Read the literature on cutoff > the term is used in all the relevant papers (even those of Markaski, Tengelsen etc.).
The SEOS-style horn is based on spherical horn math which includes cutoff, the profiles are modified which improves DI at the price of some horn loading.
Still these horns load the driver much better compared to an OS waveguide.
Docali has shown many examples including sims, so there's nothing to proof. The proof is in the physics.
Such horns are used by many, that don't care about DI, so what's the basis of your authority to judge those people?
Just to put in my 2 cents, I don't want to get into the middle of this, and I thank Marcel for taking up the challenge. I gave up arguing with you a long time ago.
I agree with Marcel the majority of the time and I disagree with you the majority of the time. So for others looking on I encourage you to listen to Marcel.
Also, within REW there is a tab called "Clarity"....when viewing my horns measurements, there is a knee, where the Clarity line drops off....the knee is pretty much....1.5....to 2 times...the fb frequency.
Clarity is based off of early to late energy ratio...
There seems to be plenty of data supporting crossing a horn over 1.5-2x fb...
Clarity is based off of early to late energy ratio...
There seems to be plenty of data supporting crossing a horn over 1.5-2x fb...
Who's the Pan!
Hook (1991) - You Are the Pan - YouTube
1. ____________
2. ____________
3. ____________
4. .......
Or,
Are we looking for another Greek god of majik?
Hook (1991) - You Are the Pan - YouTube
1. ____________
2. ____________
3. ____________
4. .......
Or,
Are we looking for another Greek god of majik?
Last edited:
If you applied this to the response behind the crossover (i.e. to any high-pass response), you would get the same result, it would show a GD rise at the knee. So what would you do? Set the crossover 1.5 - 2x higher? Then it would still show a GD peak at the knee...
Well generally the xo is going to up GD, so common sense says lets not do this in an area where theres already a problem with GD...that would be making a bad thing worse. We are much more sensitive to GD in the upper spectrums above low bass.
And no, you wouldnt get a GD with any highpass...one of the benefits of Linear phase XO's actually. Also much less insulting are the 1st and 2nd order HP filters...and to keep the response correct through XO....you'll need to be about 1.5-2x above fb lol
Camplo,
If you were to measure your horn/driver while you EQ the response flat one octave below the current knee, you would see the knee of the GD plot move along with it.
I don't recommend doing so, but that's how the relationship works.
If you apply a crossover 1.5 to 2x above fb, the knee will move a long with it too, both of them. In frequency response as well as group delay. (as long as the crossover is minimum phase)
So You didn't find any magic here, just the roll of following minimum phase of the driver + horn combo. Which after applying EQ would or could move either way. Depending on the choices made. At least if you don't apply linear phase EQ.
But one could easily EQ the phase flat of the driver + horn in question and have no group delay knee either. Since you brought up linear phase modifications...
Euhh... the group delay knee is there because the frequency is already declining. Applying a crossover right there would make it into a higher order acoustical crossover, with the accompanying group delay. I wouldn't call it a problem though. Just physics.
If you were to measure your horn/driver while you EQ the response flat one octave below the current knee, you would see the knee of the GD plot move along with it.
I don't recommend doing so, but that's how the relationship works.
If you apply a crossover 1.5 to 2x above fb, the knee will move a long with it too, both of them. In frequency response as well as group delay. (as long as the crossover is minimum phase)
So You didn't find any magic here, just the roll of following minimum phase of the driver + horn combo. Which after applying EQ would or could move either way. Depending on the choices made. At least if you don't apply linear phase EQ.
But one could easily EQ the phase flat of the driver + horn in question and have no group delay knee either. Since you brought up linear phase modifications...
Euhh... the group delay knee is there because the frequency is already declining. Applying a crossover right there would make it into a higher order acoustical crossover, with the accompanying group delay. I wouldn't call it a problem though. Just physics.
Last edited:
Its not a competition, we are just trying to get to the bottom of things. Martin king tried to sell me on group delay being nothing but the derivative of the phase curve also....and when I displayed the connection, to the time to peak energy curve in the spectrogram, he also had no come back. Mabat you are a good one, and youve corrected me several times over but when it comes to group delay....lol
Wesayso
Ive done that plenty and where the group delay was high before filtering, that area was even higher if the filter is placed there...it had an additive affect. So if group delay is lower in said area, before filter, same frequency, the same filter will result in a lower final group delay
Wesayso
Ive done that plenty and where the group delay was high before filtering, that area was even higher if the filter is placed there...it had an additive affect. So if group delay is lower in said area, before filter, same frequency, the same filter will result in a lower final group delay
Last edited:
True, because together they form a higher order acoustical crossover. The group delay you see before filtering is due to the frequency dropping off there. If it (the frequency response) were flat, the group delay wouldn't be there either.
