We all fell in the W7 pitfall

[PeterSt]

Because I saw someone writing about it (can't find it anymore, and possibly was in another topic), here's something I start to "learn" and which seems to be a rule :

The more aggressive, the more forward;
The more smooth, the more laid back.

Be careful, because "laid back" is (I think) the same as "in the back", while "forward" is literally more towards you; more in the room.

Over time I have learned to like the forward better, because it just fills the room better. It is more interesting. More 3D. But now it looks like that it is impossible to combine this with "the least distortion". This is not only the result from the W7 to Vista story, but happens to anything I do with the DAC, another amplifier, LS cables. It is always the same : the more forward it is, the more aggressive the sound will be. Too much in your face. Too much horn (which latter is two folded : horn instruments come too much forward, but horn speakers start to sound like horns).

Now, we could say that "laid back" is "in the background", and I think the term will be used for that normally. But NOW we need a new phenomenon : layers (not really new of course, but we now need it in the equation).

Actually by coincidence we "discovered" the layers, as them bing there with 16/44.1 8x Arc Prediction "filtered", while the same take, but in 24/192 or 24/96 does not contain any layers at all. Just nothing. And yes, this is about the applied filtering to the music data, at creating the files ... I am pretty 100% sure suddenly about that. It just sounds so similar to an OS DAC.

And now there is the difference between laid back and plain flat. Or maybe better : between the opposite of forward in the room (hence much more in between the speakers, if not behind them) and flat; Flat is always between the speakers (not even behind them), while the "laid back with layers" is just enjoyable because the depth is still there. Actually, the depth is there even better compared with forward, because forward is more messy in the plane from where you listen to where the speakers are. It's more fatigueing as well. More difficult to concentrate.

So, all I actually wanted to say is that changing from W7/SSD to Vista/RAMDisk for 100% sure will give you the huge change from rather (or very) forward, to very much laid back, but with the micro detail allowing the layers. But but but ...
This is in my system, and I am pretty sure with any random OS DAC you will loose those layers. This may (or will ?) leave you with a more flat sound, and you realy must wonder whether this is better than the aggresiveness from W7/SSD, which just will be more 3D compared to Vista/RAMDisk. So, be careful with your decision, and try to concentrate on the "all being dead" thing. Smooth but dead ? no good IMO.

Of course, decide for yourself, but now don't fall into a pitfall of smoothness which actually is deadness. Just watch for it, since I now I told/warned you.
Peter

[AUDIODIDAKT]

THANK YOU Nick, much appr.

You will get feedback from me and many others,
BUT, I will first get used to Vista and RAMDisk a little bit more, then apply yours and Marcins tweaks.

Please people take the time for this all, do it step by step, so you can experience it for yourself,
and like Peter said: "don't just follow the crowd".
Even Peter himself made that mistake (this will not happen again !), it was Peter who started this topic afterall,
and we should have listen to him, WAY back in december 2009.
Windows7 ? hmm ... maybe not !

Respect the fact that Nick and Marcin spended a lot of time on it, so they DO deserve PROPER feedback people.

Roy

[Nick]

Marcin,

GREAT !!!!

I think the MMCS regedit is another fundamental component of the playback platform

I was thinking I had a case of the placebos, so it would be good hear your detailed on the change. For me it changes almost everything in the sound and takes music very big step closer to vinyl.

Ps don't take the Mulimedia Class Scheduler MMCS service out, its this service that implements the registry values that you are listening to 

Nick.

[AUDIODIDAKT]

Can we please make another topic on registry settings and such.
With the purpose of ending that topic, with some .reg files.
And ofcouse improvement in sound, AGAIN.
 

Its fital to find the information that is shared fast.

PS: the same counts for the vLite thing that we got going.

Thank you

[PeterSt]

I don't know if you have had time to digest the underlying concept behind the registry tweak I posted above. This could explain at least part of the Vista W7 difference. It seems in Vista Multimedia Scheduling Service DOES guarantee CPU time slots to audio. This feature has been dropped in W7 - so NO guarantee of CPU time slice length in W7 !

