Hi Gerner,

Okay, this one first : Re: 0.9d vs u/i 0.9j.

Then your "desktop" approach ... hmm ... yes, it could be done (if you'd allow me to do that for Engine#3 only (yes you do, I know)).

This very approach in fact is already in there, and the only thing it needs is the "controls" again ("again" because I said such a thing before on another subject, i.e. many many more things are already in there, anticipating on "some" future, and it is just that they cannot by changed or set by the user yet).
It would allow the change of "Engine#3 engines" (hehe) as "realtime" as you can change the other stuff since 0.9j.

But I promise you, you will get crazy of all the options then.

Despite the above, I now think it would not be the best way to go about with things.
The 0.9j version from today can be changed to how 0.9d worked. The only problem with *that*, is that it would not be 100% the same because it just can not. This by itself is related to "the future stuff", e.g. gapless to name one (which was future for 0.9d ). However, the way the program was changed to avoid "redundancy" should make the difference in SQ, and that can be changed to just have the redundancy in (a bad thing by itself from a programmer's point of view, but when the cause is good ...).

What will come from it is a 0.9k with the character of 0.9d and the features of 0.9j as long as the Q1 slider is set to 14 (that's what the setting concerned was in 0.9d). Now, *if* there would be an unintended difference in SQ afterall, obviously this would follow from a comparison between 0.9d and 0.9k (at Q1 = 14). If there seems to be no difference in SQ there, there should be a difference between 0.9k and 0.9J.

Please note : What will happen with what I propose, is that a perceived better SQ performance from 0.9j (which started with 0.9h) is removed again, BUT, now there's the Q1 slider which can effectively be applied to the old 0.9d (that's what it would come to).
If *then* the conclusion for some is that 0.9j sounded better than 0.9k there's not much more left to do than create the "test version" I talked about in the previous post. That test version however, would be a normal fixed version, which then would contain an additional "quality option" A and B, A representing 0.9d behaviour, and B representing 0.9j behaviour.
But I'd really like to avoid that (because it would add *another* level of the same redundancy).

The only thing logically possible is make a testversion which acts like 0.9d but with the after that version added features (including Q1), and even add the option for 0.9d vs. 0.9j behaviour.
And then see further.

Ok. The additional reasoning then for "unaccording the filter type". Possibly it is more than lame, but remember, I just try to find explanations for what I hear.
As a sidenote to you Gerner : I'm back to the original filter and ran random tracks from random albums for over three hours, and none of them sounded even slightly bad.

In the earlier post I started with the bass example, but did not come back to it because it did not seem necessary. As you seem to point out (because I can't fight it oterwise) it might be necessary to come back to what I had in my mind really. Remember, this is about the bass vs. midrange and I can already tell you it might not apply to midrange vs. tweeter. But it's to get the idea.

Jitter causes more squaryness. This by itself would be audible, but never mind that.
A bass driver would have more difficulties with that than a midrange driver. Remember, a voltage jump with near infinit rise time is already hard to be followed by electronics, let alone a diaphragm. And of course, the more weight the diaphragm has, the more difficult it will be to follow the steep rise.

Digital, followed by a nos-DAC, will indeed feed that to the drivers. So, with a nos-DAC we're asking for trouble.

Indeed it would be so that an unnecessary (jitter) high voltage rise would apply to both drivers. But the more it is in the upper range of the "low" driver, the more it would be a problem for that driver. Otoh, the more it would be in the lower range of the "high" driver, the more easy it would be.
So there's a contradiction. Both would not be "time aligned" for that matter.

Where the bass driver presents this distortion from not being able to follow, the implied more squarish will result in just less afterall.
Where the mid driver is fed with the same at the same time, I propose that it does not show this distortion (or less anyway), and it produces a more "square" in comparison with the bass driver. So it will be the mid side of the slope that rises and fills the gap towards the left.

Since both drivers do not produce the same in the xover area, there is no xover.

But during this writing, it springs to my mind that there is more;
While jitter incurs for more squaryness, it can be present in the music just the same. So synthesizer music would destroy xovers ... (???).

