No, I really did not read your post !
So we both landed at 7 he !?

I'll respond to your post soon.

Huh ??
That would be the most strange ...

I forgot what Vista version you have ...

Hi Edward,

As it appeared, RealTime is just not possible in Vista. Not that I could find it. Maybe a registry tweak ? You can't do it by task manager and I can't do it by the program. In XP it can ...

"Nothing" means the program doesn't change it. So that's meant for external tweaking (like a startup parameter for the player (not the threads, that's not possible).

Not sure, but I don't think that is true. You must first start playing to see it ! IOW, it always reverts to normal when playing stops (or is not at order).

?

Theory became practice, that's the least I could say. Man, what have I done.

Usually I leave it to you out there to rave (or not) about something, but this time I happily do it myself;
If I counted right, this now is the fifth time in the year of development of XX that the improvement on sound quality is so huge, that it now is 5 times in another leage or however to say that (the first version was already better than CD players).

My Q1 slider is at 7, Absolute Phase is normal.

Okay, what were my theories hence what to achieve with this 0.9i version ? Eliminate horn resonance.
As it appears now (I am almost certain of that), the "horn resonance" as a property of front loaded horns might exist by itself, but as with more anomalies in life, you have to wake it up to let it bother you.

Horn resonance my *ss I say now, and obviously I must have thought that already when I started this little project last saturday (also see http://www.phasure.com/index.php?topic=159.0 which was the base for it).

A typicle characteristic of a front loaded horn is its relative enormeous efficiency (the ones I use are 115dB). Although I must be careful not to express things I know nothing about, one of the reasons a horn can be so efficient is because of its "beaming" characteristic. So, it beams the sound towards you instead of spreading it around, so all the energy goes towards you instead of to your walls. But there is a kind of drawback (and please note these ar MY words and not official, but anywone going against my words just has a disagreement with me) :

The mouth (actually the throat) of such a very efficient horn is very small. It depends on the horn of course, but mine are 2 inches diameter I think.
Depending on the frequency the horn is allowed to carry, such a small diameter is prone to various problems, thinking of the wavelength exceeding that diameter and blablabla, but my point with the "horn resonance" is about the reflections in that small throat.
The throat has a length, and throughout the that length the waves have to pass, not disturbing eachother. Well, I can tell you, there's even turbulence in there ...

Now I fall back to a by now more known explicit principle of mine : the standing waves. I do this on order to "visualize" what is happening really;

Those withg fine playback systems know what makes XXHighEnd benefit most : the low frequency (only those ? ) will disappear from the room. No matter how untreated and "lousy" your room is to this aspect ... they disappear. But do they really ?
No, they do not, because they still can be measured. But without XXHighEnd they are so much profound, that they make more music than your music. So there is a tolerance level. So with the "standing waves" inside the horn's mouth ...

One of the things happening in a horn, will be the "pushing" of the sound waves. Remember, they can't go anywhere but straight forward. In my horns anyway, there is a 6 inch (15cm) or so with paralell walls before the throat gets wider. Of course, the speed of sound remains the same always (I mean, it's not depending on frequency or volume (better : SPL)), but still we can imagine that the previous sent out wave is just ahead of the next, and theoretically there will be some "pushing". Compare this with a normal 180 degrees radiating driver (diaphragm) into free air, and you can feel it's different.
Because of the waves bouncing onto the wall of the throat, they will theoretically slow down, bounce against the opposite wall and form a barrier for the next wave coming up, that possibly not bouncing the walls - so trying to catch up with the previous.

Some would say the principles of this are wrong ...
and again : these are *my* descriptions, and I really don't care what others say or how it supposedly really is; I "work" with my descriptions

But things appear to be different for results ...
Like with the standing wave thing, and the bass traps etc. that should remove them, instead we can also not wakeup the beast.

When listening closely, or with much attention if you like, you can hear when things go wrong with music playback from a horn. And I say it in advance : with any speaker really. It goes wrong when the sounds are squary. Ha ! good one ! not. Because what we perceive is squary just because of the distortion !

So now it is a matter of what comes first : the squary sound creating distortion, or the distortion making the sound squary.
I can tell you, they work together ...

Let's first define the distortion we are talking about.
This is very very difficult to "make up", and let's say it took me 5 minutes to perceive it, but another year to make it clear to myself.

- A trumpet becomes more trumpet like;
- Timbre is more expressed;
- Highs from cymbals become more expressed;
- More detail is perceived;
- More fresh sound in general.

But wait a minute Peter, you didn't mention distortions here ...
Hahaha, yes I did ... Read again :

- A saxophone sounds like a trumpet. I fact all metal blown instruments with sibilance do.
- Male voices have unreal square timbre;
- At higher volumes and specific frequencies cymbals are perceived as pure distortion;
- Unbalanced fresh highs.

