The saga continues ...
I feel I have to write this down merely for myself, already because I found myself confused this morning of what I actually observed last night. So while I still can dig it up from my small brains I better write it down now.
So, started out where I left off day before yesterday with that SFS=120 and Q1 x xQ1 x DevBuf = 14 x 10 x 4096.
While nothing disturbed really, I found the bass to be vague. Or more vague than I am used to thus want it to be;
Placebo working told me that adding a USB disk should not be for the better, no matter I find it to be that myself. "Wrong approach". So I decided to try to get the sound similar to USB disk connected but without USB disk connected. Some kind of XXHighEnd settings possibly could be found. But where to begin, and how to have the apples as much apples as possible for comparison.
I decided to continue with the exact same PC and OS for a next step, and this is an XXHighEnd PC *not* setup for the LAN connection. All in there is a 2.5" OS spinning disk and day before yesterday I played music from that same connected USB disk, though thus actively used;
I copied some albums from this disk to the OS disk and played some tracks from there.
Not much difference was to be heard; good. Then it is time to pull the USB disk.
Again not much difference to be heard. Hmm ...
See ? this is already interesting, because apparently now other tweaks are at work to let vanish the all so obvious difference. So what can that be ? only Q1 x etc. and the SFS. I mean, I didn't change anything else to get where I wanted to be (hence current situation), so all it would need is (slowly) going back to the old situation - Q1 x xQ1 x DevBuf = 14 x 1 x 4096 at an SFS of 4.
But what is slowly ? I can't spend ages on this ...
Anyway notice how I always do this : Bring up XXHighEnd, press Pause, change settings and press Play again. So, no real A-B which I hate anyway, and just continue whatever you were doing and unconsciously observe. Have a next beer etc.
It was in my mind that this few days back I had set the SFS to 60. Why ? yes, right; this was the setting of JonP who spend tremendous amounts of times (something like well over a year ago) to come to the conclusion that SFS=60 was the one which really worked. Many of us used that at the time. But the other settings ? I don't know anymore. And besides, that will have been W7. But still. So SFS=60 would be the first single thing to change.
Right away I noticed what I was missing : air. Yep, when this suddenly lacks this goes quite unnoticable - I see now.
Lately I walk to the speakers 2-3 times per track and feel the woofers. So, I made myself aquainted with how what sound feels. Notice that this is how I found "strangeness" in the first place, which was with Windows 8.1 and the higher SFS to begin with. Reminder, also for myself : I started doing this because I found the "Windows 8 nastyness" went away with it. Also to remember : this requires the higher Q1 etc. setting.
I thus found myself lucky to have dialed in a setting (SFS=60) which was for the far better right away. Ha !
But the balance with the highs did not seem right. Oh, it was all OK in the "does not disturb" department, but I thought there could be more detail. Lowering the SFS I did not want, so up to the Q1 stuff ...
This went more fast than I could have dreamt;
I felt that 14 x 1 x 4096 would not work (because to "equal" to what I used all the past weeks), so the first I tried was 14 x 2 x 4096.
Aha, see ? more detail. But more of it was needed. So I tried the 14 x 1 x 4096 after all to proove some things. And hey, YES, there is the nastyness right back.
Now, since 14 x 2 = 28 and 14 x 1 = 14, I figured that 21 x 1 would be right in the middle of that. I even did that without machinery.
Spot on.
Bass did not significantly change that I noticed and detail was there in the amount I did not want more. But what I learned from this - and notice that this can well be SFS=60 specific - is that the Q1 stuff has some threshold somewhere. Maybe it is 22, maybe it is 20 but for now that 21 seems to be just it.
For the next, let's not forget that the Device Buffer Size of 4096 is actually just an arbitrary number. It just weights in the formula of Q1 (x xQ1 x DevBufSize). But also, in my (NOS1) situation the 4096 is not related to any buffer size of the NOS1 as such, and or but that physical buffer size is set to 16ms (NOS1 Driver Control Panel). How they really relate I don't even know, but maybe I should ...
But now this :
Knowing that I'm a conspiracy thinker - meaning that I like to reason towards the unprobable - I recall from a far past that the maximum Q1 "stand alone" for USB setting was 21 (above that no sound at all). Hmm ...
This was before the NOS1-USB and it was from the time that WASAPI (Engine#3) was the only Sound Engine available. No Device Buffer Setting as such was available because this was (still is) calculated under the hood and don't ask me what the result of that would be for the NOS1-USB. Or for whatever USB interface from back at the time or today.
For WASAPI the Q1 of 21 doesn't really relate to how it is done in Kernel Streaming (by me) but still there is a relation making it worth while to conspiracy-think.
Both feet back on the ground I should say that this situation should imply a similar noise pattern as with the USB disk connected. But, at least this is tweakable by everybody the same way (if the remainder of the setup is the same, which obviously can be done - and not so with a USB disk which is different for everybody). However :
Before I forget it myself : What was lacking with the USB disk was the air in the music (but with the notice that these Orelino speakers can so well express this - nothing of the kind with my old speakers).
So, net better anyway.
This was the first step in eliminating that noise-adding USB disk. The second step is applying this to the LAN setup which does not play from the OS disk. Notice that both situations (will) go through the RAMDisk as Playback Drive (and XXHighEnd origine). So no real differences to be expected although we know that everything matters.
And on the latter I should not forget that this is still Windows 8.1 which can also do a couple of things, although I will be saying in advance that I won't be able to get hold of those particular merits because I now play with "Windows 8" in a normal fashion, not emphasizing the strangenesess we know of W8.
Peter