Sorry for the blahblah, but prove me wrong ...
From a theoretical view I will be right, in practice I can't tell. But it would be an explanation for what happens ...

Peter

Hi Gerner,

This digital domain is not separated from the analogue domain. What happens in digital, will 1:1 influence analogue.

As I tried to explain, jitter implies squaryness. In its simples form this implies more energy. When this energy occurs on the slope of the xover (either side), it rises, and shifts to the "inside" of the xover. At that point it is destroyed.


Where I had my answer ready yesterday, in fact you give me a hard time doing it. Oh, I have one, but in order to apply it to the filter you have in mind (I know) I'd need additional reasoning. I have that, but something else came up which is related to the filter I use, and where I hear "things". In order not to let this too much offtopic, just a question for now : What filter types would you think this applies to (can apply to, act as one driver) ? Or do you say "all" ?

Actually, things are rather imcompareable. You said that first (with 0.9d vs. 0.9h). But if you have a good low range woofer (like near 25Hz) or a subwoofer, in that area the differences seem to be best noticeable. 0.9d gives more tight bass than 0.9j, and therefore 0.9j is more prone to standing waves again (but, also in the higher areas). It may sound stupid from a listening point of view, but a not tight bass in that low range should be felt on the driver. You feel the wave and hear the vibes of it in the mean time, or you feel wobbling and hear "low sound".

In the higher regions the main differences can be heard in the timbre. A track like Saling to Philadelphia (Mark Knopfler), played at a higher level, makes you feel the cracking of his voice in the stomache with 0.9d, or just not with 0.9j.
Now, if you'd know that this timbre thing for 95% is determined by the speaker's performance, I already don't know what is best in absolute sense, because a. a speaker can overdo timbre so easily, and b. I don't know the real timbre of a voice.

What I tend to do from a technical point of view is judging tight bass (0.9d) as better than not tight bass or judging less resonating at very difficult base squarish sounds (0.9j) as better than more resonating. Both conflict ...

*That* both conflict seems typical for jittery influence (was this proven anywhere ?) and the one jitter signature makes the bass right but highs not, and the other way around. But, it is far more difficult IMO;

When jitter destroys the bass, it keeps on bugging you all over;
When jitter destroys the highs, it depends. I tend to (start to) believe that jitter actually "destroying" the higher frequencies (say above 5K) is not so much different from what is actually happening by nature already. Destortion from the driver(s), reflections and all that happens wrongly in the analogue domain. So, that natural destroying might be worse than jitter in that area. BUT :

Disclaimer : I don't take scientific or not readings on the internet for granted (if I read them at all), and these are all my own thoughts. So just take it as total BS for now, Maybe later that changes.

When jitter influences the xover area of two drivers, it is a catastrophe. Of course it depends on the phase alignment that was achieved with the speaker for starters, but assumed the creator of it did his best to make it optimal, jitter will shift the connection between the two drivers.
In order to understand this better, let me first try to explain what I only could explain myself at last on the influence of jitter on the base 2 days ago :

First of all, I "proved" before that jitter makes a fluent sine more squarish; When samples are skipped, a volume level of one frequency stays the same too long, and when then the next sample is respected for its actual volume, the smooth step in between is lost and there is a too high jump to the next volume level. Now :
Knowing that the bass is a long wave phenomenon, I could never understand how this long slow moving wave could be "attacked" and destroyed by jitter, while at the same time it is so obvious that jitter impacts the tightness of the bass so heavily. Well, what came up to me two days ago, is that the squareryness which will emerge in that slow wave just the same, can't be followed by the relative slow woofer. It will try to make the jump, and when it at last is moving, it will even overshoot (depening on the control of the amp). Here is the wobbling bass, and while you'd still hear low frequency sound, it is undefined to the original vibe of the string etc., and actually it will be higher pitched than the original.

