Preparing for REW+BFD

[Trekari]

So..I set phase manually and then ran Audyssey today. Then I measured manually with my ratshack SPL meter and imported the data. Here is 10Hz-100Hz with Mains and both Subs active at listening position 1!

Corrected RS values are in the blue chart below. Not a bad starting place I think!

[brucek]

Then I measured manually with my ratshack SPL meter and imported the data.

Why????

[Trekari]

Because it's a chore to move my computer tower into the living room, then find the mouse, keyboard, monitor, etc to go along with it.

My desk where all my computer cables go is not exactly accessible, so it truly is easier to measure manually. It might take more time, but it requires IMO less work.

THe only drawback I see so far is that I don't seem to have a way inside of REW to apply the RS calibration curve to what I manually import?

[Trekari]

Ok..so I went back and re-tested the obvious problem area at 91-Hz or so.

With the RIGHT subwoofer firing along with the mains, I get
91Hz 69
92Hz 67
93Hz 65
94Hz 64

With only the LEFT subwoofer firing and the mains I get:
91Hz 66
92Hz 63
93Hz 63
94Hz 65

Is this a phase problem with the left subwoofer? I set the phase identically since they are equidistant from the listening position. All main speakers are Xover'd at 80Hz, and the "LFE" Xover is 90Hz. If this were a Xover problem, wouldn't it be logical to assume both subs would be equally affected?

**EDIT#2**

Apparently it was a phase problem. At the very least, reversing the phase of the second (LEFT) subwoofer then commencing playback with both enabled resulted in:

71, 70, 69, 68, 68 and 69dB readings from 91Hz through 96Hz. I'll report back with further testing after brucek admonishes me for doing this manually and tells me to haul my computer out into the living room no matter what I 'enjoy'

[HClarkx]

Trekari wrote:

Ok..so I went back and re-tested the obvious problem area at 91-Hz or so.

Is this a phase problem with the left subwoofer? I set the phase identically since they are equidistant from the listening position. All main speakers are Xover'd at 80Hz, and the "LFE" Xover is 90Hz. If this were a Xover problem, wouldn't it be logical to assume both subs would be equally affected?

Trekari, You certainly fixed the low-end with the new subs!!

The 91 Hz problem isn't due to phase difference between the two subs unless your mic was off to the left or right so that it was 6.15 feet further from one sub than the other. You will have a notch that moves in frequency and depth as you move away from the center line between the two subs. That's unavoidable. It will start at 0 db and increase in depth as you move off to the side and the frequency at which it occurs will drop. It won't get down into the sub range until you are several feet off to one side or the other. This is the same problem most dual-woofer center channel speakers have.

The above is based on ignoring reflections so may not be as obvious as I make it out to be. I.e., reflections will add their own notches and maybe mask the ones from the direct wave.

A question though, your green plot shows a 10-15 db notch while your per-speaker measurements indicate half that for one sub and nearly flat for the other sub. Something doesn't add up. Were the measurements take with the RS on a tripod and not moved?

Was the RS meter pointing straight up at a fixed position where your ears would be and with nothing else near it? That's important since the only way to make the mic omni-directional to sound coming from the subs and reflecting off the walls. It will be more sensitive to sound coming off the ceiling, but that's unavoidable.

Your ears are nearly omni-directional at sub frequencies so you want the mic to be omni-directional as well.

It would be interesting to see the plot with Audyssey turned off. The notch might be 10 db deeper. If so, that gives you some idea of what's causing the problem.

Audyssey will reduce the notch well if it appears at all measurement locations. If it's only at some measurement locations, those will be improved possibly with a bit of a peak introduced at the same frequency at other locations where there was no notch.

It would also be interesting to see the pattern at each of your seats. If the notch only affects some seats, you might not be able to improve it without degrading sound at other seats. Or moving subs or such.

Can you give us a floor plan of the room with all openings to other rooms shown?

