No. Why ? your system is just fine. If you'd only reconize that the cracks are just there !
Two topics in this area :

1. You'll get a message when the cracks are there, and this message is justified;
2. With the last versions (the ones with the Mem checkbox) all is just more accurate again.

I too perceive the "cracks" far more than before. Btw, they are not cracks, but ticks.
I have told you about what I found in your example files, and this is just not right. Or it is, and you should perceive those ticks.
Strangely enough it listens like vinyl very much, although I hardly can imagine this is the intention for those particular albums.


This is another matter. Firstly, you are not alone, and secondly I myself am unclear on what I think of the versions after 0.9u-8.

Also, think of how hard it is : I start to hear transients which just are in the data (them being wrong by itself IMHO), and now I must think this is worse for accurate playback ? that's a tough one ...
But also : who says that those transients should lead to ticks ? can't my DAC follow ? and, will the oversampling DAC smooth them out ? (hmm ... I should try that for fun). Officially not, because this is not a matter of too steep transients. They are just plain vertical in the digital data, and go over a range of 60000 within the max of 65536. There is nothing to smooth out, although the analogue part will round a bit at the start and end of it.

On the other side, we must think of what is happening really; we try to follow the digital data, which actually just *is* about small squares. Do we even want to follow those ? I say yes, that is, up to now that is what I tried to achieve. But there may be a stage that is over the top. With 44.1KHz anyway ...

As I said earlier, I will build in the choice for the 0.9u-8 way of working vs. the later versions. It is very easy and it won't deteriorate either way.

Hi Rolf,

I think your observations are right.

As far as I reasoned it, the lower values indeed should sound more digital, but this is (should be) because of more accuracy.
When indeed you perceive a more digital sound, this can be fought back by means of the Core Appointment settings. For me clearly setting 3 does that. Beware, because the Core Appointment settings will not workout the same over systems.

With Q1 higher, hence assumed less accuracy, you will get a more liguid sound. But with less detail, and bass may get more fumbling because of it. "Liquid" is not the best description because for me that goes along with the most accuracy and e.g. the most square sound from synthesizers. It finishes off though, things start to fit.
YMMV because when the accuracy - that going along with needed speed of e.g. amps (and speakers) - cannot be followed, you'll get ringing and destortion by it. In that case it is better to have Q1 higher, so things can follow and *then* it's more liquid.

With an oversampling DAC you can hardly incur for following squares from synths, because the os DAC already will destroy that into sines, but at the highest frequencies it may work out better (because in the nos DAC they are just too square (a 22050 HZ sine just *is* a plain square in the digital wave)).

All together it is rather complex, and you have to know your system in order to know what you can - but also should achieve.
If you want to know what your amps can do according to speed, move your pre-amp out of the way (but be careful !), and try what you think of that. If it is clearly worse (too harsh) you'd know that you shouldn't aim for more accuracy, and better go for the more rounding.

In each occasion, try it with UnAttended playback. Only that is under my control for SQ, and only there you can have a reference yourself for later (or earlier) versions. Also, UnAttended is more accurate by itself (or in a controlled fashion anyway).

Peter

Hmm ... I don't recall that I said that the lower the Q1 the "better" jitter becomes. Besides that it isn't proved by all people.
I did say though, that as far as I can recognize, the soundcard's buffer should be as low as possible. So maybe you have that in mind ?

The upper range of Q1 as well as the lower range has its limits. These limits are actually there right now, although I must honestly say that the upper limit is related to somethings that is factullay unknown (by me), but together with that it is my reasoning that it shouldn't matter anymore above the upper limit from right now. One exception exists to this : that one exact Q1 number your DAC may respond upon for upper limits. This though, will always be too critical and too much depending on the performance of the PC itself. Too keep it simple : the DAC's buffer will underrun at this one special number.

Ok, despite all I still could increase the upper limit, just as I could increase the lower limit. The point is though, that increasing the upper limit goes along with a "resolution" that never suffices. Think of an logarithmic scale, and one step above the current limit would be a double sized one. What I said above still counts though, and it implies a necessary scale of something like 700 steps above the current 30, to reach 31. Relate this to the current 35 steps, and you'll see the unuseable thing coming from it.

Much more easy, would be a checkbox to double the resolution within one chosen number. Thus, when you have chosen e.g. 28, you then can choose from 27.5 to 28.5 in a 35 step resolution.
I always intended to make this, but it never came to that because other improvements caused more change to SQ. As a matter of fact, it stopped me from playing with Q1, while actually that should be done at each version that claims improvement on SQ. But as you will know now, it just takes too much time to do that after each upgrade (for me anyway !).

