Averaging Listening Positions when doing multi-channel equalization. (1 Viewer)

[Larry Chanin]

Hi Wayne, Greg and Peter:

Isn't this the formula for room modes (peaks) not nulls.
Do you agree?
Can someone post both formulas to avoid confusion?
Thanks very much.
Larry

 

[brucek]

Larry,

Are you saying the mains are high passed at 12dB and the sub is low passed at 24dB? If so then in theory you'd calculate the midpoint within the bandpass, but our hearing acuity isn't too good down low WRT phase so averaging them is 'close enough' IMO, or 18dB/octave - 135deg..

I find this interesting. It's new news to me - could you elaborate please (wayne or greg). One of the features of a processors' bass management is the ability to compensate for phase/time differences in multiple speaker situations. A typical crossover for processors is a 2nd order 12dB/octave HPF and a 4th order 24dB/octave LPF. When we enter a simple distance from the listening position for each individual speaker during the setup of a processor, this compensates for a speed of sound (~1foot/msec) and adjusts the phase accordingly to the appropriate channels knowing the steepness of the crossover, does it not? It's just a computer, it certainly has the ability of making such simple decisions. Why would I add extra phase delay in my subwoofer signal after the fact. I can understand once the system is operating to make fine phase adjustments, but I don't really understand the extra initial default adjustments when we're using a bass managed processor.I can understand it if you're saying these should be dialed in as a starting point when using a subwoofers internal crossover LPF in an older unmanaged system...

brucek

 

[Larry Chanin]

Hi Bruce:

I think not. You could readily measure any audible noise with your RS meter. You would also hear it.... Are you saying you think you have a 60Hz hum in your room - no way.
This is a typical peak caused by your room that you'll be able to EQ out very easily..

Well, I guess you're probably right. I had my sound pressure level meter all the way down to the 60 dB scale, and I checked for ambient noise prior to the measurement. I didn't see any needle movement. It just looked suspicious to me that for almost a half a second the level stayed about the same at 60 Hz. I have two aquariums in the room a smaller one in the back. They have filters and air pumps operating at 60 Hz. I turned off the larger one, but the smaller one was so quite it didn't. As I mentioned I didn't see movement on the meter, but I figured that perhaps ETF5 was reading the meter signals much more precisely than I could by eye. You know, the needle has inertia of rest and the signal is changing faster than the needle can react, but perhaps not faster than ETF5 can record the electrical impulses. Anyway it sounded good at the time.
There are, of course, two ways to check. 1) See if I can equalize the peak away, or 2)turn off the fish tank accessories and run another set of measurements to see if it goes away.
One thing is for sure, it has nothing to do with an individual speaker since the condition repeats itself regardless of the speaker being measured. Unfortunately this supports either the room response or noise theory.
Thanks again for all your help.
Larry

 

[brucek]

Larry,
Yep, and I don't mean to disagree with you, but I think my reasoning on this is sound (pardon the pun).
If your mains are outperforming your sub above 35Hz, then your sub is either challenged or is in the wrong location for the room. Perhaps an SVS or a Servo-15 is in your future.
If you try and produce bass frequencies as low as you are from multiple speakers at fixed unoptimized positions in a room, it is a sure recipe for cancellation trouble that you may not be able to EQ away.
Generally, as has been discussed, a speaker should have a response about an octave below a crossover selection. So, a crossover of 80Hz should be fine for a speaker that extends to 40Hz, and if a crossover of 40Hz is selected, then the speaker should extend to 20Hz. If it does not, its natural rolloff will come into play and interfere with the processors controlled rolloff.
My suggestion would be to cross those mains at 80Hz and move that main sub until it gives a better response..... - at least as a test sometime......
I would also set the Subwoofer Peak Limiter control for minimal effect...
brucek

 

[Larry Chanin]

Hi Bruce:
Thanks for taking the time to explain this to me.
I'll give the 80 Hz crossover a try. I assume the sub should be set to 80 Hz as well?
To be honest with you undoing the Subwoofer Peak Limiter feature right now that I'm running high volume test tones and about to start fiddling with equalizers scares me silly. If I felt I knew more about what I was doing I'd be more comfortable undoing this safety feature.
I'll probably hold for a while on moving anything until I finish all the testing and start running the Bijou. The interconnects got lost in the mail :frowning: and my supplier says he'll ship a replacement set tomorrow. So I'm still waiting before I can do any serious damage.
Thanks again.

