Highres music files sometimes don't play with XXHE and NOS1a

[Rmalits]

Hi Peter,

I have got a few highres music files which I often played with various network players without any problems, but with XXHE and NOS1a a very loud highfrequency noice is produced for about 1 sec and then it stops. Do you have an idea, what the reason could be for that?

Before sending you one of the few music files which did not play any more, because they produced horrible sound, I played again the one, which had this bad behaviour two days ago.
The result: Today everything is ok. Nice music is produced...

I had this problem a few times only and until today I avoided to to play these files again with XXHE.
It only happend with highres files like 96/24, not with RB files yet.

What could be the reason for this?

a) Maybe the loudness set on XXHE? I believe it was set to max, when this happend last time? Today I played it with -9.
I remember bad SQ once when I played a not well recorded RB files. It was kind of a pretty strong "knistern" (in German, english maybe:  "cracking" every second or so) when playing it with full volume on XXHE (it was easier to compare it to my Linn Klimax network player this way). With less volume there were no cracks any more.
What loudness would you suggest on XXHE when used with a preamp?

b) Could it be a hardware related problem, e.g. USB cable?

One more point:
Yesterday I again checked the DC offset of the NOS1a. When switching on this switch, one of the displays often shows -1 (on the very left), what's probably an error indication.
So I tried a couple of times to swich the NOS1a and the DC messurements off and on again until both displays showed appropriate values. And these values were good (only .1V off the values you emailed me).
Could this (-1 in the display after the first switching on) be an indication for some kind of hardware instability?

Kind regards
Richard

[PeterSt]

Hi Richard,

To start with the DC Offset : That should always be on par with the values we measured over here. If not - and especially when it is really over like your mentioned -1 (+1 the same), there won't be headroom for the output signal to fully develop, and I guess that could incur for the sound you describe (but hard to tell of course, unless you know that you played with that offset, which thus implies clipping of the signal).

But my thoughts go elsewhere, and now I'm asking others just the same ...

What I myself, by pure coincidence, noticed one single time only, but which was less than two weeks ago, was I'd say the same and it was with Hires. One thing : I could predict that something would go wrong right after hitting Play, because some other anomaly happened first;
In my case I selected a Hires track (I stopped playback right in advance of that (Unattended that was)), pressed play and ... heard the last few seconds of the previously playing (Redbook) track. I don't recall whether I heard those last few seconds again, but I think that was the case (read : the track played right through the end and did not "forget" its last few seconds to play). Now notice : this has to be about some internal buffer which was not freed or cleared or whatever what caused it to play again, and looking at those "few seconds", in my case this must have been about the Q1 which was at 14 x (Q1x) 10, thus 140. So what happened technically ?

The internals of PC + software + DAC ran into that last buffer of Redbook and at the start of Playback it thus sets all for 32/705600 (16x upsampling/filtering and with 16 bits as the base). However, the track I selected for real was a 24/96 (or 192, I forgot) and because everything and all selected upon those first few notes of music (the former track), it went wrong right after those few seconds (all being set for 16 bits base, now running into 24 bits base). That is, this is my reasoning and I can't be sure.
All I know is that I now play with the larger buffers for two months or so (also the SFS of 120) and that it may have been 3 years or more that I did not play with those larger buffers. And, looking at the way the clock (44.1 or 48 based) is selected (which originates from 2.5 years ago), possibly something is amiss with the larger buffer sizes (this is all quite complex). Anyway I never experienced this before.

A second time this would go fine anyway, because the clock (oscillator) in order would have been selected already (and yes I know about the tick again at any start of playback of 48 based material). So like it could have been your situation, Richard, my second attempt went fine. But I also know how very coincidentally things can go wrong, like the moment of you stopping playback while the track is at its end and the software itself also "stops" (and now strange things may happen).

So the clear questions :
a. Do others run into this, perhaps since "we" use the larger buffer sizes (again) ?
b. Does anyone recognize this last buffer (of the previous track) being played ? Notice that when the buffers are set small (like Q1 of 14x1 and/or the SFS of 0.2 etc.) that amount of last buffer plays for maybe 0.1 seconds only and will hardly be audible.