By kind of coincidence this occurred to me a few weeks ago at 0.9h without then knowing what happened, at Hatfield's End and Stonehenge 4, which "suddenly" showed a crazy volume increase in the sub low area, which I measured at a difference of 26dB (!) compared with the normal frequencies. I knew the record enough to know that this wasn't there before. And, since the normal level I played was at 90dB, the 116dB coming from the (rather continues on that album) low frequencies made it difficult to breath (this is serious).
Where I after this occurrence reasoned out the influence of jitter on the bass as explained above - went back to 0.9d and already noticed the difference on other tracks, I played the album again "knowing" what to expect, and indeed the crazy uplevelled low frequencies were gone.
It was another pure coincidence that my son went out of the room during playing this album, and came back within a few seconds to scream about the creapy things hapening in the hall; it now appeared that all doors were rambling, which I know they do when a good volume level of 15-16Hz is fed to the house. So, without eloborating too much further, 0.9j (already 0.9h would be the same on that matter), had shifted up the 16Hz to somewhere in the 23-25 range.
In addition, because of this jitter distortion a sine becomes squarish, additional energy is added, which may be (partly) the explanation for the 26dB increase of SPL.

Now back to the xover IMO being destroyed by jitter, let's assume we now "know" that jitter can shift a certain more or less steady tone, like that from a bass guitar. And, keep in mind that a bagpipe is not much different, though in an another frequency area (and composed of far more complex harmonics, but never mind that now);
What you'd see is that there will be an unnatural peak in areas of the frequency band, that actually *are* peaks with increased volume level. This is not only because of more squarishness and increased energy, but merely because they add up to the already "playing" tone in that area. Thus, where all "instruments" were playing together creating the homogene sound, now some frequencies add up at another level, and it sure was not mixed like that. The whole point (IMO) is, that when such a peak emerges at the slopes of the xover area (say left side), the carefully calibrated phase alignment of the two drivers will be completely destroyed. Because of the peak at the left side the roll off will shift to the right, and at the xover point there will be a peak now, and right from the xover point there will be a complete cancellout because of the antiphase of the both drivers will just do that. For sound this means a peak with right next to it a dip, and where this might cause four subsequent piano tones to jump (right to left) from normal to too high to nothing to normal, the actual effect for a gliding tone IMO is even worse;
E.g. a trombone gliding from low to high will show a stepping in the increased volume, followed by a giant step to nothing, followed by a step to normal. Effectively this is a squarish sound again.
But now take an instrument that as sibilance from itself, and the sibilance is created by harmonics, the "harmonics" emerging by inteference in air (and so recorded). This high frequency sibilance in the xover area will come to you as a very squarish sound, and it sounds as plain destortion.

Whether the above explanation for things is total BS or not, it is not different from all the same stuff I can just emulate. Thus, I can play with the xover, and I can play with jitter (XX) at the same time. Things do change largely, and the only thing I can say is that the explanation above I tried to find for it, just fits things.


The real message is, and this was in between the lines of the first post here, that what I hear on this matter (and it is of MAJOR influence), hardly can ever be perceived in the same manner in "your" system(s). We could never trade experiences and track examples, because "you" xover(s) will be at a different place and may be of another type.
As said before, in my system things seem to get profound after making the xover(s) theoretically better.

All together, my own judgements on versions of XX become kind of worthless for that.

For those who break their brains and ears with Q1 sliders and all ... better go back to 0.9d.

The Q1 slider was supposed to be a replacement for what changed from the versions after 0.9d, but I am kind of definite it can't be done like that. The Q1 influences sound allright, but it can't bring back what 0.9d had.

What happened might be too complex to explain and it might even be too early to be able to.
Anyway from a technical point of view the versions after 0.9d were expected to sound different, coming from a necessary change in the program in order to be able to cope with the future. And for those who want to know : Although the Q1 slider with its setting of 14 was equal to "a" setting accordingly in 0.9d, the program had still changed, and changed the sound for its own matter. The slider cannot change that ...