So here we are. Distortion all over the place.
But mind you, this "distortion" is very much acceptable in most cases, with which I mean that in most cases you won't know the difference with real life. A male with "unreal square timbre" ... how to tell if you don't know the man ? and worse : how to tell if you heard him singing through speakers only anyway ? Yes, think about that.
There is more in this than you might believe, and only few people will have ever heard the difference by explicit tweaking. I did, at working at the xover of horn speakers (well, out Bert did, but I was there), and just by tweaking the xover for better or for worse, a thing like timbre changes all the way. From completely gone to full square.
So, once you know the differences which can be made, you have learned how all can be attacked.

But also you have learned to hear it. And in the end you hear it everywhere. But also it is still perceived as rather normal, because it can't be different anyway ...

So what happened with 0.9i for results on this ?

The obvious the distortions have gone.
The result is creepy;

First of all, now there is balance balance and more balance. This can be sensed all over.

Then I had to create a new phenomenon for myself on audio playback : fragile. Fragile not in the sense of breakable (or just not broken) but with a finesse where "finesse" is far too rough.
The freshness what I was used to as coming from a horn speaker has changed into sprankling. The harmonics of bells etc. now are not overwhelmed by the over freshness of a cymbal.

Square natural tones (what about negro males) do not destroy the "background tones" like form a cymbal singing at the same time.

There's one phenomenon coming over me each other minute : water like. The horns seem to drool water now. Hard to express better.

There is a stiffness in the silky tones. This one is difficult too. Silky would be true, but silk also is weak. Now the silk is not weak anymore.

There is a so much "founding" bass now. With this I mean that you visualize the bass underneath everything, the further sounds not riding on that bass, but floating above it. The bass now can separate from the rest. It makes it all "deep down earth" bass. Bass can have color.
Individual bass vibes are inifitly better expressed now.

Generally, it now seems that all music creates tears in your eyes. Okay, mine;
Where before you had to have the mood of getting into the music, including the music had to allow for it, now you can't get DEtached anymore. Okay, me.
All is so much homogeneous, that with each and every song you realize that someone has been thinking that song over for many days or more. This is so strong that even right now I have tears in my eyes only because of thinking of last night (no music plays currently). Wow.

Already before (earlier version) I noticed that a distorting guitar would not destroy the complete somg anymore, because it would be in its own corner from then on, not disturning its neighbour players. Now here a major change has happened :
The distorting guitar, producing square waves (!) is now not destroying its own sound anymore. Like the example of the square male voice  who could destroy cymbals.
Rough heavy synthesizers before layd a grey sound over everyting. Now it's just their own square waves you hear (nos-DAC only !!).

In gerenal you can say that not any instrument can smash out the other. I did not know this really, but it now appears this was so. Any ferm drum hit would destroy its surrounding at the moment of the hit. Not so anymore.

The for me best of all is this :
Where I listened to versions of XXHighEnd with so much crazy detail that it became unbearable, now all that detail is back or there is even more.
It is very hard to explain what went wrong with an earlier version, because detail is detail and detail is good. But it was not. It was a version that let fall apart the instrument, as Gerner could describe it so beautifully. We listened to the individual strings of a guitar (interesting) but not to the guitar anymore.
In this version with my setting all the detail iss there, but not over expressed. Again hard to explain how it now is better, but strangely enough there is even more detail than ever before, but now it is good detail. Most probably because the detail now is accompanied with "body". A sort of : the detail has a base tone, and the harmonics of it let you perceive the detail as a whole; in the earlier version the base tone was lacking, and the harmonics were the only tones coming to you. I don't know really.

Horn speakers only ... No !!!

I did not change the properties of the horn speaker. They remain as is, just like the room remained as is, but standing waves disappeared anyway.
Again I say that obviously I knew this in advance, at starting this small but ever so important project. All in other words : this has few to do with horn speakers. But in there is a beast that can be awakened. Don't.

The obvious question is how do we wakeup that beast then.
The answer is more simple than applied :

With digital sourced music we deal with squares. Yeah yeah, you know.
Ha ! no you don't. Not what I mean.

What you probably *do* know, is that the digital data is comprised of volume steps, lateron applied to Volts. It really is volume only, that implying frequency (combined with the time domain -> time proceeds, and volume changes, for CD data 44100 times per second).
What you also (or even better) know, is that the possible volume steps are 65536 for normal CD data. Half of that is for plus Volts, half for minus. When time proceeds, this creates the "wave". This wave really looks the same on the computer screen in digital form, as it would look on a scope in analogue form. If really 1:1 applied, it is.