 

[brucek]

Larry,

I assume the sub should be set to 80 Hz as well

Could you explain this statement. I hope you're not saying you're setting the internal low pass filter control on the sub itself. The LowPass filter on the sub itself of course should be dialed to maximum, so as to be "out of the way" and not to interfere with the already managed frequency range that is sent to it from the processor. Some subs (I'm not familiar with yours) also have a disable or bypass for the internal low pass filter.
Is this what you're indicating?

brucek

 

[Larry Chanin]

Hi Bruce:

I'm sorry for being unclear. I meant I assume I should set the Mains crossover point in my Lexicon to 80 Hz and the crossover point for my sub in the Lexicon to 80 Hz. Yes, I would be setting the sub's crossover controls to the maximum.

Thanks.

Larry

 

[Wayne A. Pflughaupt]

brucek,

I find this interesting. It's new news to me - could you elaborate please (wayne or greg).

I’ll be the first to say I don’t get the phase/distance/time thing. This is Greg’s field of expertise – perhaps he’ll have some succinct comments for us.
I can tell you, however, I made some experimental changes to my set-up after seeing a post Greg recently made on this topic, and the results were astonishing. He claimed that if the subs were near the mains and the high and low pass crossovers were 24dB/octave, the phase setting should be 180 degrees. That description fits me to a “T,” so I thought I’d try it. I have an outboard crossover that doesn’t have a phase control, but it was an easy thing to flop polarity on one of my subs.
I had always taken my 1/6-octave sine wave sub readings with the mains off, and knew I had existing problems with a 1/6-octave null at 71Hz and a 1/6-octave peak at 45Hz. I took new baseline readings with the subs in phase and the mains on, and noticed that the two problem frequencies were worse than with only the subs on.
However, I was amazed to find that when I reversed the sub’s polarity the 45Hz and 71Hz problems virtually disappeared (the mains were still on). Measured response went from ±30dB to ±6dB!
There’s still a lot about this I just “don’t get:”

• If this is a time/distance thing, how does the fact that the subs are a few feet from the right main, and about 12-15 feet from the left main enter the equation?
• It is readily observable that if you reverse the polarity on a typical full-range L/R speaker set that the bass is “sucked out,” even though no changes to time or distance have been made.
• The bedroom tests I mentioned in the earlier post (the one where expected nulls didn’t materialize) I had the two subs in separate corners (sharing a common wall), and I reversed polarity for one test. The expected “suckout” didn’t happen, but some other strange things did – namely that with broadband pink noise the SPL level was 9dB down at any location equidistant from the two subs, and gradually increased as you moved toward one sub or the other. Again, there is no difference in time or distance with the subs in phase or out of phase, but the results were very different.
• If these scenarios are somehow related to a time/distance situation, then it could be duplicated or eliminated by applying digital delay – right? Someone please convince me.

Regards,
Wayne A. Pflughaupt

 

[brucek]

Larry,

I have an outboard crossover that doesn?t have a phase control, but it was an easy thing to flop polarity on one of my subs.