A bit of a background story :

As I told elsewhere, the improvement of XXHighEnd caused and causes now audible anomalies in the other parts of the systems "we" use, that by itself implying improvements of the system parts concerned. And, since these parts now are under the control of people in my direct neighbourhood (the speakers, the amps, the DAC), it appeared that a latest change to the speakers -which was from the theoretical view good- would not work as intended. With 0.9d it does though, and it is just this that makes me say that 0.9d must be better.

In the topic ver .9D vs .9H from SeVeReD it was actually anounced, and although I never listened back myself until yesterday, I was near sure SeVeReD was right.

We must be careful, because things start to be so so delicate, that it starts to become nearly impossible to state that what I hear on my system, will be expressed in the same manner in yours. Otoh, so far it always worked, no matter what amps or speakers "you" have.
Mind you, as I introduced with 0.9i and the Q1 slider, we now operate at the level of resonances. Resonances in the diaphragm material (as far as I can tell) that can be over-expressed in front loaded horn speakers, and that influence the xover area largely. This latter may be a kind of premature, but as far as I can reason out myself at this time, this implies a huge impact to sound quality.

For now I can indicate one phenomenon going wrong with the versions above 0.9d u/i 0.9j : too much dynamics.
I think few people will admit that dynamics can be too high, but it really can and causes the most strange anomalies with it, which (so far) cannot be logically explained. Too much dynamics itself already is kind of impossible to explain, and for later reference : possibly this implies just *better* playback in the base, but physical materials might not be able to cope and things get over-expressed (like ringing in an amp, or a diaphragm which can't follow).

So far for now.
Peter

Thanks both.

The *32 won't be there because it's compiled for both. But since (external) DLLs won't be recompiled, there might of will be one "old" 32 that's touched during the process. There are hundreds being called ... all not mine. 
So, IMO there is a chance that "enable 32 bit Applications" does the trick. Did you try that ?
If that fails, I could compile the lot for 32 bit only and see whether that helps. Awkward, but it may get me on track somewhere.

The only unsigned driver I would used and which is addressed soon after - or during startup, would be "your" primary sound device, or the device you lastly used at a succeeded earlier startup (which can't be the case hehe).

The MS page won't load here at this moment. But I will look into it.

Important : When you are a first time Vista user of XXHighEnd, do NOT use this version to set your reference of the sound quality XXHighEnd is capable of. Use version 0.9d instead ! OTOH, please read this topic as well : 0.9d vs u/i 0.9j .
It puts things in a perspective, and it might well be that afterall 0.9j sounds better for you.


Changes :

• Engine#3 now allows Pause/Continue.
• Engine#3 now allows the dragging of the position slider.
  Please note that this repositions with an accuracy of one second.
  When the slider is put near the end (last few seconds), accuracy will be extra off, playback may stop or restart at the beginning of the track.
• All settings logically requirering it can now be set during playback with immediate effect.
  A small dropout in sound will occur though, and the playback continues at a position rounded to the second.

Bug fix :
It only now appeared that changing the Absolute Phase (Invert) did not take effect at the start of the next track. It required stop/start. This is now solved (and takes effect immediately).

No changes on sound quality are to be expected with this version. However, do note that an earlier remark about Doubling not being the best idea from off version 0.9h actually seems legit again from off version 0.9i with the "proper" Q1 setting ("proper" is up to you).
It looks like Doubling has gotten a bit back from what it was created for : more dynamics. Please decide for yourself.

Another thing :

IIRC offline it was tested to remove the four microsoft.directx.* DLLs. But now I think of it, removing the other 3 msvcm* DLLs might just give better chances. Keep in mind though, that this DLLs come with the "install" just to avoid the problems with Vista systems which do not contain them ...

Even more important might be the contents of the microsoft.vc80.crt.manifest file. I mean, it contains references like processorarchitecture="x86" which sounds all but good to me. This manifest file is only there to coorporate with the msvcm* DLLs, so the manifest file can go (away) too.

Now, if all with Vista64 out there could try this ...
If your Vista64 coincidentally needs those DLLs back in I'm sure another kind of message will popup.

Btw, if the "solution" from my previous post helps first, it just might be because of the manifest file told the system to "be" a 32 bit application.