What you most probably don't know, is how jitter impeeds square waves. Or more square then original. Oh, maybe you knew it, but not me, and that's what came to me in creating 0.9i.
Jitter is the skipping (or repeating) of (in our case) audio samples. When time proceeds, a next volume step is applied. Let's say that for a certain (composed) tone the step rises with 50 (out of the total of 65536). So from one sample to the other, the volume rises with 50. Now, since this coincidentally is a squarish tone like from a trumpet, and where we *know* that squares tend to rise very steep (a real square would try to do it in 0,0000ms which physically nor electornically does not exist) at the next sample another 50 would be added. what is jitter to this respect ?

As said, the skipping and repeating of samples. Hmm ... what happens to our volume steps ?
They want to rise with 50-50-50-50-50, but instead they go like 50-0-100-0-100
So what really happened is that we created a double sized square. Thanks very much.
(note that in practice my sample is a bit overdone, because jitter acts more randomly than I proposed).

Now things become VERY important and quite a few properties add up, most probaly in a more than lineair way :

1. Squares have more energy than a sine (probably the most important aspect).
2. Squares create aliases. This means a square can be found back audibly somewhere else in the spectrum (not good !). Do note that for a trumpet this is just real life. So, the harmonics of it (hence aliases !) will appear in mid air. This is how we appreciate a trumpet !
3. Squares in the analogue domain scratch echother, unlike sines which slide along eachother.
4. In a horn the additional energy can't escape and will bounce to the horn wall.

Mind number 4 : Additional energy ...
But also think of the diaphragms where the additional energy was already put to ... Don't forget, it emerges in the DAC ...
So possibly (or most probably) a horn is more prone to these anomalies than a normal radiating speaker, but a normal speaker would suffer form the exact same in the base ...

By eliminating the additional squares, you loose the anomalies. Simple !!
So, that's it.
:lol:

It is not all *that* easy of course, because the only physical property to "manipulate" would be jitter. Yeah, well, that's exactly what XXHighEnd attacks. Jitter.



Lastly, a bit off topic, but for you good to recognize things. I learned them from perceived better or worse playback throughout times;

When the accuracy is better, one of the first things you'd notice is flanger. The more flanger, the more accurate the playback is.
Flanger is the slowly overall increasing and decreasing of volume. The Leslie (organ) became famous for it.

When accuracy is better, detuned voices become obvious.
This is a very very strange phenomenon, but just true. Like with the flanger, it is just my own notices;
When an instrument is detuned, and then merely *temporarily* detuned (like a singer seeking for the right key), this suddenly becomes VERY obvious.
In the end the principle would be the same as with flanger, because it is about slowly changing ... ALWAYS VOLUME. Haha. Think of what I said before : the only thing really changing is volume. Frequency (tuning) is just a derival of that, implied by proceeding time. So that both flanger and detuning come together seems logic.

Peter



PS: Didn't feel like looking for typos in here. Sorry.

Adrian,

I realize I never took the effort to look into this. But how are you doing ?
You are still using Windows2000 do you ?
And I *still* did not try that ...

Currently I lost track of any possible necessities (for you) to switch to a higher version than the "f" you mention here. Do you have them ?
(obviously I was working on Vista only for the past months)

Peter

Hmm ... WMP doesn't even (officially) want to comply to the Exclusive Mode playback of MS itself. Haha.

EAC : http://www.exactaudiocopy.de/en/index.php/resources/download/older-versions-for-download/

which is for "older versions", but as far as ai can see the latest is "pre-beta".

Download it, don't make any fuzz as of yet, and try whether it then works (but don't use "bulk copy").
If it works, start to worry about the settings before further listening. Otherwise you might be listening to a bad rip anyway.

Peter

As you will have learned by now, the possibility to change the priority of Engine#3 is in there (0.9i).


Not really. Many things depends on the systems.
For #1 and #2 the priority of the Player is more important than for #3. But in either case the priority of the Player could degrade the Audio Engine. So actually the prio of the Player should be as low as possible (like shutting off other services to gain on quality), until it can't do its job in time (which would be preparing tracks, and even timing things).

Anyway, as far as I can tell so far, setting the thread priority to High for #3 really makes a difference.
Also, Vista is more "out of control" at doing things behind your back than I would like.

Lastly, IMHO currently we're at a next level of sound quality, so more "micro" stuff is becoming audible anyway. I'd still won't shut down services, but the Thread Priority is the least we should use now.
Times change ...