Long story short, it sure looks like you have a reflection problem. Are the subs both facing a back wall that is parallel to the wall they are on? If so, I'd bet you can measure the distance from your subs directly to the mic location and then from the subs to the back wall and then forward to the mic. The difference will be 6.15 feet. Either that or the distance from the subs to the ceiling and back down to the mic is 6.15 feet further than the direct path. Ditto for the left and/or right wall. The back wall will give a sharper, deeper notch than the ceiling and walls.

From this you can see some issues becoming evident. For instance, you don't want mid-room with a sub in front of you and a wall behind you. There are sub locations that will kind of spread out the reflections so that none of them are horrendous. The trick is keeping the subs and your mains equidistant from the prime seat, though adjusting sub phase carefully helps a lot.

Harrison

[Trekari]

The RS meter was in fact mounted on a tripod and was not moved during any of the measurements. I was sitting in the seat to the right of it... but I was in that spot for *every* measurement.



The kitchen is open to the living room, with a 34" or so high counter behind the love seat. The 'hallway' going down is into the kitchen, and towards the front door.

The hallway going Left is towards the bedrooms.

The ceiling is a cathedral ceiling, with the sofa and TV being positioned directly underneath the apex, which is 11' high. The ceiling does not slope back down to 8' high until the sliding glass door and the bottom surround speaker (about 8" below that speaker, actually).

The mic was positioned on the center seat of the sofa - the wall to it's (on the image at least) down-left was perhaps 18" away, and the wall directly left of it is 1' behind the sofa.

Now that I have the new subs positioned in what I'm hoping can be a 'final' spot (ever try moving 130lb subs that are on spikes and discs?), I plan on doing some serious measuring work in the near future. I currently have another 1/4" TS to RCA cable pair coming along with a Y-adapter to split my sub pre-out to go into L and R inputs on the Behringer, then to each sub.

Thank god for you guys who know what you're talking about/doing. You know FAR more about audio and acoustics than I do, and I appreciate every input you offer, even if I don't necessarily like your replies

**EDIT** And yes, the low end definitely got fixed I should also mention (if it matters?) that EVERY speaker is actually aimed at the center listening area of the couch, including the subs.

[HClarkx]

Trekari, Good information. Very helpful.

The wall that is a foot behind the sofa can be a problem. It's not the cause of your 91 Hz dip because it's not 6 feet from the mic to the wall and back to the mic (is it?). I'm assuming it's more like 4 feet. That will give you notches at your ears at about 280 Hz, 560 Hz, 1120 Hz, 2240 Hz, 4480 Hz, etc. (assuming exactly 4 feet). Audyssey MultEQ may be taking good care of those higher frequencies even though it isn’t eliminating the more severe 91 Hz one.

If the back wall is found to be a problem, you might want to look into an ornamental wall-hanging or something like that (heavy, with some sound absorbing material behind it) for the wall behind the sofa to reduce sound reflections that bounce back and partly cancel the direct wave. Even if Audyssey MultEQ is taking care of this problem, it can be a bit more effective if the problem is reduced to start with.

Your ceiling might be the problem or much of it. If you trace lines from your subs to the ceiling and down to the sofa, one on either side of the ridge line, following the equal incidence/reflection angle theory, you will find that the ceiling is kind of focusing any sound reaching it directly from the subs back down to the sofa. That is, you are getting more SPL at the sofa from your ceiling than you would from a flat ceiling. If it's 6 feet further (or 9 feet further ... 0.5, 1.5, or 2.5 wavelengths at 91 Hz) via the ceiling than it is direct from the speakers to the sofa, I'd bet on that being the major contributor to your 91 Hz notch. Not easily fixed!!

The openings to the hall and kitchen are potential problems, but are unlikely to cause a sharp notch at just one frequency. At most they would cause some ups and downs across the sub range but only small ones. The left wall is a possible minor issue and might be contributing to the notch but alone isn't likely the cause (it’s weakened by the fireplace).

I'm looking only at first reflections here ... the direct wave and the first reflection meeting at the sofa. Secondary reflections aren't likely to be all that damaging though they can aggravate a first reflection problem.

