As in the subject, I am running Roon build 1357 into NWB into Meridian 818v3 into Stax SRM717. They connection is by SPDIF at the moment. Roon is doing no processing, nor is the NWB downsampling, NWB volume is at 0dB so (I guess) it should be a bit-perfect signal going to the 818v3.
There is light distortion on peaks which are not even very peaky. When I turn the volume down on the NWB (by 3dB) the distortion goes away. And then if I crank up the volume on the 717 the distortion does NOT return.
The latter should prove it is not my beloved headphones doing it.
So that leaves either the input of the SRM 717 distorting, or the output of the 818v3 distorting, for whatever reason. The SRM 717 can handle more than 4V rms without distorting, and this unit has just been restored by Spritzer, so hopefully it’s not already faulty. And it seems an odd problem for an 818v3 to get.
Could it be that the NWB is sending an overly hot PCM stream? Is 0db on the volume dial in Mosaic going to be bit perfect. or do I already have a misconception there? I found this posting, which would seem to point to this being the problem:
If dCS ever make such a bridge again, I really hope it has a display!
DCS do make a bridge which comes with a display The bad news is the Vivaldi Upsampler isn’t the cheapest of replacement options!
More seriously, the first thing I’d check is all the connections on your Stax headphones.
I saw a similar issue when the cleaners had pulled mine loose a while ago. The headphones sounded dreadful. If this doesn’t fix it, if possible, see if you can borrow another electrostatic amp and/or Stax headphones (the ‘and’ route should rule in or out either the output card or the input card on the 818v3 being at fault).
My gut feeling is that it shouldn’t be the NWB. If, for the sake of argument, you have a 16 or 24 bit output from the NWB, surely you can’t go to a number above those 16/24 bits because you only have that number of bits available (unless you have a carry overflow situation) - and the DAC shouldn’t distort even when all the bits are set to 1?
[I’m working on the basis that the DAC scales its maximum output voltage to the highest value each sample can store]
It might be worth checking whether there is some consistency between different types of files: MQA vs non-MQA or DSD vs DXD, which might indicate a compatibility issue, but I’d check the above first
Sounds like a problem with intersample overs In which case the problem has to do with the oversampling that goes on in the 818v3 before D/A conversion where some samples are clipped as the analog converted level would exceed 0dBFS.
If the 818v3 doesn’t have the headroom to address intersample overs properly, then adjusting the digital source signal level down either in Roon or on your Network Bridge would be the only solution I believe.
I think DSP for the volume control on the 818v3 (for its analogue outputs) is carried out after increasing the bit depth, before the DAC and the output stage. Are you saying you suspect the 818 allows for intersample overs and refuses to increase the volume if it knows it is likely to clip?
Well, if the problem has to do with intersample overs, ie. the interpolated samples exceeding 0dBFS, then by lowering the digital source level, say, by 3dB for example, you’re giving the upsampler/interpolator an extra 3dB headroom before any digital clipping happens.
Whatever digital signal processing that happens after that shouldn’t matter. Same with the volume control, especially when it’s in the analogue domain like on the SRM717.
I completely understand - and you make a very good point.
What I wasn’t so sure about is that Ludwig says, if he decreases the level, the problem goes away (which supports your point), however if he increases the level back to what it was, the problem doesn’t return.
Hence me wondering whether the connections to the electrostatic headphones might be causing an issue as this has happened to me with a similar result.
Thats right, it makes sense because he’s decreasing the source volume digitally, which presumably stops the intersample overs from occuring, and thereafter increasing the volume on the analogue side, so it remains distortion free.
I’m pretty sure the NWB is bit-transparent at 0dB. Otherwise MQA decoding on the 818v3 would fail. It would also fail if you lowered the NWB level by 3dB as well actually! So, this really needs to be solved within the 818v3 itself.
In fact, a quick search revealed the following on Stereophile’s Meridian UltraDAC review;
In Variable mode, the maximum output level at 1kHz with the volume control set to “87” was 6.54V from the balanced outputs, 3.23V from the unbalanced outputs. (Although the volume control goes up to “99,” the Meridian clips with full-scale data at settings higher than "87.”) With the Ultra DAC set to Fixed output mode, the maximum output level was 4.85V from the balanced outputs; ie, 2.6dB lower than with the volume control set to “87.” All subsequent measurements were taken in Variable mode with the volume control set to “87” or “86.”
I bet the 818v3 is the same . So you definitely want to test changing the 818v3’s volume level, either set it to fixed, or if variable, then set it to 87 or lower. In any case, you might want post about this issue on Meridian Unplugged and see if anyone else has experienced something similar.
By the way, the 818v3’s specification suggests it already has an Ethernet streaming port? Why not connect that up directly to Roon instead of using the NWB?