Hi Frode 

Engine#3 is kind of "fragile" when it comes to official layouts of WAV files. To give an example : files created with WaveLab won't play, while it will with the other Engines. So the best advise would be to just try another CD of normal origine.

Note that currently only 44K1 16 bits files can be played anyway, with unpredicted results on other formats.

Supposed that all your ripped CDs don't work, how did you rip them ?
With EAC they should just work, unless you're really reading errors continuesly. Then the errors are in the "header data" already, and you're lost. Again, #3 is fragile to that matter.

Peter

This guy will be out for a break.

If you have important questions to answer or things to discuss, tomorrow is the last day for that.

The last weeks I've been trying to get FLAC working, but the documentation on it is too poor (lacking examples) to get it going with the speed I wanted it to go. Maybe tomorrow, but probably not.

Might you have problems with Activation during my absence, don't be afraid; When I'm back I will settle whatever went wrong.
The server for it keeps on running ... until a power outage occurs. Happens once a year maybe, and from a statistics point of view this will happen the next 10 days.

Peter

Hi Frank,

As you might have read in between the lines the reason why it's in there now, it quite different.
It was the perceived phase shift occurring in 0.9h opposed to earlier versions, so when that would be true you'd need the reversing for that as a base (and the 60% recordings with proper polarity would be correct again). Do not forget : this would be so with the Q1 slider at 24.

I had it ready for a month or so, originating from detecting correct absolute phase. Although I've developed an algorithm that might suffice, I still think it can't work, just because of the sole reason of instruments being mixed for polarity, as you told. And, in the end it comes to a preference which varies per track; The drums might be in correct phase after adjusting, but the voice might not, and the latter might be the most disturbing. You said that too. So ...

It is actually of no use to allow the changing of the polarity for two reasons :
1. Currently it can't work real time, which would be VERY convenient in checking was the best setting would be for "you";
2. Whatever trouble you took to find it out, you'd have to apply it yourself the next time because it isn't remembered.

So there you go. You said it all.
But again, the reason why it's in now is very different, and I propose to use it as one definite setting only, depending on what the Q1 slider does to it.
Ultimately it will be remembered, like many more things can be remembered. Take the example of the sax in the latest sample I uploaded ( http://www.phasure.com/index.php?topic=160.0 ) and it really doesn't matter much what the setting to make this bearable destroys on other aspects, if only this sax stops resonating. So that would be the Q1 setting to be remembered ...

Remembering these things for 100% sure will be in there, but there's a few other priorities to deal with first.
And you can bet that Q1 is called so because there will be a Q2 ...

Peter

One more thing : Before well respected horn speaker builders triple over this : This is just one example of which I have one more only. So it is exceptional that this kind of distortion happens anyway. The real message is though : even a distortion like horn resonance is incurred by the player. I said that yesterday when I heard this example for the first time, and I could prove that to be right today. 

Hi Bert,

Couldn't perform two things at the same time. Needed two references (links) in two topics to eachother at the same time.
Thought you all might be sleeping ... 
It's there now.

Peter

Major breakthrough ...

First some credits :
SeVeReD because of his wonderful evaluation of version 0.9d vs. 0.9h and his effortless time he has put into that;
Gerner because of his (mainly offline) explanations and sparring going against that;
soundcheck for his wonderful means of being cosy at a meeting after again, effortless sparring on the best audio playback.

These three people made me see a means of how playback quality could be influenced, which started just some 40 hours ago ( http://www.phasure.com/index.php?topic=159.0 ) 

I want to emphasize once more how important it is for me (and yourselves, hehe) to express yourselves in any detail you can find, like e.g. SeVeReD did in the topic I referenced to. Without that it would be me only, and it would bring a minor part only on improvements. So here it is :

This applies to Engine#3 only.

From the combination of the three people mentioned I could derive that a certain "parameter" influences sound quality majorly, while I myself on my own did not hear the difference so far. It should according to my own theories, but I just couldn't hear it. Obviously it needs the attention to certain music or tracks where things go wrong, or just make *the* difference.

Atually you might need front loaded horn speakers to get the real hunch of what's happening, but since "things" are happening anyway, I think *anyone* can notice, although I would not not how then.

Front loaded horn speakers have the kind of annoyance of being subject to resonances. Not all over the place, but just at certain frequencies and special combinations subject to the properties of the horn concerned. You can hear them as a distortion or you just don't (a bit similar to the color wheel of DLP projectors). And mind you, this just might be the very (a)live like why such a horn brings joy.
But it can go wrong big time also ... See http://www.phasure.com/index.php?topic=160.0

Version 0.9i holds a parameter "Q1" that is able to vary this. As you might expect from the reference above : the lower the better. But beware, there may be drawbacks. All this is jitter related, and there might not be one setting that suits all (types of music).