Well that was my point in a previous post. I'm sure those phase delays of 135degree for 18dB/octave are correct for external crossing, but when a processors bass management takes care of it, the knowledge of its own HPF/LPF is known and a simple proper phase delay would be present and taken care of by the processor. Alternately, it has no knowledge of your speaker distances, so you must enter this info on setup, so the processor can add a group delay to give time alignment to the main seating position used. "Time" is used as the fixed variable to adjust for distance because it alters phase with frequency. If I make the rough approximation that sound energy travels at 1ft/msec, then if I enter a fixed 1 msec delay to all signals that leave a signal port, then this is the same as moving a speaker 1 foot away from the listener. With a fixed delay there is a linear increase in phase shift with frequency, so if you pick a frequency of say 200 Hz which has a period of 5 msec, then 1 msec would be a 1/5th of the delay producing 72 degrees (72/360) of phase shift. The obvious frequency would be that 1000Hz produced one cycle of phase shift (360/360). This is how time is used to time align speakers.
If I introduce a fixed "phase" adjustment though, of say 135 degrees for all frequencies, you can see how that is different from using fixed "time" as explained above.
A crossover will introduce certain phase anaomolies depending on its type and slope and I'm fairly sure that a processor handles these since the designers know their crossover and what phase changes it would introduce if not adjusted for....... Again, this is my take on it.
brucek

 

[Greg Monfort]

>You mean I can't equalize 700 pounds of water out of the way?
====
Based on the plot, nope, your speaker doesn't have enough dynamic headroom.

GM

 

[Greg Monfort]

WP:

>Help me out here – I have no idea what you’re referring to.
====
I was alluding to your confirming my response on phase in another thread and wanted you to check on the null to confirm as it's been a long time since I fooled with standing waves in a bounded space and didn't have time when I posted to give myself a refresher course.

My memory is pretty much shot though and should have kept quiet until I'd done so. Anyway, I just went back and edited my posts so please re-read them again.

Sorry for the confusion. :b

GM

 

[Greg Monfort]

Brucek:
>I find this interesting. It's new news to me - could you elaborate please (wayne or greg). One of the features of a processors' bass management is the ability to compensate for phase/time differences in multiple speaker situations. A typical crossover for processors is a 2nd order 12dB/octave HPF and a 4th order 24dB/octave LPF. When we enter a simple distance from the listening position for each individual speaker during the setup of a processor, this compensates for a speed of sound (~1foot/msec) and adjusts the phase accordingly to the appropriate channels knowing the steepness of the crossover, does it not?
====
Yes, AFAIK based on other's comments in other threads (don't have a processor), but if WP's experience is typical then it doesn't do it for the LFE's XO phase shift WRT the other speaker's acoustic phase, so the sub would need adjustment.
====
>If this is a time/distance thing, how does the fact that the subs are a few feet from the right main, and about 12-15 feet from the left main enter the equation?
====
Well, it does, in the form of comb filtering in the overlapping BW between the sub and furthest speaker. Ideally, the sub would be equidistant from them to minimize this, but then this causes more compromises/problems than it solves in most cases.
====
>It is readily observable that if you reverse the polarity on a typical full-range L/R speaker set that the bass is “sucked out,” even though no changes to time or distance have been made.
====
That's because they are reproducing ~ the same BW, whereas between mains/sub this only occurs in the overlapping BW, which shows up as comb filtering.
====
>The bedroom tests I mentioned in the earlier post (the one where expected nulls didn’t materialize) I had the two subs in separate corners (sharing a common wall), and I reversed polarity for one test. The expected “suckout” didn’t happen,
====
There already was a suckout due to opposing corner cancellation, and why 'stereo' corner subs are frowned upon if max boundary gain is desired.
====
> but some other strange things did – namely that with broadband pink noise the SPL level was 9dB down at any location equidistant from the two subs, and gradually increased as you moved toward one sub or the other. Again, there is no difference in time or distance with the subs in phase or out of phase, but the results were very different.
====
Sure there is, it's you moving the mike in relation to them.
====
>If these scenarios are somehow related to a time/distance situation, then it could be duplicated or eliminated by applying digital delay – right? Someone please convince me.
====
Yep, you could digitally move one to the other side, but it would only be valid for one seating position. :frowning: Best to just stack them together.
GM

 

[Larry Chanin]