But I keep it in mind. It has been quite some time that someone mentioned explicit testing about the Q1 anyway.
And mind you, this will make me to try 28. One never knows ...

Thanks.

Correction : I said "Library data", but I meant "Galery data".

Meta Data (the extension is not official, and just used by me).
In there the Library data is stored.

One mta file is a representative for related music data, like a wav, but also a jpg.
In fact, anything that's stored in an album's folder, will become mta files, once put in a Library.

Yeah, I read that first this morning. Already forgot that this "story" was in a separate topic.


Even that didn't trigger my sleepy head.
Coffee was good though.

For the remainder, I guess it is too difficult for at least me to comprehend what you can exactly do with your settings, how your soundcard mixer should or should not influence your DAC, how the soundcard can play offline (I understand this) while mixer settings influence playback, combined with the soundcard slaved to the DAC but the DAC (D70) should produce the sound.


The only conclusion which seems logical to me would be that you use the soundcard as the DAC in this slaved situation (the P70 playing). But since you will not be doing this ... as said, I just can't comprehend.

   

I checked Upsample instead of Quad.

Sorry ...
Will solve it !

I just tried (without sound because I'd have to reroute cables for that) ... when you set the DAC to 96/24 it just works.
It does NOT work when the DAC is set to e.g. 192/24. This is a matter of the inappropriate error messages.

Note that with setting the DAC to 96/24 this does not imply you can play 96KHz only. It is merely a self protection from setting things wrongly, and in some occasions it's really doing something.
So now you can mis-use the 96/24 setting to allow for Quad anyway.


PS: Don't forget to set the FF to 128 samples for that.

SPDIF.
I can use USB as well, which is in the DAC converted to SPDIF before going to the chips, and it sounds way different (and far from better). Whether this is related to the FF reclocking I don't know.

Probably more things are going on, and reading the net it looks like more people come to the conclusion that USB is not Walhalla in those cases where a choice is possible.
Anyway, and as I said more often before, in almost all cases it is impossible to compare, because things just cannot be compared. For example, if the FF could take USB (in and out), I myself would already be much closer to an honest comparison.
Or, if my DAC would have I2S internally, and both SPDIF and USB would convert directly to that, the same.

I am fairly sure that - with enough people hence equipment - around, we could find the absolute best connection, and the connection will show a nature by itself. Actually this is what I created this board for.
But it really takes many more people.

It's already good that now ONE person is around who can slave his soundcard to the DAC. Didn't know that ...

Must think about this ...
If you, in the mean time, have more input to this, I'm very glad to get your opinions. Everybody would.
It is the most interesting anyway.

Peter

Right. That one I missed (in my context, not in your story).

To be honest, I don't know. Few people will be able to do that, and in here you are probably the only one doing that.
Must the answer be Yes then ? anyway, IMO you should do the test with a normal CD to avoid possible other anomalies.

Hmm ...

Maybe I misunderstand or misinterpret, but if this is about the comparison with (passing through) soundcard and without, why do you think there shouldn't be a difference ?
I'd say everybody knows that soundcards (at passing through) just make a difference. Try your mobo's chippy.

Do I misunderstand ?

I am sorry Leo, but I can't get what you are actually saying.
Maybe it isn't even necessary to make it more clear to me, because I just know something is wrong here. I will solve it.


On this occasion I don't know what you want to prove, apart from the obvious. Did you find the Release Notes I was referring to ?
This is from 0.9u-5 :


But, I obviously now should prevent you from being able to set this.


This is from 9.9u-8 :


Does it make better sense now ?

By my own experience it is true that for the Fireface SQ may be influenced by their drivers. Therefore I *DO NOT* change to the newest versions when they become available. If there's anything that sneakes into your system for the worse (hence without ever noticing it, and then never being able to find the culprit), it is this. IOW, once you are satisfied, don't change !

FWIW : I am using driver version 2.82, firmware Dec 6 2007.

Het Leo,

The Upsampling not working is unintentional of course, and what I get from others is that you should lower your DAC Is setting (to e.g. 96K). I didn't try that myself, but obviously the message is (still) tripping on wrong data.
I can ensure you that it works, if you only have the right settings (hence those are abusively anticipated upon).


What do you mean "don't start anymore" ?
Note that if you by accident switched the advised prio settings (like setting ThreadPrio to low) it is true that indeed Engine#3 won't start anymore. This changed a few versions back, and is intentionally (and mentioned in the Release Notes concerned).
Otherwise, yes it is true that the latest versions seem more demanding (which proves the good btw).

?