Hi Nick,

Thank you very much for stressing the importance of this, and I sure think something will come from it.
I have always been aware of these settings, including the W7 change. However, I recall playing with this (or wanting to do it), but didn't see the sense because I kind of work around it in the first place, and (from off) the new Vista Audio stack these things are arranged for automatically. So, say overruled.

BUT (don't laugh, don't cry) ...

What I am only thinking of right during making this post is ...

and (from off) the new Vista Audio stack these things are arranged for automatically

... this counts for WASAPI ... which we actually don't use anymore.
So for Kernel Streaming (I now think) this really should matter, be it Vista or W7, but for W7 probably more. IOW, Kernel Streaming (never officially supported, but now officially dropped although still there) for 100% sure won't know about audio priorities. And *now* the Registry Settings will be important. But still ...
At this moment I can't be sure whether these settings will ever influence. So, what is "audio", and how does the system know "audio" is occurring ? Will it be supported by KS in the first place ? And thus *again* I have the feeling not much can be done with it, untill the opposite has been proven (didn't Marcin just do that in his last post ? -> I'm not sure ... Marcin ?).

In either case, I want you to know that I digest everything, and moreover try to combine it with my experiences so far. Things can get rather complicated, and usually one such an element doesn't bring me much (to proceed on) just because I may have things against it (like what I just said above), and I usually can't find the time to play around with whatever it is. So, if you only look at 0.9z-3 which I want to put up nearly every single day for over two months now, you can imagine that I can't play around in between. But if it has been proven, then it's no playing around anymore. Then I dive into it instantly. Ok, still when time permits.
Also, I usually need the time to interpret things by means of stupid listening. And just for fun :

What was my impression about W7 ? (Dec. 2009)
What did I repeat again last week ?
What did Josef find out now ?

You see ? what Josef came up with, was just what I could hear, and with some "PC knowledge" I can kind of predict what will be happening. But, I *never* could find the time to sit back and even thing about a tool (+ proper settings) to investigate it. But as soon as someone else does it ... here I am.

This post is a bit vague to you perhaps, but what I actually want to say is : now it is known (or proven) that this "task swicthing" whatever thing actually is there - hence in between each other lione of coding so to say 1000s of other lines of OS code is processed which I sure doen't want, we now must get that out of the way first. It just deserves all the priority, because it has such a heavy impact. And yes, one of those Registry Settings may even be able to "undo" this crazy stuff, but I still like to learn why it happens in the first place. For example, why is this influenced by some stupid SFS parameter which does NOTHING. But does to a super high degree. This *is not* about I/Os. It is only about using more or less memory, and that by itself does something. Something which to my best knowledge shouldn't do anything at all, unless it is about internal memory mapping (like old DOS days could use more memory by means of indirect addressing. Is it still in there ?).

Additionally, and almost off topic, I had the great "opportunity" of not being able for the life of me to get my Asus Extreme-EVO whatever with i7 running without super distortion. I just gave up on it. In the mean time though, it has been the greatest opportunity to just clearly listen to what happens with what SFS because it just influences. With 12 ? there is no way I can get sound other than trrrrrrrr. With 70 ? MMMV but ALWAYS after 4 minutes the trrrr started to jump in, ending in only distortion a minute or two later. With 220 ? chances were very very small I could come through those first minutes. Why ? why why. It's just the same thing (I'm sure), but this time without any escape. Yes, I better spend a week to find out, but for me myself and I, I better exhange the lot to something which works. OTOH, you will be buying that same Mobo/processor, and you'll end up with the same.
And W7/Vista was not a single difference, except for the processor speed in W7 being constant at 40%. Is *that* good then ? I don't think so. But Vista ? that for 100% sure is wrong, the speed always being at max, just because the SFS is large. So, load the whole track into memory, and suddenly something thinks it must go at work. With what ? sure not those I/Os. Doh, sequentially search for that memory address. That is what it looks like. The more memory, the longer it takes.

Ok. Summed up, suddenly so many things are going on, that I can't tell head from tail. In the mean time we *do* improve significantly, but I promise you it will really happen when I understand what is happening, and how to really eliminate it. Just give it some time.
But also : keep on finding out yourselves, because that is the REAL thing which helps. Time time time.

