All,
First a few general, boring things.
It has become clear that this "channel integration" thing really is THE thing to hunt for. When I wrote about this the first time I was only used to it two days I think, so that was extasy all over. But actually -at running more and more albums- it is an amazing thing you just haven't heard *anywhere*. Yes, I'm pretty sure about that.
This too excels with certain types of music, and since this is all about separation - think about King's Singers. I listened too Good Vibrations of them the other day, and really thought it was the most beautiful I EVER heard.
Further (but this is really the last one) improvement has been made on the sound in general, which partly came from the lot being in its cabinet now (I perceived somewhat more body from that). I must add to this that this time I used the analyser to squeeze down every bit of noise which is in there, and what comes from the DAC can not be measured anymore (at the end of the intterlink) from an FFT. IOW, this is beyond -144dB, as far as the analyser goes, which thus shows -144dB. And I must say, I never heard completely *no noise at all* from my horn speakers (which as you may know by now are 115dB sensitive, and with the amps at full gain of 27dB).
The larger deal tough came from the DAC and gain stage now being "integrated" by means of two "click connect" PCBs, while the proto of the gain stage was connected with soldered wires and two connectors. And yes, that all matters. This time it brought the snare drum on Six Blade Knife (Dire Straits) to more reality, while it was sounding like firm cardboard before (what one can hunt for
).
And this brings me to the volume control ...
Let me (again nag) first that this just took over 2 weeks of wasted time. Wasted time to the sense of that it has to go (hence fit) in the cabinet (of which I'm producing a 100 at first) or not. I mean, I can't produce the definite cabinet when I don't know what will be in there.
So as I told the other time, I immediately ordered one of which I thought it would be good (I recall that was on a Thursday) and heard nothing. The next Sunday though I received an email it would be shipped the next Tuesday. And so I received it last Wednesday. A week later than could have been.
The last two days I have been working on it, to finally come to the conclusion that it can't work not being in the signal path. This is about adjusting a resistance somewhere which should be on a tracelength of 3mm, while here I can't do it otherwise than over a length of 20cm. So, distortion all over, and it didn't even attenuate.
At wrapping all up yesterday, I thought to at least check what would happen if I would use it as intended. Thus, in the signal path as any normal attenuator;
I have been measuring the whole afternoon, which merely was the result from not seeing bad things at all. Oh, I saw a decrease of almost a factor two on harmonic distortion, but you must know that isn't all *that* bad, while we're at -105dB, and this factor two implies 6 dB (thus now -99).
The only thing I could find as strange was the slowly increasing noise level beyond the audio band, but since this is inaudible for both the frequency (25KHz and up) and the level (-130dB), I actually could not "see" where it would be wrong. Not in hours of time.
And so, at listening time the volume control was in the chain, and I ran a couple of tracks that I'm familiar with.
Now, with the explicit notice that all is relative, and that I'm used to what I'm used to (which is, say, a far too high standard), I could hear immediately that it didn't "work". The width of the soundstage seems compressed (yea, must find some new phenomena sometimes haha), bass doesn't go deep at all, but worse, is coloured, I sensed some standing wave behaviour in the low areas and for sure perceived standing (buzzing) waves in the higher regions. It was mushy. It didn't fit anymore. Piano notes shouted. Te music seemed too loud to be comfortable.
I imagined "you" listening to it, and I imagined that you still could say "wow", not knowing better. But, it would be critical. It would be critical to "the best DAC ever" and I would not dare to bet on it.
At this moment I can't decide;
*If* it would be in there, there would always be a direct path to the (different set of) outputs as well. But heck, why to do it in the first place ?
1. To comfortably attenuate the sound ?
Maybe, but in the very end non-sense. When talking about other players it is not much of a problem to begin with, and when talking about XXHighEnd a lot can be done to make it more comfortable. It can be made more fast (responding), it can be made more direct (like choosing the number of steps to in/decrease). But no, there won't be a physical knob. So what.