I've spend many posts replying to people to explain the difference between acoustical crossovers and electrical crossovers. Most people believe when they are applying a second order crossover, the driver in question automatically will follow that curve as well. But the reality is that the roll off of the driver or horn + driver is added to the electrical crossover.
The resulting curve will be the acoustical crossover. It doesn't matter if one picks a certain named slope from a program in a digital solution or if one does it with coils and caps.
As long as it (the digital counterpart) doesn't apply phase compensation as in FIR filters the minimum phase behavior rules.
If you look for posts from Mark100, where he talks about applying crossovers you'll see that he EQ's the drivers flat a specific amount beyond where he wants to cross before applying the phase linear crossover. That EQ work to flatten the frequency response beyond where it falls off by itself also flattens the group delay in that area. The EQ needed to do that is minimum phase or IIR EQ. It makes it easier to get a phase linear result.
Last edited:
Camplo - you have it wrong and you should listen to what people are telling you. GD, Phase slope, "peak energy" are all the same thing. Just different ways of looking at the data. The audibility of excess group delay is not a resolved issue. With any resonance there is GD. Do we hear the GD or the SPL peak, if they arise from the same issue?
I think he just gave up explaining. As I'm sure he did realize that all those graphs are a derivative of the same impulse.
Or as Dr. Geddes called it just minutes ago: another way of looking at the same data. How do you think a program like REW is able to show you that data... It can make all those graphs from one impulse measurement.
Because it really is just another way of showing you the same data. Some are more clear than others, but they all come from that same pulse. All those ways to show the same data are to help you find your way. As soon as it starts making sense for you
.
You contradict yourself here.
But it doesn't matter, even if you add $1000.000 instead of 0.02c. You only amplify the power of my comment.
I know where I come from and foresaw such a comment not recently but years ago, long before both of you even knew my name.
A protegé won't let down his virtual proponent, now would he?
If the content of your argument was only remotely obvious, you wouldn't have to make it in the first place.
But your biggest blind spot becomes visible by focusing on me.
It's not about me, it never was.
I dont understand why I should just ignore time to peak energy....I cant see it not being an audible issue... what it sounds like is being said is that there is no difference in sonic quality of a 1st order vs a 48db xo.
Having group delay/phase/time to peak energy be essentially the same thing (which I understand) does not = ignore it....there is correct...and incorrect time to peak energy....as Dr Geddes just said, the audibility of excess group delay is unresolved...
But its being argued by myself and those who disagree with me as being resolved lol
Time shifting harmonic information is not good...because of the phase relationship, you end up with a new harmonic character vs the original
Having group delay/phase/time to peak energy be essentially the same thing (which I understand) does not = ignore it....there is correct...and incorrect time to peak energy....as Dr Geddes just said, the audibility of excess group delay is unresolved...
But its being argued by myself and those who disagree with me as being resolved lol
Time shifting harmonic information is not good...because of the phase relationship, you end up with a new harmonic character vs the original
Last edited:
Camplo - My advise would be, and I really mean this on a constructive way, to find a good read/book on control theory and control systems.
Or an online course or something these days.
That way you will get a much better idea what is at play.
Skip all the extremely boring math, unless necessary (obviously), but try to understand the bigger picture and most importantly the relationships.
These things are just really hard to explain to someone in a couple of forum posts but are a fundamental essential to understand what problems are involved.
Another good read is Loudspeaker Handbook by John Eargle.
I think there is a better updated version these days by Beranek and Mellow
Or an online course or something these days.
That way you will get a much better idea what is at play.
Skip all the extremely boring math, unless necessary (obviously), but try to understand the bigger picture and most importantly the relationships.
These things are just really hard to explain to someone in a couple of forum posts but are a fundamental essential to understand what problems are involved.
Another good read is Loudspeaker Handbook by John Eargle.
I think there is a better updated version these days by Beranek and Mellow