Thanks,
Peter

[PeterSt]

Oh boy ... Ehm, for Nick, a small follow up on my previous post, after following the link from Roy to my own Dec. 2009 topic once again, this time actually reading what I wrore in there :

One of the major flaws I discovered (but it's a general thing I think) is that "media" (hence audio) doesn't perceive the highest priority anymore. Thus, when a piece of kernel code is supposed to fill a buffer while another piece of user code is doing something else, that user code now has preference. STUPID ! This is just where Vista was good at, and now it is destroyed. I don't say this can't be solved, but in the mean time it now *has* to be solved, where Vista was ok with it. So thank you very much.

You see ? I knew that too (by emperical finding).
But also : I forget things some times.

The heat of the battle ...

[PeterSt]

Can we please make another topic on registry settings and such.
With the purpose of ending that topic, with some .reg files.
And ofcouse improvement in sound, AGAIN.
 

Its fital to find the information that is shared fast.

PS: the same counts for the vLite thing that we got going.

Thank you

Hi there Roy,

It may sound counterintuitive to you, but I rather keep this "mess of things" in this one topic. I already am sorry that I didn't "join" the RAMDisk topic from Mani into here;
It is all about the same purpose, and already know I have difficulties in what topic I read which.

In general you are very right, but please not this time (no, you are still right, but I prefer to have all in one topic for now).

Thanks,
Peter


(I may change my mind later )

[PeterSt]

All,

I thought of it before, and now think I really should mention it :

While we're debating about W7 vs. Vista - or may think by now that Vista is the solution - pure from theory W2008 should be the better solution;
It is Vista, it is more lean, it is more decent/stable, it is more towards "server" activities (hence less fancy stuff), *and* those who tried it, were ever so enthusiast; I didn't do it, but I guess a search for "2008" may come up with some posts about it.

I am sorry if this comes as additional confusement to you, but better late than sorry.

Peter

[AUDIODIDAKT]

OK, by me  

If all can be find just here, its FINE and works for me too !

Its gonna be a VERY nice topic then

[PeterSt]

I may still merge these two topics. Just look at the last few posts in the RAMDisk topic. It's just all about the same thing ...
But I don't know how it will look like (the merging will take place on date/time).

Btw, this sure is not Mani's fault, who only created the topic for its specific subject, which IMO is justified (but nobody responded to, so far).

[manisandher]

Peter, please merge whatever makes sense.

Why do all the really interesting developments come when I'm away from home and can't get involved? FYI, I'll be away from home a lot in the next few weeks - SQ should improve in strides

Mani.

[Telstar]

Thanks guys. At this moment (some 10 hours of subsequent listening and trying back and forth) we are as far as that the RAMDisk with Vista brings the least "distortion".

I had to do a second test last night to be sure of that and, it's true. Peter is right once more. Vista and ramdisk is the way to go. The differences on a resolving system like Peter's are huge, from fatigue to enjoyment.

It remains to be checked vista32 vs 64bit (hope it doesnt) because I think we need the latter with 8+gig of ram to be able to load enough tracks. I'm already thinking 16gbs but I didnt check yet if my current mainboard supports that.

[Telstar]

Did somebody make a direct comparison between Vista/Ramdisk and W7/Ramdisk already?

(I like W7 more then Vista...)

Yes, we (me, Peter, Roy) did, twice. Unfortunately Vista wins hands down.

[AUDIODIDAKT]

Music should NEVER sound agressive, music should sound enjoyable ALWAYS, even with lesser sounding albums !

For now, and maybe never, w7 is NOT the OS you want to use for high-end audio playback.
When experience this thru a NON-filtering DAC like the NOS1, this becomes very obvious, I couldnt stand w7 for 5 minutes.
While enjoying Vista for hours and hours, but also only with RAMdisk. (very important)

Take a Piano part with high notes in it, being striked pretty hard, ..................its so obvious.

[AUDIODIDAKT]

Also a tip for people who still think W7 sounds better, should maybe raise the db in their tweeters, by adjusting the resistors.

I modded many industry loudspeakers, including several expensive onces and its very common that industry lowers te output of the tweeter just a little too much
just to get that smooth sound as being reffered to (this for hiding flaws in their own audio systems), you got no details for sure,
otherwise it could never be smooth sounding.

Just beware of this....................