Comb Filtering - Quick way to Tell


Hi Everyone,
Lots going on here, with this odd Golden Ear vs. Tekton debates. I personally don’t care, except there’s a lot of theory being bandied about as proof of the superiority or inferiority of a design. The arm-chair experts are really giving me a headache because my eyes can only roll back so much before my optic nerve gets kinked.

I have no interest in either brand, really, but let’s talk about it in practical terms.

Comb filtering happens when you have simultaneous arrivals of the same signal which are slightly out of time alignment. Really, any multi-way speaker has to deal with this, but arrays in particular. Much has been written and debated, but let’s boil it down to this:
Comb filtering affects what your ears hear in a particular location in space. The frequency response at location A can change rapidly a few inches away at B. In fact, we can use this in a process called interferometry to get very precise measurements of acoustic offsets.
The point?

If you are concerned about an array having comb filtering, move your head. That’s it. If it is a problem, small changes in your listening location will cause significant changes in the frequency response. Stand up, sit down. Is the frequency response changing much? is it pleasant or uncomfortable/weird sounding? Do this for any speaker.  Listen at the center and both ends of the couch.

But please, listen realistically! Listening 1’ away to an array may yield vastly different results than at 9’ so it is important to understand the manufacturer’s design and intention, which is really true for all speakers.
Which speaker should you buy? The one you like the most. For heavens sake, listen to them and look at them.

We should also point out, as I alluded to elsewhere, there are lots of interesting ways to handle potential issues with comb filtering, so the idea that you can measure center to center differences, and diaphragm diameter alone will tell you what the speaker will sound like at the listening location is simply not enough information. Your head and ears however can do an excellent job of discerning these issues.
erik_squires
Post removed 
. If you look at the horizontal response plot in the link, that is exactly where they begin to decouple and soon after their off axis response plummets

Except with a vertical array, horizontal "decopuling" as you call it (not a term I’ve ever heard in speaker design) is exactly what would never happen even in the most poorly designed arrays. What you are seeing is the normal off-axis drop off of the tweeter, combined with the unusually large mid-woofers (see JA’s comments in the measurement texts). We usually refer to this as lobing, and is a type of comb filtering inherent in almost all multi-way speakers. A combination of the woofer’s output, crossover and rapidly changing acoustic offsets with angle. We can argue about who does it better, and whether speaker A vs. B’s coverage sounds better, but to call this a fault of the line array itself is simply not accurate.

A better place to see this line-array comb filtering is the vertical plots, and comparing it to how a similar 2 way would function. In this sense the Enzo’s array offers superior coverage below axis, and, like many other speakers, above gets wonky.

As I try to mention, comb filtering of the type you are complaining about happens very very quickly, in a matter of a few inches, and goes back and forth. That is, as you move the microphone the nulls and valleys shift rapidly. This smooth drop off across the horizon is anything except that. It does, overall, suggest a narrow dispersion which in many rooms will be a good thing.

But based on the charts, I don’t see any of the types of array / comb filtering mistakes that have been discussed for it.


Best,

E
Nice thread Erik and good, simple explanation for Comb Filtering.  
Way back in my Marcof/SpeakCraft days, we built a tower with a 15" in a 2way and 3way version and 4 Audax HD's on top.  Comb filtering was minimal and this was a fairly smooth and refined speaker. 
On the tweeters, we tried different variations of series parallel wiring as well as taking specific drivers out of phase and changing the slopes on these tweeters in sets. The end affect was that Comb filtering/lobing effects were minimal. So my experience is closer to Eriks.  
Tim 
Post removed 
Kosst,

Comb filtering due to array specific issues will be shown in the vertical, not the horizontal plots. The reason is simple: The acoustics offsets to all tweeters is identical across the horizontal plane.

Across the vertical plane, the acoustic distances to each tweeter is changing. That is where comb filtering would appear.

Everything else I see is due to "normal" 2-way design choices.

However, I go back to listening. Listen to see if the off-axis performance is what you like or not.

@timlub

Thanks for the vote of confidence! 

Tim, sounds like you are discussing something similar to a Bessel Array. One of several ways to deal with real comb issues. Another simple way of course is to use smaller number of tweeters as the frequency goes up, which is what Tekton seems to do.


Best,

Erik