2. To attenuate with smaller steps
Yes, but this is very relative. This one works with 64 1dB steps, and using the knob isn't even comfortable, knowing that you need 10 full cycles or so of the knob to go from 64 to 0. I didn't try it with the remote, but imagine it takes ages to find "your" volume, because the steps can't be addressed directly, and it takes a "wait time" when the up/down key is continuesly pressed to start moving, and after that it will go too fast to stop where you want, or will go too slow to be comfortable.
Allright, where this is 1dB, XX works with 1.5dB. Well, I tell you that the annoying thing at the moment is those small steps when you want larger. It really really never occurred to me that I couldn't find "my" volume in those 1.5dB steps. So ... if you really think you need smaller steps (like Telstar) better think twice if this may be nagging, while in reality you are after the best sound quality. So, if I *tell* you this is nagging, maybe it helps ! haha
3. To protect from static
Is this only stupid theory ? I mean, when we play loud with our well matched amps, we play near full volume, and the static will be as bad as without the volume.
Of course this is different when we use 700W amps on speakers who need 50 only, so we only talk about that (and these people sure will exist, so why not take it as a legit reason).
For me it is clear that we just need an inherent solution for this. This is not a panic-mute button (you won't be able to press it from shock) but something which detects it and just cuts the signal. In theory this can be done in-DAC, but I think it will be too complicated to have a program running in there which can "see" this, like the Crack Detection in XXHighEnd can. The Crack Detection in XXHighEnd by itself works allright, weren't it that it is hooked in the chain too soon. I mean, after the Crack Detection process there's more which can mangle the data, so theoretically it can be done later. However, while it can be there at the latest at filling the buffers towards the OS and driver, there's always the driver at latest which can do things wrong. Not that we need to expect it will do things wrong (I never encountered that except for one report of a user with an explanation which goes in that direction), but we can imply it ourselves so much. Thus, when we squeeze down the latency so much that we aren't able to pass on the bytes in the proper sequence - meaning just skip one - we'd have it. On this matter, notice that an audio sample comprises of more bytes, and if one is left out you'd have static. This too I never encountered, which may imply that the OS (or driver ?) is very much responsible for keeping these things together, so if we really could rely on *that*, then I could guarantee it in the software.
But then this is only *my* software, while the DAC won't prevent you from using Foobar etc. ...
This brings me to the last subject for today : the inputs.
As it looks now, it won't be possible to use the available digital input (SPDIF) for measuring means. This never has been a subject much, but for me it is a kind of important. This is in the area of people "testing" the DAC (could be a review), while without such an input there is not much to test. The input is physically there allright, but it implies such a bad "stream" that this itself masks all the good stuff from the DAC. Most probably this is because the SPDIF input will route via the OS, or better : I can't get the SPDIF input passing through to i2s (which I use), unless I switch on the "Listen to this" in the device's properties. So, while the both are available on the same board (think like a sound card), the OS apparently is used for the routing, and nothing than harmonic distortion comes from that (like over 30%).
I didn't get around yet to testing the analogue inputs. But I need to add something to this :
I guess I made a thinking error on combining two "connection" situations where either at this moment won't allow analogue inputs to work (but the combination does while it can't exist as combination). The one situation needs a space in the cabinet which isn't there because the PCB with the clocks occupies it (and which I won't move farther away because it should be as close to the DAC as possible), while the other situation will be in the PC, but it will have a too jittery clock connection because the clocks are in the DAC cabinet. And please keep in mind : this "inputs" thing was about "your measuring" (so not really recording as such) which by itself needs the same clock for playback and the recording (of that playback).
Hmm ... something to sort out. But maybe I'll leave it be; It ever was a nice "withgoing" feauture, but if it's not "withgoing", then bye.
Anyone there who needs 400 Neutrik connectors ? you can also use them for outputs ! hahaha
Ok, I'll see what I can make of it.
First the Crack Detection I'd say.
Peter
