find a brickwall filter in normal mode. In theater mode, it is slightly different, but also particularly steep. This is reflected in the impulse behavior, which “rings” a little less in theatre mode. We see subtle differences with the Director in between.

Then the phase behavior of the dac. This is also different in video mode than in regular mode. The phase in video mode is between 3 and 6 degrees, which is very good. So the processing really does something. And that is audible, of course. What you like better, we leave to you. Realize that phase is all over the place in your listening room (reflections, absorption, speaker placement, etc.). Not to mention the effects on the recording side. So don’t jump to conclusions here. We measure it because it gives interesting insights into the filters and effects of DSPs.

The Jtest – which is totally unreliable, but can give an indication – shows around 3.5 – 4 ps of jitter. Very good for a Jtest, but as we said, it is not a reliable test. The only really good way to measure jitter is with a dedicated device – our Wavecrest for example – and then also directly on the clock, or via a BNC / SMA clock-output. Unfortunately, we don’t have a differential probe that can measure at this level. Otherwise we would certainly have done it.

Finally, something about dynamics and channel separation. These are all extremely neat. Which is not surprising, considering the low noise floor and neat construction of this unit.

The Director is supposed to bring calmness and blackness through extra isolation from the inputs and the dac. It’s audible when you add it. And we also measure a slightly lower noise floor. Especially towards the end of the spectrum. There we see without the director also a peak at 48 khz (wordclock). That is gone with the director.

Limit of the measurement setup

We estimate that we have reached the limit of the Prism dScope III with the MSB dac. A noise floor of -165dB is just unprecedentedly good. This shows that MSB really knows what it is doing and pays attention to every detail. There are few manufacturers that emulate them in this regard. And realize that there are two more models above this Premier…. we just don’t dare ask for those anymore honestly.