Of cause the Comb filtering is there like in any reputed itself louspeker but it s not the problem that would manifest itself in your case. Remain you that “your case” was the following:

“Moving your ears 1/2" closer to all drivers in the system is not the same as moving a single driver 1/2" closer to your ears.”

The tip would be that in the acoustic systems that have the configuration like yours and mine the exposure to the bug will be particularly vulnerable. Reminding the homework:

1)      What is the bad

2)      Why Romy's system is more susceptible to the bug then Jessie's

3)      How to deal with the bug.

The presence of this "bug" is clear. Apparently your Raal tweeter is somehow making your system more vulnerable to the issue than mine. I'll refrain from guessing exactly why, as I'm not familiar enough with the dispersion pattern of this tweeter or ribbon drivers in general.

I have to admit that I have a kind of superstitious aversion to mixing driver types (that rely on different ways to generate SPL) especially across the HF/MF transition; this same aversion extends to filter slopes.

jd*

noviygera wrote:

I think clearing this up this may be of benefit to other users who are building multi-way horns. Let's say I've got two horns mounted on a frame. There's a high horn above and a mid horn below and they are physically time aligned, meaning the magnets of the drivers are about at the same vertical axis. I want to cross them over (either passive or active or line level passive) at 2kHz. I want to try 6db/octave slopes and 12 db/octave slopes for SQ comparisons sake. The sound wavelength @ 2kHz is about 7 inches. 6db crossover have a 90 deg. phase shift; 12db crossover have a 180 deg. phase shift. And the QUESTION IS: Switching between 6db and 12db crossovers slopes (and other slopes like 18 and 24db), do I also need to physically move the mid and high horns in relation to each other to have proper time alignment? Or should I leave the horns the way they are?

Romy the Cat wrote:

Herman, I completely do not understand what you ask. The physical aligning of magnets is absolutely irrelevant. The crossover slopes are irrelevant.

Romy the Cat wrote:

Why would you need to do switching between crossovers slopes?

Romy the Cat wrote:

All that you need to alight is the arrival time of the summits of two sinusoids in the same period. I think the best would be not just tell you how to do it (it was told many times before) but to encourage you to listed your multi-way with and with alignment and to learn what the auditable difference might be.

noviygera wrote:

I want to listen to both 12db/octave and 6db/octave passive versions and see which one sounds better. And keep the one that sounds better.

It depends what you would like to listen. If you would like to assess how a given driver/channel sound with first and second order then do not need to worry about time alignment for most of the cases. If you concern what happened with imaging and other fine element of channels performance then an alignment need to be done.

noviygera wrote:

I am trying to move away from my DSP crossover to a passive crossover. Currently I have a DSP crossover @ 2Khz @ 12db/octave. I will make two versions of the passive crossover: a 12db/octave and 6db/octave version. I have the capacitor and inductor values calculated. So I need to know when I switch to the passive crossovers, do I also need to re-align the horns?  If so, would you recommend listening to the system and moving one of the channels (lets say the highs because it's easier to move) until the sound becomes coherent? Then let say I switch the crossover slope, do I have to re-align them again? The summits must be aligned at the listening position, right?

OK, I see now. a few things you need to consider. 2Khz @ 12db/octave is not the same as 1Khz @ 6db/octave filter of the 1.5Khz filter. Then might be mathematically would give you the same cut off but my experience indicate that each driver/enclosure have own reactance to those changes. There are many variables in place here, and it the cheval is horn-loaded then there are even more variables.   Then there is more complexity. When you get to second order filters then the question is what kind filter it would be. Bessel filter have constant phase and very much resembles the first order filter but with faster roil off. However you will be able to write Bessel profile only if you have fixed load impedance or if you use passive line-level filtration. If you filter at speaker level then the impedance will be running in there wild and only God knows what kind filter Q will be effective at any given moment. You might find it noticeable or not, however.

When you switch from one filter to another then you misalign your channels, even if you move your filter a few cycles up or down you misalign your channels and you need to do it again. You do not need to ask me. Whatever method of time aligning you use – pulses or RTA it will clearly indicate to you that any change of crossover set channels in misalignment. Yes, you need to align at the listening position but do not go overly crazy over it as at the most parts you shall be able to make judgment about first vs. second order just by listening approximately-aligned delivers. After you make the decision, then you fine-align them.

BTW, you shall be able to make a decision about the first vs. second order crossover base upon use of your DSP crossover. That is what the DSP crossovers are very good and comfortable for.