Hi Bruce:
I've tried to study this a bit more based on your great explanations regarding room modes. I found an excellent resource at Systems Development Group Acoustic Calculator. Below is a link to a picture of an Exel spreadsheet that I created by copying their format and using your formula. (It of course agrees with the web-based calculator.)
Predicted and Actual Room Modes
I've added a column (in white)to the table to show actual measurements that I have taken off of the this Low Frequency Response. The curve in blue is the Left Main that started this thread. You'll notice that surprisingly (to me anyway) there is a significant correlation between the predicted results and the actual measured results. With two notable exceptions. First, nowhere is there a predicted 60 Hz peak. Second, and this is to be expected, there are some missing predicted peaks around the 100 Hz area where the fish tank is doing its thing.
So, I'm thinking again that something besides a normal room mode is going on at 60 Hz. Maybe that noise theory isn't so crazy?
Thanks again for those great explanations.
Larry

 

[Larry Chanin]

Hi All:

Earlier in the thread I frankly was quite confused about Room Modes and Nulls, and obviously my poortly worded questions reflected that fact. Now, I believe that I'm beginning to see the light. So, hopefully this question won't prove that my optimism about my emerging understanding was misguided.

It would appear that the frequency of room modes and the physical location of nulls (at those frequencies) is totally dependent on the room dimensions. If this is correct, then the distance between speaker drivers and boundaries has nothing to do with the room modes frequencies and null locations. What it does effect is the amplitude of the room modes.

Is this correct?

Thanks.

Larry

 

[Greg Monfort]

>So, I'm thinking again that something besides a normal room mode is going on at 60 Hz. Maybe that noise theory isn't so crazy?
====
Is there an exposed side of the fish tank that's parallel to the back wall and ~18.83ft away? Any AC hum coming from the HVAC ducts, lights, whatever? A problem with the soundcard?

GM

 

[Greg Monfort]

>It would appear that the frequency of room modes and the physical location of nulls (at those frequencies) is totally dependent on the room dimensions.
====
Assuming something isn't blocking them, like a full 'wall' entertainment center, which defines the new room boundary.
====
> If this is correct, then the distance between speaker drivers and boundaries has nothing to do with the room modes frequencies and null locations. What it does effect is the amplitude of the room modes.
====
Right, if they are at the same frequency, otherwise they make their own. Then there's all the standing waves between all the other hard parallel surfaces.....

GM

 

[Mark Seaton]

It would appear that the frequency of room modes and the physical location of nulls (at those frequencies) is totally dependent on the room dimensions. If this is correct, then the distance between speaker drivers and boundaries has nothing to do with the room modes frequencies and null locations. What it does effect is the amplitude of the room modes.

Is this correct?

EXACTLY!

The frequency of the modes will not change, only how we excite them. This is why I have been asking about different peaks to identify which modes are causing the issues. The whole confusion before on my comments about not always using corner placement comes down to evaluating the application and what will happen with these room modes. Axial modes, the most prominent, will occur at multiples of the 1/2 wavelength which equals the dimension. If you start writing down where these will occur, it is quickly observed that the lower in frequency the fundamental, or the longer the room dimension, the closer these peaks will be to eachother, making for a smoother response. Smaller, mathmatically divisible dimensions will have more gaps in between room modes.

Now on to applicaiton... MOST subwoofers, dual SVS owners, ContraBass owners, etc. excluded, are output and distortion limited below 25-30Hz. The first thing to realize is that an overdriven speaker sounds BAD, especially bottoming drivers, and should be avoided best we can. The next concern should be the response. Smoothest response before EQ is what is desired, as peaks can be very offensive as well.

With output limitation being so common, most subs will benefit most from corner placement where they will get the maximum boundary re-enforcement. The goal is to keep the sub within its operating limits, and there are a variety of ways to help the situation. Moving the subwoofer closer to the listener can sometimes be just as effective.

When looking at the effects of room modes, we have to consider the results of exciting them. The spacing of the room modes and the response of the subwoofer have to be considered. While it is not always the proper choice, if we do know where the seating locations are, there are some interesting things which can be done with placement of the subwoofers along a dimension to reduce the excitation of specific modes, and to keep the seats out of resulting hotspots. I believe you can find a general explanation of this at the Stereophile Guide to Home Theater website in Russ Herschellmann's articles. There is also a spreadsheet which helps predict some of these effects. Yes, there is a bit more to it, but it does help to understand what is observed in measurements, and aids in knowing how to attack the problem.