The below is an attempt to let "a" part which is not suitable for 64 bits run in its 32bit mode. The whole program then will btw, but I don't expect SQ issues from that.

In the mean time I will look further and try to find that part.
So, somewhere viaviavia (most probably) a DLL is used (this would be external software) which just was not made for 64.

The text that goes with it (which I just copied from somewhere) is this :

Open up the advanced settings of the AppPool you're running in and notice the option "Enable 32 bit Applications".

If that is set to Yes, all may run (as a temporary solution).

Yes Dave, I could agree to everything, but I have a few remarks though :

Yesterday I spent 1.5 hours on a reply to you in your Foobar topic, and when I was near finished I wanted to try something with Foobar which caused my system hanging and I lost my post. Wasted time, and I couldn't find the guts to start all over again.
Today I spent a lousy 14 hours on an improperly working xover, and bothered two other people with it during that same time. I had to switch off/on my Fireface to solve the xover problem ...

With this I only want to say that people may for whatever always good reason feel confident with an OS. It won't be me wasting their time by forcing "an" OS. So yes, I hate it when it comes to that anyway. I will solve it, once I stop waisting time on ...
computers.

Hi John,

Currently it is more a matter of just not having a machine with Vista/64 and it is far from easy to create one.
Do you actually "work" with Vista ? IMHO there is no way of getting things running in a normal way. Mind you, I said "work", not play audio because that's okay.
To give you an indication, it took me a full week of downloading the development stuff it takes, without knowing in advance what to get. If things don't work, there is no way to find out why. And please remember, what I do, so far, nobody is doing, so Google will only bring my own posts.

The signature of what happens with 64 is that something in the code causes the compiled code to crash immediately on 64. There is no way to debug because it doesn't even start. This implies that (for me) there's nothing else left than finding someone with the same situation, but who solved it.
As I said before, the gaming world has enough problems with it to kind of "know" it is unsolveable by normal means.
Otoh, all bugs (assuming that) can be worked around, but *now* time comes in. I mean, the trial and error it will take will be very time consuming.
Anyway, I'll see what I can do.


On the plugin thing : can't be done because too many things must be arranged by the, say, "preloading" of stuff. It's not only about passing through audio samples that creates audio quality ...

Thank you for offering your help John,
Peter

You are right about everything;
As long as the longer cable gives you a connection, there's no problem. And *if* there is a problem, you'd just get errors. This is a normal "data" connection you are talking about, different from SPDIF ...

Btw (unrelated to your subject), for my XP machine I use a run of 10 meters Firewire towards the reclocking (Fireface) soundcard. It seems to "grey" the sound a tad, but not more than Firewire interfaces differ amongst eachother. IOW a good Firewire interface card is more important than the length of a cable (the worser interface being less able to control the longer run of cable).

One other thing :
Putting the HDDs in the toilet might be a nice gadget. But before you do, maybe you could open the cabinet they're in (I don't know anymore what you have), shut down the fan in there, leave the cabinet open and keep track of the temperature of the disks. Nowadays disks just can have it and you must be with your ears on them to hear them. Mine (4) are right in the open stacked onto eachother without (air) space in between them ...

Hehe, it took me a second read to realize that your daughter wasn't even in the room itself ...
But you know, it is really one of the tricks to find out whether things are in good balance : go outside and listen there. Like in the garden with doors open; when thinks sound the same as if you were inside things are really good. When you hear too much bass or treble, things are not good. Never mind high frequencies have more trouble getting around corners.

In general : one should not compare on the sweetspot (only). Many more things are unveiled outside it. Example : the width or how much the sound comes from the speaker or not. Just be in a straight line opposite the e.g. right speaker (normal listening position would be the most extreme). Now listen how much you hear from the left speaker. When things are not good, you'd hear the right speaker only. When things are good, you'd hear both just the same and sound keeps coming from the middle. Of course the right speaker will sound louder because it is closer to you but there still will be a balance. Btw, even with (front)horn speakers this can be done.