Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Douglas R. Allen]

  I ran across a post on another site that claimed to "smooth out" the phase cancellation of putting 2 - 1-15 and horn 90X45  boxes that even though they were properly splayed out that if you put 3 ms of delay on the outside speaker you will smooth the comb filtering out and improve the overall response.    Did I miss something in the past 30 years or so?

  Is there a basic program I could download to model 2 speakers side by side with a 100 , 200 , 400 hz tones to see how it effects the low frequency coupling?  First with both with no delay then again with 3 ms added to the outside box?  I always thought it would be best when you have 2 speakers close together you'd want to have no delay on them to keep the frequencies coupled for as long as possible. Certainly the low frequency range where they are closer than 1/4 wavelength.  I would think 3 ms or 3.5 feet of delay would keep that from happening. Down to around 90hz or so anyway.  Is this sometimes done with good results?  I would think it would lobe or steer the bass? I also would think it would make a Frequency related to delay delay comb filtering sound.  I realize that 2 - 90 degree boxes combined side by side is a bad idea but any thoughts on this? 
 
  Here to learn  as I may have simply missed this though the years even if it sounds odd.

Douglas R. Allen

[Dave Garoutte]

  I ran across a post on another site that claimed to "smooth out" the phase cancellation of putting 2 - 1-15 and horn 90X45  boxes that even though they were properly splayed out that if you put 3 ms of delay on the outside speaker you will smooth the comb filtering out and improve the overall response.    Did I miss something in the past 30 years or so?

  Is there a basic program I could download to model 2 speakers side by side with a 100 , 200 , 400 hz tones to see how it effects the low frequency coupling?  First with both with no delay then again with 3 ms added to the outside box?  I always thought it would be best when you have 2 speakers close together you'd want to have no delay on them to keep the frequencies coupled for as long as possible. Certainly the low frequency range where they are closer than 1/4 wavelength.  I would think 3 ms or 3.5 feet of delay would keep that from happening. Down to around 90hz or so anyway.  Is this sometimes done with good results?  I would think it would lobe or steer the bass? I also would think it would make a Frequency related to delay delay comb filtering sound.  I realize that 2 - 90 degree boxes combined side by side is a bad idea but any thoughts on this? 
 
  Here to learn  as I may have simply missed this though the years even if it sounds odd.

Douglas R. Allen

It seems like a bit of delay (based on the center to center distance) would help the outside speaker blend with the inside one if you are in the CENTER of the audience.
But the point of the splay is to put the outside speaker energy to the OUTSIDE. 
The primary concern has to be at the overlap of the two patterns, so no delay would make more sense.

[Robert Healey]

  I ran across a post on another site that claimed to "smooth out" the phase cancellation of putting 2 - 1-15 and horn 90X45  boxes that even though they were properly splayed out that if you put 3 ms of delay on the outside speaker you will smooth the comb filtering out and improve the overall response.    Did I miss something in the past 30 years or so?

  Is there a basic program I could download to model 2 speakers side by side with a 100 , 200 , 400 hz tones to see how it effects the low frequency coupling?  First with both with no delay then again with 3 ms added to the outside box?  I always thought it would be best when you have 2 speakers close together you'd want to have no delay on them to keep the frequencies coupled for as long as possible. Certainly the low frequency range where they are closer than 1/4 wavelength.  I would think 3 ms or 3.5 feet of delay would keep that from happening. Down to around 90hz or so anyway.  Is this sometimes done with good results?  I would think it would lobe or steer the bass? I also would think it would make a Frequency related to delay delay comb filtering sound.  I realize that 2 - 90 degree boxes combined side by side is a bad idea but any thoughts on this? 
 
  Here to learn  as I may have simply missed this though the years even if it sounds odd.

Douglas R. Allen

This is a technique Bob Coffeen used (calling it "optimal misalignment") and taught. He had a demo where he would play pink noise for an audience on a pair speakers that rotated so you could "sweep" the audience with the comb filtering, and it did sound smoother with the delay. Bob found it increased intelligibility in systems that required multiple loudspeakers. With EASE, you can see that it was steering the worst comb filtering artifacts to less critical areas of the audience.

