Before nothing happens anymore in here, here is my final verdict :
Will keep on using it.
It has been a bit of a long search (without any help of any of you - just saying
so careful), but what it requires is what has been my idea about it all along : the longest buffer sizes.
SFS=2 ?"Our" SFS of 2 is unlistenable. Well, all is relative of course, but there is too much detail and the overall dynamic range seems to be too high. So what I noticed is that it has become sheer impossible to listen to acceptable levels. Don't understand ? well, the distance of softest parts of a track to the loudest parts is too high. Partly this will be because all is beamed to sharply or something and louder passages become "nasty". Not distorting, but tireing because of too much detail, which *really* is the case to begin with. So, we have the examples of spitting women and tongue-clicking men, and this comes all the way forward. Not good.
SFS=430While the SFS of 2 will direct us automatically towards all the buffer sizes being small (at least with Phase Alignment this is so), with an SFS of 430 we can do all we want. But what is
needed here is the large buffer size. In my case I stopped at Q1=20 and xQ1=30 (can just as well be the other way around) at my Device Buffer Size of 4096. This is completely acceptable for Unattended Playback (not Attended), were it for no clicks and no stops (although a stop occcasionally happens when a track is finished).
Important : With 8GB of memory this can't work. With 12GB it will (I use 16GB).
A Clock Resolution of 10ms or 15ms is allowed - under 10ms is a no go again (for SQ merits). I use 15 because all towards the direction of "faster / shorter" creates the nastyness hence too much detail.
I never fooled around with XTweaks (yet) and they are of the same setting as my W7 install, except for the "Cool when Idle" which I don't trust under W8. So that is On (while the normal advice = Off). Load balance is at the lowest value possible (43).
My Windows 8 XTweak "Stable time" is On. Off did not work out for the better (difference is marginal but there).
The "soft cracks" I talked about earlier, now appear to come from the low buffer sizes. That is, I never heard one anymore;
It is to be kept in mind that
something is not right for W8 because it is too fragile on those ever same soft cracking (moving a screen can already cause it). But, this is when the OS is working on "your" task of audio. Not so with the all over higher settings - then totally nothing happens for tens of seconds (or for minutes with the lower sample rates).
So Peter, how about the sound now then ?I *had* to approach it from a completely different angle. So, from W7 most (up to almost me) agree that the higher buffer settings make the sound dull. This was certainly my idea about Windows 8 - even far more so. But it's a psychological thing, and that's why the other approach needed;
I think I will be able to talk you into this (lay down on my coach) :
* What comes from W8 is way too much detail. Can it be too much ? it certainly can; the wrong things will be emphasized and what actually happens is that the balance is not right (this is not the normal balance we talk about).
* So we dial in our dullest setting, and actually start off at the other end. This is a good idea, because we know that when we know go back, at some stage we will meet nastyness. But, what we look for is our detail. Not dullness.
* It should be good if you forbid yourself to start out with the lower buffer settings. Remember, it's a psychological thing and really *everybody* will judge the more detail as better. But you shouldn't even learn about it ...
* With the large buffer settings dialed in, you already will perceive that more detail. So, way more than with W7. So, actually you will find yourself satisfied right from the start with these settings. However, might you still run into nastyness, you can always go some higher (theoretically you will be able to use Q1=30 and xQ1=40 and SFS=470 or so, BUT you may run into troubles on the technical side).
* Of course we are all so that we want to see what can be achieved more in the detail department. So, you will now go look for that threshold where things start to get nasty. This is difficult because this *very* much depends on the music, and you should not dial in something for a cosy song and such, but for something of which it is known that it can go very wrong. Use The Beatles (Sgt Pepper would be a good one, NO remaster)- just because I say so.
You can use any parameter now, once you started off with the settings I told about above. Just depend on my statement that each parameter dialed down will bring more of that detail. That Windows 8 Stable Time is a special one, because it makes the attacks harder and only that. It seems to express hits on the cymbal better (edit : when it is set as not stable ! - remember, my chosen setting is "stable").
Right. For me none of the above dialing down worked out for the better, with the logical conclusion I should go higher instead. However, at this moment and after using my current settings for the last 4 days, I don't need that.
Hey, the sound, remember ?All right. No absolute judging, and only relative to Windows 7 SP1, and of course assuming the "dull" settings as described above (outside those settings my observations would change drastically) :
a. Way more detail;
b. Infinitely better channel separation (I can expect that this happens with the NOS1 only);
c. Interaction of harmonics which just wasn't there before at all.
d. A strange effect of the sound being farther away than before, while looking into it it's only way more forward.
e. Sound comes way more from outside of the speakers.
Ad c (Interaction of harmonics)
One of the first XXHighEnd versions showed that Mike Oldfield wasn't able to play on key at all. I should have checked that Tubular Bells by now, because I can expect this to be so again. Not sure though, because at the time that XXHE version was regarded "wrong". However, I'd say that when this 0.9z-8-1a version shows Mike Oldfield to be off key again, he just is.
The phenomenon as listed under c. is hard to describe in words; It exhibits as flanger (which is softer/louder change as a Hammond-Leslie exhibits) and unveils I'd say everywhere. Read back in this topic how I suddenly could see Neil Young being closer and farther from the microphone - something I never heard in the same tracks before. Notice that this is a way rough exhibit of this "flanger" phenomenon, so let alone the more detail this exhibits in. As I said - everywhere.