What strikes me is the resemblance with Richard's Hires, though with the notice that me playing Hires material is rather rare (I'd need to run into some Hires by accident).
So what I am actually looking for is dozens of people who experience(d) the same regularly or even once, so I can state definitely that something is not right in there (and solve it).

Thank you all and thank you Richard for posting,
Peter

[AlainGr]

Hi Peter,

I do have those "ticks" at the beginning of a song, and I also play a hi-res file between 2 redbook files from time to time, at the same settings as you have (q1 14, q3 0, q4 1, q5 0 and SFS 120). No recall of last notes playing from the previous redbook (before the hi-res). So I can't say a thing about the reported potential problem...

Alain

[Rmalits]

Hi Peter,

this morning I had a different kinds of strange DC offset messurements on my NOS1a: both values shown were about 150, like e.g. 149.5 or so.
Therefore I had to switch off and on my NOS1a a couple of times until it showed the correct values. There was no -1 today.
So will always check the DC offset right after switching on the NOS1a.

So my question is: Is there a maybe serious hardware problem coming up inside my NOS1a?

I was playing a lot highres files these days. No problem any more with horrible noise instead of music. All is fine now. But I don't play with full loadness, -7.5 or -9 instead.

When this happend the last time a few days ago, it was not a part of the previous song play what was played, it was just a very loud high frequency noise for about one second and then nothing more.

Kind regards
Richard

[PeterSt]

Hi Richard,

The DC Offset as you experience it, is all normal. I mean, that it does not lock in automaticaly is normal, and the procedure to follow is the proof of it (try until it's right and you will always get it right).
So indeed, randomly after switching on the gear (or switch off the PC or unplug the USB cable !) the offset may be your share and it is always to be corrected (thus also after shutting off the PC etc.).

But why not let all on always ? all is explicitly made for it ...

Best regards,
Peter

[Rmalits]

Hi Peter,

as I watch the DC offset more frequently now, I could see that the values change a bit more when the NOS1a is warmed up.

The values you messured and sent me in an email are: +01.5/-01.2mV
When starting the NOS1a from cold I get about these values, maybe 0.2 mV off.

When NOS1a is warmed up I got these values today : +0.5/-1.7mV. So there is a shift there a bit more than just 0.1 or 0.2mV.
The closer the values are towards 0 the better, would be my opinion about it. But the second value moves 0.5mV more away from 0.
Is it still ok like this?

To leave everything switch on always is not really practicable for me because I have got to much hifi gear, e.g. 3 DACs and so on...
And everything is fed by two 230V regenerators from PS Audio plugged together serially. For all that I have reserved one phase of the 230V power supply in my house. Due to this extensive setup I switch off everything, including the PS Audio regenerators when I don't listen music for more than a few hours.
I could plug the audio PC into another 230V phase and let it run for ever.
That could work without loss of SQ, because there is galvanic separation in the USB port of the NOS1a as I read in the forum.
But I would not do that with the NOS1a because I believe that clean power is very important for DACs.
What do you think about it?

Kind regards
Richard

[PeterSt]

Richard,

So there is a shift there a bit more than just 0.1 or 0.2mV.

You say this as if I told that it should be (or can be) 0.1 or 0.2mV only. But did I ?
In the email I sent with the here measured values I think I said (or at least intended to) that it easily can differ 1mV (in either direction) when things warm up. So that is 1mV, not 0.1mV.
Don't observe it when playback is going on, because that influences (very low frequency looks like DC changing).

The closer the values are towards 0 the better, would be my opinion about it.

Well, sort of. But with these levels it really does not matter. What would be good though is when the distance between the two is minimal (2.7mV in your case, shifting to 2.2mV as it seems). But anyway this is out of our control unless you want a capacitor in the signal path.

+ and - 20mV is regarded to be still OK (distance is 40mV now) although these are not my own laws.

Anyway, maybe read the initial email about it plus what it refers to. I just looked it up, and this is the text :

In the Guide you will also find how to cause the DC Offset of the NOS1a to be correct. We measured yours at +01.5/-01.2mV.  Notice that this is when cold and a few minutes after being under power it may change 1mV or so (+ or –).

OK ?