Enough for one post...

 

[Larry Chanin]

Hi Greg:
Thanks for the response.

Quote:

The face of the entertainment center is 20.5' from the rear wall. While most of it presents a solid "wall" to the room, in the middle of it over the HDTV is a large hole that exposes the front wall of the room to sound waves.
A view of the hole (and the large aquarium causing the 100 Hz dip). The space behind the center channel speaker is entirely open.
If we recalculate the room modes using a 20.5' length the closest mode to 60 Hz is at 55.1 Hz, but none of the other predicted modes for this dimension match up with the observed results. In other words, with the big hole in the entertainment center the room "sees" a 24' room and the measured results confirm this.
Thanks.
Larry

 

[Greg Monfort]

>Yes, there is a small fish tank that is in the rear left side corner of the room (when facing the TV) at that approximate distance. It is not exposed meaning that the filter and motor is inside a closed top.
====
But if there's line-of-site between it and another parallel surface it will have its own standing wave modes. If so, then have someone stand in front of it to see if affects the 60Hz peak at all.
====
>The sound card is high quality and has never exhibited any problems whatsoever. As I demonstrated with the spreadsheet, for the most part there is good correlation between the predicted room modes and the actual measurements. I doubt I would have gotten consistent readings with a bad sound card.
====
Does the soundcard have a filter and/or shielding to ensure it doesn't pick up any AC hum from the computer? If so, and there's a problem, then does this mean the SC wouldn't otherwise work fine?
====
>I see where a 18.83 foot dimension corresponds to a 60 Hz room node, but I thought that if the 60 Hz peak was a room mode that it would be unaffected by the distance of the source (fish tank) and a boundary? There is no room dimension that is close to 18.83 feet.
====
As I noted, there will be standing waves between any two parallel surfaces or between a driven cone and a parallel boundary. How high in amplitude they are will depend on how rigid the boundaries are. Glass reinforced by water is extremely rigid. Don't know if this is it, just tossing it out as a possibility.
====
>If we recalculate the room modes using a 20.5' length the closest mode to 60 Hz is at 55.1 Hz, but none of the other predicted modes for this dimension match up with the observed results. In other words, with the big hole in the entertainment center the room "sees" a 24' room and the measured results confirm this.
====
Understood, I was just giving an example that would change it. 55-60Hz and the subsequent modes are close enough together that they would ~sum and change the amplitude and Q of the peaks/nulls.
Nice looking room BTW.
GM

 

[Larry Chanin]

Hi Mark:

Thanks for the response.

The frequency of the modes will not change, only how we excite them. This is why I have been asking about different peaks to identify which modes are causing the issues...

When looking at the effects of room modes, we have to consider the results of exciting them. The spacing of the room modes and the response of the subwoofer have to be considered. While it is not always the proper choice, if we do know where the seating locations are, there are some interesting things which can be done with placement of the subwoofers along a dimension to reduce the excitation of specific modes, and to keep the seats out of resulting hotspots. I believe you can find a general explanation of this at the Stereophile Guide to Home Theater website in Russ Herschellmann's articles. There is also a spreadsheet which helps predict some of these effects.

As you may have surmised from reading this thread, your questions regarding the exact value of 60 Hz peak (whether intended, or not) has caused me to critically evaluate whether or not I am dealing with a room node in the first place, and the exercise has been most instructive (even if I reach the wrong conclusions. )

As a rookie I have discovered that dips aren't always nulls, they might be caused by the placement of fish tanks in front of speakers. Likewise, peaks might not be room modes, they might due to noisy fish tank filters, or other causes.

I've taken a quick look at Mr. Henderson's spreadsheet, and its reassuring to see that it agrees with the one that I posted earlier. In particular, it fails to show a predicted 60 Hz room mode.

Do you have an opinion whether the observed 60 Hz peak is a room mode, noise or some other explanation? Can the Time Slice curves that I posted earlier, showing the persistence of the 60 Hz peak over time, be used as evidence supporting noise as the cause?

Thanks.

Larry