Along with this goes the vibrato (this is the change of key). So, where this wasn't the case at all in certain by me known tracks, now it is all over. Notice that I am not talkoing about vibrato in a voice or from e.g. a violin where this is an ever applied means, but the more "flutter" like from a turn table. So, when the revolutions of the turn table are not 100% the same (the speed changes somewhat) the effect of that would be the key slightly changing.
With the notice that this would theoretically lead to the audibility of "long term" jitter (which does exist !) it is merely like because of room interactions with the sound itself, doppler-like effects may happen. This is (my) theory only, but when the sound wave is pushed faster than the normal speed of sound, actually some doppler like thing would be happening, and a bass hit would be able to change the pitch of a high frequency bell.
Notice that -whether true at all- this is only partly because of our electronic means of music reproduction, because in the studio (when all playing together) or live, the same thing would happen.
If you combine all of the above, you suddenly end up with "music". Think like a drum computer needing random being slightly off of the beat to mimic human drumming, and this looks the same to me. And hey, just making up a few things, because it is new to me.
The "interaction of harmonics" I dedicated to this subject, is about the so-totally-clear buzz of for example two nylon strings can create (already one string can do that). It is like a zzzing which happens because the different harmonics come together at a certain stage of (decading) level and varying interaction with the guitar's cabinet. The "come together" means a slight frequency change of say one of the "harmonics" which now suddenly falls together with the harmonic of a different part of the instrument implying a sheer add up of the level at (of) that frequency.
I don't think I can describe it better, but these jump outs make things live. Real. It, btw, resembles my earlier expression about suddenly "seeing" how the neck of the guitar (etc. etc.) is held and manipulated, because *everything* is audible. Everything is relative too of course (to W7 in our case), but nothing of this kind was there.
Envision an album like Love from the Beatles (which I don't regard a remaster as such) which contains original tracks (surrounded by different lead-ins and -outs), that I played that album quite a few times, and that I of course heard the songs themselves hundreds of times - and that I did not recognize anything of it anymore. Like I was listening to a different mix.
Ad d (A strange effect of the sound being farther away than before, while looking into it it's only way more forward)
I don't know how this happens and never heard anything like this. It could be a wrong thing by itself, and it should be speaker positioning related. This is allowed of course, but I exactly never changed anything about that positioning for as long as I own these speakers (4-5 years ?). But, for W8 I already had to make a change (see earlier post), so maybe it needs more.
It is an effect of what the eyes see don't match with what the ears tell. Or that reality seems to violate physics. So, a few observations :
- I do not suffer anymore from not being able to listen from a(n actually too far) distance and no stage being there (hearing the closest speaker instead, also knowing that my being at that distance will be opposite to one of the speakers only). So, that all is OK now (it needed a more beam up of the mid-high). However, I hear the sound from where the speakers are, and not from "just in the room". If I *look* at the speakers, I see the sound coming from the plane where the speakers are.
- That channel separation remains the most obvious. I now think this just goes along with the above observation. It's mutually exclusive.
- Knowing that the sweetspot is some 4 meters more towards the speakers, how can it be that the SPL is higher at this larger distance. Yes, higher. We would say "oh Peter, this obviously is because of reflections just behind you at that end of the room you are". But this is not the case. I would hear that by various means.
And do notice that the phenomenon of "mystery feet" as how we call them by now, is able to create the exact same SPL everywhere in the room, which already can't be, but just is so. This looks like a similar trick.
- Also the most strange is that I have another listening position which is at 4 meters from the speakers and which is too close, knowing that the speakers are over 6 meters apart ... which just is perfectly allowed now. This doesn't feel like being too close. It works perfectly.
- Read the above ? Right. And *now* try to feel my discomfort about the sound being as forward as hell. Discomfort because I can't understand it.
So, sound comes from the place of the speakers, but is not layd back at all. The most contrary. And btw, this coincides with the low buffer values and the too high dynamic range which is perceived. A bass nicely plays, while a voice shouts right in front of your nose.
So, the latter not so with the high buffer values, but the least I would expect is everything jumping in my face being at the too short distance. Not so.
To wrap it up (finally), many of my observations are inconsistent or at least they are as long as I can't explain them. This doesn't prevent pointing out Windows 8 as the clear winner at this moment. At least one very important thing for me is consistent : my ideas about the large buffer sizes. The smaller are not to work out for the better but even I had to admit that under W7 they did. But if you compare certain tracks and perceived the total clearness from W7 and SFS=2 while under W8 and SFS=430 (etc.) I can only be clear about the the roughness of W7 (at the low SFS !). Clear and undistorted as it seems, but rough. Under W8 it compares as totally dull and 1 meter of blankets. But listen through it and hear the silkyness of the highs which no even 1cm blanket would let through.
Keep in mind "with certain tracks". Otherwise I never heard so much cymbal before. With the dull setting they carry still somewhat more metal than W7 SFS=2, which more metal IMO was needed.
Make it all small buffers and you will have a most clear signature to the sound, which is a metal one. And really, you won't last for one album then. Well, I won't, and I assume all NOS1 users won't; "We" are not used at all anymore to a signature to the sound. Have that back in an unavoidable fashion, and you will NOT carry forward with this hobby. Not anymore.
Well, let's see for how long I can keep up this story !
Peter
(not checked for typos)
