XXHighEnd

All,

I think it is confirmed that when Inverting the Absolute Phase things can go wrong badly.
I had this myself, and cracks can be very loud and can be persistent. It occurs at the peaks of sound (full digital volume level).

I thought I solved it, but appatantly it can behave differently per track.
Save your speakers and do not use it !

:sorry:

Also ... more than 16 bits currently is not supported generally, but I did not even test it;
Might you feed XX with more and it seems to work, then with Inverted Absolute Phase it for sure will go wrong (at the peak levels).

Peter are you "off line" and on Holiday now?

However it is I have no clue of what you are talking about, but when you talk there is a reason to it.

I did not experience any of your findings. The reverser works fine here. So?


Gerner

So, for anyone who's had things "go wrong" with the phase inverting option; how bad is it?  Like possible speaker destroying bad? hehe... maybe I'll wait. hmmm

No, it won't be bad for your speakers. Also (if at all) it will happen only with tracks that contain "full volume". If the recording is right, this won't be in there are all, and if all is "maximized" without clipping the data (!!) there would be one small "crackle" (IMO hardly audible) throughout the track. However ... with bad "recordings" (digital masters), the sound can be clippen for even a second long, and that several times throughout the track. It still won't damage the speakers, but it is just not good for listening obviously.

I will do some better testing myself now, and let know whether it's really wrong or not.
Peter

Ok, I'm working on that now. But Johan,

At last I tried your PMd Truffazz track ... But there's a warning here;
With or without Invert it cracks. BUT, this track is clipped at many places (several per second). Maybe the cracks are a tad louder with Invert, but that might be because of the inverted phase itself (it might express better the "dynamics" of a crack, hehe).

I don't say the problem is not there, but be careful with what you actually listen to.
It can well be that clipped data (which is just about the volume reaching the digital max) is audible with today's "quality", whereas before this just was merged in the fuzzyness (?).

Peter

I think that's precisely what the Invert box means to me. I never thought it had anything to do with phase (per se), but what I hear is tighter focus and crazy wild dynamics.

Indeed I looked at a few of the songs (with a WAV editor), that "crackle" for me with Invert checked, and I can definitely see that there is a lot of clipping. I started going through a whole bunch of songs from my library, randomly, and so many songs contain clipping. What am I supposed to do about that? I don't hear the crackling with the Invert unchecked.

What are you going to do about that ?

Nothing much. But the truth is that with Inverted I already lowered the output in that area (which would be DSPing  :() and it doesn't help so far.

But by now I collected a few CRACKling tracks myself. Now how to deal with them ...

Solved ...

http://www.phasure.com/index.php?topic=204.0

Trufazz sounds ok (only recorded cracks)! Great job!

Johan

Well, the crackling (with the invert) is now gone, but sadly the SQ is not the same. It has gotten worse (more "cloudy"). I prefer 0.9k (with the Invert checked).

Edward,

Do you mean with Invert the SQ has changed ? that's possible.
Without Invert ... impossible.

Or ?

Hello Peter,  :prankster:

I don't hear the difference between having Invert fase checked and having invert fase unchecked. What is the best thing to do in most of the case's?

Grtzzzzzzz  :grin:

Hi Gerard,

It depends ...
Many recordings have been recorded with the wrong absolut phase. But also, many individual microphones etc. are connected with the wrong polarity. This can imply that the whole track is okay, and only the singer is wrong. etc. etc. etc.

If you don't hear the difference, this is because you did not learn how to hear it. Generally things which blow start to suck or things which need to suck now blow when wrongly recorded. A voice does both, but a drum just as well. It will be in the attack and decay where things are audible (but upside down windblowing also is strange). Note that IMO it is not so that physically the suck/blow thing happens (it would require the track to play in reverse order), but what happens is similar; where you speaker oughts to push, now it draws. So it happens at the micro level.

I, so far, never checked the Invert, just because I don't like to be bothered by switching the thing each 5 minutes (and back because things got worse). However ...

As it occurs to me acoustical phase can change per jitter influence. Thus, the Q1 at -4 might require Inverted Phase in order to be normal again.
Note this is not proven (let alone science), but it just seems to be so that in certain areas of the Q1 value, Inverted Phase is better.

That we can be fooled by "different" dynamics is another thing. A blow which sounds like a suck will be less dynamic. So, things that originally suck are going to blow now, would give more dynamics. Generally you could perceive it better, but it would be fooling.

Ok. Sucked enough for now. :)

Yes. I only compared them with Invert checked.

However, I did (very briefly) try 0.9lb with Invert un-checked (in comparison with it checked) and I almost could not tell a difference. The impact of checking Invert was less obvious than it is with 0.9k.

Hey Peter - you haven't replied to my last response, so I was just wondering where you stand on this. I just tried version 0.9m and you have done good things here. However, the SQ with Invert checked is still different than 0.9k. In some ways I like it (0.9m) better, and in other ways it lacks the detail of 0.9k. (Of course, when playing 0.9k, I have manually set the priorities and affinites the same as 0.9m achieves automatically.)

So I was wondering, is the Invert function in 0.9m the same as it was with 0.9lb, or did you make further modifications?

BTW, with 0.9m and Invert checked, there is no crackling and gapless works. Success.

True Edward ... had to find an appropriate answer ... or in fact, I have/had one, but you may not like it that much. ;)

In the later versions -which exactly I can't tell- had slipped in some debug code for this matter. So, this was active in production unintendedly. :blush1: This debug code did just *not* activate Invert, but the math to correct a 2 byte positive integer (as what is used in Redbook) while half of it represents negative (volts), was still active but now wrongly applied. In fact, with Invert the lot was DSPd a bit ... and certainly did not invert the phase.  :blush2:

Now it is working properly, and what you might perceive from it is untouched data, whereas before it was mangled with (though unintended). So now the phase is just inverted, nothing else. So what you are dealing with now is that maybe you have to learn how to hear the difference.

But careful now ...
The Phase Inversion, though originally setup to satisfy those who are disturbed by wrong absolute phase, when finished was not officially implemented at first because of the perceived idea of it not being useful because it hardly occurs that a whole song is recorded with wrong absolute phase, and merely individual elements (singer, guitar, the snare drum) are. However :
When the Q1 was introduced, it occurred that at certain "levels" of Q1, now Invert suddenly could make the difference of the world. Btw, this was when the bug as described above did not get into the production version yet.

Moral : When Q1 starts to be, say, nasty, it is exactly the point where Phase Invert will drop your jaw. That is, it looks like the general experience from people; not only my own. This phenomenon by itself implies that there are also Q1 settings where Invert hardly makes a difference (think of +90 vs. -90 degrees).

Phase is the most important phenomenon within audio IMO. Not so much to the sense of "proper phase" but far more in the area of properly aligned.

I doubt that ... although you could have done it technically. :)

I think I mentioned this before. I also felt that the detail I was experiencing (with 0.9k) might have been inaccurate (or crazy exaggerated in certain frequencies) and it did sound (dare I say) a bit too digital (or maybe what you refer to as squary).


That's exactly the impression I was getting while playing around with 0.9m. Phew! I thought maybe you forgot to activate it or something.

OK, well, thanks for the explanation. All makes sense now.