Hi pedal,
(btw, is there any importance in the picture underneath your text ? I see a read cross only ...)
The answer is not all that simple (for understanding) I'm afraid;
Let me first say that
if you do not want to use the Digital Volume from XX for your own reasons, for 44.1/16 playback it doesn't matter at all.
The point is in the
if though ...
I determined that applying the digital volume is better than the analogue volume. Never mind why for now, but I reasoned it out en so far it looks ok.
Additionally this - by accident - allows you to get rid of the preamp. And, as this is a major smooth-operator
holding back transients, creating noise and all, it slowly becomes a matter of "needing" the Digital Volume to just have a better sound quality.
Btw, that a preamp holds back and therefore better is to be avoided is notthing new, but -so far- we couldn't really see how to do it right with the ditgital volume and really too few bits available for it. Now, what is applied in consistency in XX is that with 16bits input
and 16 bits output all is - or just is not sufficiently legit.
Btw, keep in mind that with 16 bits in (say, the CD) and 16 bits out (your DAC as it sadly seems) everybody on this planet will say STOP !
And when nothing "special" is applied at digitally attenuating, they are very much right ...
And mind you, all who apply official digital attenuating (like in the pro world (which includes RME and the like) will do so by means of internal 40 bit or more float processing, and XX does nothing of the kind ...
With this I only want to say that we're way off the usual tracks here, and my own theories including the practice as of now, tell me it just can be done. Ok, just, or just not;
As I said before, the just / just not is highly related to the gain you apply and the sensitivity of the loudspeakers;
At first, when I couldn't exploit my own 18 bits of the TwinDAC+ (XX could not do that yet) I determined that as long as I kept above 48dB attenuation (so, towards 42, 36) all was okay. But mind you, this was absolute listening without the possibility to compare the 16 bits DAC (useage) to the 18 bit DAC. I thought I could hear anomalies at 42dB, but I took them for granted NET. Together with that I would never play that soft, and it's merely -30dB or -24dB. At -30dB too, I again I thought I couldd hear the anomalies, and again I took them for granted, and again NET.
Please keep in mind that there is a "balance" between more attenuation and the SQ getting worse on one side, and the less attenuation and less distortion (that's what it plainly comes to), or IOW the louder you play the better you can hear distortion obviously, BUT the distortion by itself gets less. At 0dB there's no distortion left on these matters ...
With the repeated "NET" I refer to the overall improvement, and *if* I can hear distortion somewhere a. I'm not even sure it is "that" distortion I hear and b. because of a. and the overall better result, I don't care.
Mind you, this is all with the 16 bit DAC.
With the 18 bit DAC though, the distortion (if audible at all at the given level)
officially comes forward 12dB "later". Thus, when with the 16bit DAC audible distortion is there at -42dB, with 18 bits the very same distortion would be there at -56dB. Now here the mentioned balance comes in again, because when you normally play at -30dB, there is no way there's audible sound at all at -56dB. Try it (not with your ear in the speaker, but at normal listening distance).
To make a long story short, I am sure that I hear the difference between the 16bit (used) DAC and the 18 bit DAC both at -30dB. Also, when you A-B this, it becomes more apparent what it actually is you hear with the 16 bit DAC and which you took for granted. It's a kind of roughness that I can *not* really explain from missing bits as such, but of which I know by experience that they are needed to fill out "holes" at certain frequency levels (or better : from certain instruments). This is similar to how I eliminated "roughness" by just adding gain to the highest frequencies in my loudspeaker filter (which lineairly goes up from 5K to 20K by 16dB !).
With 18 bits I'm satisfied at -30dB, but maybe with 20 bits I'm more satisfied ...
Suppose that 18 bits would be the "standard" for legitimally play back at -30dB, then from that follows (it's just math) that with 16 bits all is as legit at -18dB.
When you are not able to play at -18dB (because it's too loud) the relatively simple solution is to adjust the gain of your amp. NOT by means of attenuating the existing gain !
The latter comes down to using the digital attenuation in the first place. So when you just don't use that, there's really no problem with 16 bits files that a more than 16 bit DAC can solve. Not that I know of.
Doubling or Upsampling to 88.2 KHz was created for better SQ (mainly Doubling !) but if it works out like that ... YMMV. For me it does not, because it needs a larger buffer in the soundcard (from 48 to 96 samples), and -again- NET it is not better because of that.
Please keep in mind, all of the above is related to 44.1/16 files only.
Once it would be a standard to have 96/24 etc. files, the "problems" are obvious, because you just won't be able to play them.
I'm not sure if you'd need any advise on bringing back your DAC to April Music, but if you spent that amount of money with the idea it could do 96/24 ... well ... The proof that you could be mislead is sufficiently available in this topic.
. Along with that, the person involved not really to blame for it as well.
For the future : when the specs of a DAC do not specifically mention input samplerates and bits, better assume 48/16 as max. For that matter the CD10 does mention 96/24 for input explicitly. Whether *this* then is true (over USB !) I don't know.
I hope this is something for an answer pedal.
Peter
