Cancellation between main speakers

[travman83]

Hello everyone,

I've been tinkering with my HT audio for several months, but this is my first time posting. I have taken some REW measurements for my FR, FL, and FR+FL speakers from the MLP that are stumping me. These measurements are post-Audyssey and without subs. The speakers when measured individually (green and orange) don't show any noticeable boundary cancellations, but when measured together (blue) I'm getting a significant cancellation at ~100hz. Any thoughts on what this could be from or how I can address it? The speakers are equidistant from the listening position making a rough equilateral triangle and about 2.5 ft from the back wall. The left side of the room opens into a larger area and the right side is walled off with the speaker ~3.5 ft away. Any thoughts are appreciated!

 

[Sonnie Parker]

Hmmm... don't know that I've ever measured both together... I've always only measured left and right separately. I don't think you would want to try to equalize them together. The only speaker I've ever combined, measured and corrected together is the subs, which are non-directional and about the only way to filter them correctly.

 

[travman83]

Hmmm... don't know that I've ever measured both together... I've always only measured left and right separately. I don't think you would want to try to equalize them together. The only speaker I've ever combined, measured and corrected together is the subs, which are non-directional and about the only way to filter them correctly.

Thanks, Sonnie. Good point! What drove me down this path was I noticed an audible drop in the bass region when listing to music, which didn’t show up in any of the individual speaker measurements. So I then decided I would measure both simultaneously to see if something funny showed up and got that result. My only thought is maybe I have direct sound from one speaker interfering with reflected from the other. Room treatment might be in my future.

 

[FargateOne]

...of course you have checked polarity of the 2 speakers...just in case...

 

[travman83]

...of course you have checked polarity of the 2 speakers...just in case...

They appear to be correct. Although maybe it’s possible one of the drivers is wired with reverse polarity in the speaker?

 

[Sonnie Parker]

You can take a 9 volt battery with a short speaker wire and test the drivers.

 

[travman83]

You can take a 9 volt battery with a short speaker wire and test the drivers.

Good tip. Just did that test and everything checks out phase wise. I’m not sure what’s left to do other than tame the wall reflections.

 

[DanDan]

It looks like a massive cancellation. I wonder how widespread it is? Our ears are nearly a foot apart, depending on how much praise and compliments........ Plus a big lump of meat and bone blocking. One wonders if this cancellation occurs at either or both in that context.

 

[JLM1948]

The left side of the room opens into a larger area

I believe it's the crux of the problem. This communicating room either results in antiresonance (Helmoltz) or provides a massive delay (100Hz is about 100ft travel).

 

[BenToronto]

The cancellation seems a bit large. One of those dirty little secrets (at least a secret from the hyper-engineer types who miss the forest while fixing the trees) that - of course - there will be ranges where the two speakers will be in antagonistic phases and double-trouble in the lower frequency range esp around XO... and then DSP differences and then distance differences....

May seem odd to say, but the joint response may not mean much, at least for sine waves, as per your post. But stereo white noise is the indicator you really should be concerned with. My guess is that even with slightly different FRs on the two sides, instruments are still well attached to their locations because an instrument is a whole structure of harmonics, not deviated by a freq on the L or R.

 

[bamtron]

Easiest way to prove if room problem is to move the listening position microphone 2 feet toward the speakers. This will change the path length, which would change the frequency of the null. If the frequency doesn't move, then the problem is not the room.

 

[JLM1948]

and then distance differences....

The OP clearly stated that the mic in positioned in the median plane, so path differences are minimal, and don't justify cancellation at 100Hz; half-wavelength ca.5ft.

 

[JLM1948]

Easiest way to prove if room problem is to move the listening position microphone 2 feet toward the speakers. This will change the path length, which would change the frequency of the null.

Considering the distances, there is no way it can result in dip at 100Hz.

If the frequency doesn't move, then the problem is not the room.

Room modes are good examples of response problems that don't move when the mic is moved.

 

[bamtron]

I am, of course, talking about path differences of room reflections. If, say, rear wall reflections are causing the null, then moving the microphone towards the speakers shortens the direct signal path, and lengthens the reflected path.

 

[MediumRare]

First of all, you must have all bass sources operating to optimize the bass in your room. So having both speakers going is certainly required. Second, if you use a sub, you must have it engaged as well. One of the benefits of having a sub cross over at 80 to 100 is to provide some additional pressurization to smooth out nulls resulting from the mains bass contribution.

So, first step, try with your sub and report the REW FR.

After that, you can try moving the sub to see the effect - even a foot to either side will make a measurable difference.

After that, try moving the mains a few inches farther apart or closer together. (Toe-in won't affect bass.) If it is cancellation you will be able to measure a difference. In the end some EQ may be required, but no more than 6 dB.