Best regards,
Peter

[PeterSt]

I could plug the audio PC into another 230V phase and let it run for ever.
That could work without loss of SQ, because there is galvanic separation in the USB port of the NOS1a as I read in the forum.

No, it will not be without loss of SQ as electrically it is worse (now you need to bride bridge DC Offset of the two separate (virtual) mains rings).

Also, I'm sure by now you make too much fuss about it, as you can easily check the DC Offset and get it right (by means of switch too, but replugging the USB cable will work even better (this is not in the guide, but taken from experience over time).
So just switch off everything if you like. But, always make the DC Offset right again.

Best regards,
Peter

[Rmalits]

Hi Peter,

thanks a lot for all these infos.
Yes, the tolerance you emailed me is 1mV and not 0.1. I was confusing that somehow.
So I will messure and set the DC offset every time I switch on the NOS1a and will not worry about it any more when it is in that range.
Btw, I listened a lot to highres music the last days and no bad noise any more since I always take care about the DC offset. Therefore I believe, the far off DC offset probably was the reason for the bad noise.

And I will keep the Audio PC on the same 230V phase as the other Hifi equipment.

Kind regards
Richard

[PeterSt]

OK Richard, very good ! Btw :

No, it will not be without loss of SQ as electrically it is worse (now you need to bride DC Offset of the two separate (virtual) mains rings).

Bride ? Bridge, I meant.

Peter

[Robert]

In my case I selected a Hires track (I stopped playback right in advance of that (Unattended that was)), pressed play and ... heard the last few seconds of the previously playing (Redbook) track. I don't recall whether I heard those last few seconds again, but I think that was the case (read : the track played right through the end and did not "forget" its last few seconds to play). Now notice : this has to be about some internal buffer which was not freed or cleared or whatever what caused it to play again, and looking at those "few seconds", in my case this must have been about the Q1 which was at 14 x (Q1x) 10, thus 140. So what happened technically ?

I have had this happen twice when going from 16/44 to 24/96/192 files.

I'm also having issues with some 24/96/192 tracks skipping mid song on some HiRes albums unattended. This does not occur with prior settings Q1 14x10 3,4,5, 1,1,1 SFS .12 Clock res 1ms or in attended mode. These hires files have been tested through another player and work fine. In prior settings some Hires albums only play one track then stop this is connected to low SFS(under 2) as higher SFS works fine.

Sometimes when listening to various single files(mixed 16/44, 24/96) in attended then going to unattended a box comes up which says Proxy not cleared but it goes on to play after ticking box away.

The recent new settings you have proposed do sound much better than the one I mention above.

Robert

[Robert]

My new SFS setting of 4 has mean't the return of these issues. Hires albums only playing 2 tracks and stopping with whole album loaded. Skipping mid tracks when playing a whole album 24/96/192 and whole album loaded. I have enlarged Ram for larger Hires albums.

I have put all my Bios settings back to default for max power. Has not made a difference. Core of 3-4 best sound.

I need to experiment with alternate SFS settings I guess. Any other suggestions welcome.

Robert

[PeterSt]

Hi Robert,

I think that there's coincidental chance that things don't work out at specific combinations. Without further explaining for now, what happens when you change that SFS of 4.00 into e.g. 4.10 or 3.90 ?
With this proposal I assume that higher SFS settings work, but lower as well (I see something of 0.12 in earlier posts ?).

Also notice that no 4.00 is supposed to be exactly that from my point of view (oh, it can work out for the best but then this is unintentional and at least not held against anything else close (like 4.10). So don't expect such a change to be devastating (but keep in touch with it).

Best regards,
Peter

[Robert]

Peter,
        I do need to try different settings for SFS I did not think of trying 4.1 or 3.9. I did try 160 and this was interesting.

In terms of settings should I leave Q settings as they are and just try SFS?

Are there any other settings that have an influence on SFS?

Robert

 

 

[PeterSt]

Oohh, Robert, difficult !

I don't think I have guidelines for combinations as you ask for. What I do know about my own trials lately, is that I sustain with the Q=14 and xQ1=10. So somehow that is always for the better (than xQ1=1), thus that is my base lately.
I also sneek in the convenience factor, because a high SFS implies high latency on volume change etc.

But if you find something interesting you will let us know, right ?

Regards,
Peter