I am not so sure it is valid with modern systems. When this technique was developed, the time domain was not being measured and most of the systems that were being designed then were old horn-type clusters that were completely time incoherent to begin with. Achieving intelligibility in installed systems was seen as more important than fidelity. Modern installed systems are much more music-focused, and we know a lot more about how the time domain affects fidelity and have an active goal of preserving it - especially with the use of coaxial designs and FIR filters.

Either way - it is worth a try, especially if you have a cluster of traditional 15" + horn boxes.

EV used to have an Excel based cluster prediction program - not sure if it is still available.

[Helge A Bentsen]

I've done that several times, but I used delay AND level to match boxes with reduced splay angle.
There is a formula somewhere in Bobs book to calculate splay angles when you reduce level between two identical boxes, I just use Smaart and roll with whatever measures "right" at the time.

[Caleb Dueck]

I've done that several times, but I used delay AND level to match boxes with reduced splay angle.
There is a formula somewhere in Bobs book to calculate splay angles when you reduce level between two identical boxes, I just use Smaart and roll with whatever measures "right" at the time.

I've played with this a few times.  It merely steers the nasties a bit outside the pattern of one box, but screws with impulse response and comb filtering.  Whether or not it's better - it depends.

[Ivan Beaver]

When you have 2 sources arriving at different times at the same time you WILL have combfiltering.

YES, you can "fix it" for 1 location with delay, but you will make other locations worse.

You have to choose what is most important.

[Bradford "BJ" James]

When you have 2 sources arriving at different times at the same time you WILL have combfiltering.

YES, you can "fix it" for 1 location with delay, but you will make other locations worse.

You have to choose what is most important.

Ivan,
Back when I had a Versarray and hung out on the PV forum, there was a hot headed PV designer who always advised to run either the outside or inside pair a few dB lower to kill (or maybe just slightly maim) any comb filtering when using 1 pair of wide dispersion trap boxes/side. Doesn't that simply move the area of combing more towards the listening area of the attenuated boxes?

[Caleb Dueck]

Ivan,
Back when I had a Versarray and hung out on the PV forum, there was a hot headed PV designer who always advised to run either the outside or inside pair a few dB lower to kill (or maybe just slightly maim) any comb filtering when using 1 pair of wide dispersion trap boxes/side. Doesn't that simply move the area of combing more towards the listening area of the attenuated boxes?

Changing the level doesn't shift the nasties, that's what delay does.  What lowering the level does is makes the nulls slightly less deep.  Rather than have +10 and -10 signals combine for 0 (0dB SPL, IE pure comb filtering) - you have +10 and -7 signals combine for -3 - it's still bad, at the same locations, but the nulls aren't quite as deep.  The peaks are slightly less high, but that isn't as noticeable. 

[Ivan Beaver]

Ivan,
Back when I had a Versarray and hung out on the PV forum, there was a hot headed PV designer who always advised to run either the outside or inside pair a few dB lower to kill (or maybe just slightly maim) any comb filtering when using 1 pair of wide dispersion trap boxes/side. Doesn't that simply move the area of combing more towards the listening area of the attenuated boxes?

As Caleb says, the freq of the combfilters will not change, just how deep and high they are.

You get maximum combfiltering  when both arrivals are equal in level.  As soon as either one goes down in level, the depth of the notch (and the peak of the sum) is reduced, but the freq stays the same.

The basic math is this.  The spacing of the notches (the peaks are between the notches) = the difference in time of arrivals.  So if the arrivals are 1ms apart (about 1'), they will be spaced at 500Hz intervals.  The first notch will be at 1/2 of the time difference, or in this case 500Hz.  1ms=1,000Hz. 

So you will have notches at 500hz, 1KHz, 1.5KHz, 2KHz, 2.5KHz, 3KHz, 3.5KHz etc

[Bradford "BJ" James]

Great explanation. Thanks.

[ProSoundWeb Community]