 

[JLM1948]

I am, of course, talking about path differences of room reflections. If, say, rear wall reflections are causing the null, then moving the microphone towards the speakers shortens the direct signal path, and lengthens the reflected path.

Since cancellation at 100Hz implies a length difference of 5ft, moving the mic by 2ft towards the speakers will be hardly distinguishable.

 

[MediumRare]

Since cancellation at 100Hz implies a length difference of 5ft, moving the mic by 2ft will be hardly distinguishable.

That is incorrect. You're not listening to entire wavelengths, you're hearing the pressure waves at your ear opening - which include cancellations and modes at the point in space your ear is. Play a pure tone and move around your room measuring the SPL at that tone. You'll see huge changes in small movements. With full-spectrum music those changes are masked, but they are definitely measurable,

 

[JLM1948]

That is incorrect. You're not listening to entire wavelengths, you're hearing the pressure waves at your ear opening - which include cancellations and modes at the point in space your ear is. Play a pure tone and move around your room measuring the SPL at that tone. You'll see huge changes in small movements. With full-spectrum music those changes are masked, but they are definitely measurable,

It may with a 1kHz tone, but here we're dealing with 100Hs, where the interaural difference is much less discernable. Now if you think instant pressure does not contain frequential information... Please.

 

[MediumRare]

It may with a 1kHz tone, but here we're dealing with 100Hs, where the interaural difference is much less discernable. Now if you think instant pressure does not contain frequential information... Please.

Only considering bass frequencies here. Do the test for yourself.

 

[johnnysix]

I find that when measured together, mine produced a massive dip at 50Hz. I can’t under stand why stereo pair self-interference isn’t discussed more. I had to employ subwoofers to defeat it. (Just got 2xKEF KC62, but had REL T7i’s before. )

 

[Taryn Frenzel]

Not an expert but can tell you what I have played with. I have messed with this as well (sort of). There is a function in REW (I think the beta version only) that has time alignment tool (hit controls at the right and then time alignment tool). Its supposed to be used to align your subs to your mains around the xo point but I have also used it to tweak any two curves to see what they sound like together. You can measure your left (however you listen to them which is either at xo point or full and right separately and then together and put the left in the top drop down and the right in the bottom drop down and it will calculate a curve that shows what the two together will look like. You can display both together and you will see it is a very close laydown. It will show you the 100 hz null (calculated) that you are seeing. Now take either the left or right speaker and slide the delay bar either negative or positive to see if you can eliminate the null by adding a delay to one of the speakers. You can then simply put the distance added to your receiver or subtracted from whatever YPAO had and record both together and the null should be gone (gives it to you in ms for minidsp, and ft for a distance adj). You can also do this by trial and error but much easier to do this in REW so you have a starting point. I don't think you have to worry about the time delay as we are doing this in the bass region anyway to time delay out subs and 100 hz is technically still a bass region. I have read that people can't notice a delay of 15 ms at the low end and 150 ms at the high end. And 5 ms would be a large adjustment so likely you won't notice anything except a stronger bass beat which can be 100 hz. Keep in mind, this new curve could alter your subs influence in the 100 hz region so be sure to add your sub back in to insure everything is behaving. Otherwise you can use the same process to time align your sub as well by adding or subtracting distance to your sub to have an optimum curve.
There is always a trade off. You will sometimes lose a little in a different area but hopefully smooth and not a null. And then of course, listen to see if you like it. There is one video out there that shows how to use this tool for aligning subs to your mains which is where I noticed this calculated curve matching my measured curve.

 

[JLM1948]

I find that when measured together, mine produced a massive dip at 50Hz.

That implies that there is something very wrong in the installation.
50Hz has a half-wave length of about 11ft. So if your speakers are separated by 11ft or more, there's gonna be places where cancellation occurs, but never in the bisector plane, that is if there is no other acoustic phenomenon occuring (Helmoltz or flexible panel resonator, labyrinth),

I can’t under stand why stereo pair self-interference isn’t discussed more.

Because it's been beaten to death in acoustics books. Anyone who's studied acoustic science knows about that.

 

[bamtron]

They appear to be correct. Although maybe it’s possible one of the drivers is wired with reverse polarity in the speaker?

Have you checked connections at the back of the amp?

 

[JLM1948]

Only considering bass frequencies here. Do the test for yourself.

I'm not talking about how nodes and antinodes are located. I'm saying that , IF this dip was due to negative interference, it would not change much with distance. Just simple geometry.

 

[JLM1948]

maybe it’s possible one of the drivers is wired with reverse polarity in the speaker?

That could be an explanation if that speaker was operating only between 80 and 120 Hz. Or maybe there's a huge phase rotation in this range... Do you have something like a digital EQ in the chain?
What speakers are they?