Version 0.9d vs. 0.9h pointed out that phase is highly related to this all. So much, that I felt it was necessary to shift Absolute Phase 180 degrees as an option.
Twisting Absolute Phase means : blowing becomes sucking. This is related to punch. So if you feel you are in lack of punch, check Invert.
Edit : Careful ! : http://www.phasure.com/index.php?board=2.0

There are so many things related here that I already know I have to add "Edits" to this posts. So watch them.
Anyway, what might happen during the process of lowering or highering the Q1 slider, is that Absolute Phase will change. It will have treshholds, and they depend on your DAC. Passing the threshhold theoretically will imply that at that moment (position of the slider) you will have a bit of nothing. Somewhere further (upper or lower) there will be the most impact. Hard to explain and possibly inaudible for you.

The best thing to do at changing the slider would be to Stop playback (the Stop button) and press Play again. Do note that just changing the slider will have implications on the next track played (just the next in the Playlist sequence), BUT the workout will be different. This is kind of deliberate, but maybe merely because it would need another explicit setting whether you'd want "it" to happen or not. Again hard to explain, but the difference is the initilization of the DAC with consistent platyback throughout the Playlist and choosen settings, versus the initialization (and applied first track) and different appliance for the second and next track. I did not test this, but there is a theoretical difference ...


On the way I allowed the priority for Engine#3 to be changed. Just set the Thread Priority to "High" which can't workout worse IMO. Leave the Player Priority at "Nothing".
Additionally found as an issue, was that the ToolTips accordingly told that the priority was changed immediately after applying it, which appeared just not to be so, and which happened after the start of the next track really. Now it does immediately indeed.

Furthermore, now a warning message indicates that previous settings have been lost, which always happens if an upgrade is installed later than a week or so from the previous "install".
Also, the issue going with this, about the tracknames having a width not respecting the coverart displayed (at first install or the first time startup after an upgrade), is solved.

Edit : The data on the track currently playing (like track length) has been moved to the Info tab. In due time another soluition for showing e.g. the track length will become available.


Both the Q1 slider ad well as Switching Absolute Phase 180 degrees (the Invert checkbox) are not applied instantly. Fot the Q1 slider, see above, and for the Invert checkbox counts that it is applied at the next track, and that it will not influence sound quality by any means because of the processing. So, sure it influences the sound, but not because of any processing needed to twist the Absolute Phase.

Lastly, there is this :
The 0.9d and 0.9h versions were very different for sound quality. Now, if you want to compare those versions i.e. just have them back by means of this 0.9i version, choose these settings :

0.9d :
Q1 slider = 14
This is not 100% similar to 0.9d, but what has changed more in there (opposed to 0.9h) should not matter. If it does anyway, please let know.

0.9h :
Q1 Slider = 24
This would be 100% similar to 0.9h.
When you additionally check the Invert checkbox and therewith invert Absolute phase, you could have best of both worlds 0.9d and 0.9h.

When you pull the Q1 slider downwards opposed to setting 14, leave the Invert checkbox unchecked as a base. When you pull the Q1 slider upwards opposed to setting 24, IMO you should check the Invert checkbox.
The stuff in the middle is up to you (and somewhere there will be a treshhold).

Happy tweaking !
 

Here is an example of frontloaded horn resonance, most probably not in all horns expressing the same, hence at the same frequencies, but it should count for the Orphean horns, still somewhat depending on the xover filter used.

Important : This is not about expressing anomalies in front loaded horns (if at all) but merely to show that resonance-like anomalies can be attacked just by software and that they are incurred by jitter and thereof derived harmonic distortions.

Do not hesitate to run this sample at higher levels, but do note that it will bring some temporary brain damage at least, then. Also, try to recognize the (harsh) instrument in the first place ...

You will notice that the instrument comes from the right speaker, right from the throath (with horns).

As of this writing, it is implied that you are running version 0.9h or earlier, so now you should try version 0.9i (or higher).
Run Engine#3, leave Invert Off, and set Q1 (Quality Level 1) to 4 (when something goes really wrong set it higher as close as possible).
Re-listen.



(see http://www.phasure.com/index.php?topic=161.0 )

Well, the last word about this hasn't been spoken.

I have just met 5 people in one room who were judging XXhighEnd (0.6h) as lacking of punch against a well tuned Linux player. Including me that was 6.
So I suggested to try it with 0.6d, and everybody agreed again; punch was back. And some more things.
The gear we listened to is near equal to my own; the room very different.

I'll build in the "tweaks" so everybody can be satisfied.
Or judge better.

Even more important might be the understanding of what's actually happening here ...