I’d measure all three seats. That will provide more info and maybe even cause you to decide to live with the problem.

How many mic locations are you measuring for the Audyssey MultEQ setup? You don't want to let the mic get too close to the back wall. Your ear location is as close as you want to get to the back wall in your case.

The subs don't need to be aimed at your center seat. The low frequencies flow out from the subs in all directions. You could turn your subs 90 degrees and hardly know the difference (though it might take a blind a/b comparison to prove that). The mains definitely should be canted in and pointing directly at your center seat. This is the best angle for most speakers.

I’d go after more information to narrow down the problem, but I think if you want to solve it and don’t want to give up the fireplace or slider, you may end up thinking about moving the left sub to the left of the fireplace and move the right one a bit further to the right, maybe even to the corner near the kitchen (where the bookshelf is). The phase would need to be advanced a bit because they would be further from your seat, but you might lose that 91 Hz notch.

I’m still concerned about the green trace not matching your RS measurements. Can you clarify that?

I'm with you ... my subs are about 250 lbs and 43" tall ... hard to move and not many acceptable locations. I lucked out with a very agreeable wife and some available ideal locations.

Sorry about my writing problem.

Harrison

[Trekari]

House is on tv right now (great show!) but I'll be back in a while to post more information.

For now, here is a link to photos of my current living room setup (just pretend the JBL sub isn't there )

http://www.hometheatershack.com/foru...html#post74968

I'll answer your questions as best as I can when House is done.

Thanks again for your time and help!

[Trekari]

HClarkx wrote:

Trekari, Good information. Very helpful.

The wall that is a foot behind the sofa can be a problem. It's not the cause of your 91 Hz dip because it's not 6 feet from the mic to the wall and back to the mic (is it?). I'm assuming it's more like 4 feet. That will give you notches at your ears at about 280 Hz, 560 Hz, 1120 Hz, 2240 Hz, 4480 Hz, etc. (assuming exactly 4 feet). Audyssey MultEQ may be taking good care of those higher frequencies even though it isn’t eliminating the more severe 91 Hz one.

**EDIT** It's actually more about 32-36" (I measured to the general area the tripod meter was at)

If the back wall is found to be a problem, you might want to look into an ornamental wall-hanging or something like that (heavy, with some sound absorbing material behind it) for the wall behind the sofa to reduce sound reflections that bounce back and partly cancel the direct wave. Even if Audyssey MultEQ is taking care of this problem, it can be a bit more effective if the problem is reduced to start with.

Hopefully you've seen the links in the previous post. I DO have something on that back wall. I'm not sure it's enough (and of course we....ok, ok, you aren't sure what the problem is yet)...(of course I'm not offering any theories as I'm not qualified to do so, so you're the only intelligent one here)?

Your ceiling might be the problem or much of it. If you trace lines from your subs to the ceiling and down to the sofa, one on either side of the ridge line, following the equal incidence/reflection angle theory, you will find that the ceiling is kind of focusing any sound reaching it directly from the subs back down to the sofa. That is, you are getting more SPL at the sofa from your ceiling than you would from a flat ceiling. If it's 6 feet further (or 9 feet further ... 0.5, 1.5, or 2.5 wavelengths at 91 Hz) via the ceiling than it is direct from the speakers to the sofa, I'd bet on that being the major contributor to your 91 Hz notch. Not easily fixed!!

I'm not sure where to measure since the ceiling is slanted, wouldn't that make it very difficult to determine the reflection point without a lot of trigonometry? I do understand the concept that it would act as a focus. The only way to solve something like that would be some serious acoustic treatment, since just hanging some decorative tapestries aren't enough to stop bass waves, right?

The openings to the hall and kitchen are potential problems, but are unlikely to cause a sharp notch at just one frequency. At most they would cause some ups and downs across the sub range but only small ones. The left wall is a possible minor issue and might be contributing to the notch but alone isn't likely the cause (it’s weakened by the fireplace).

Hopefully you saw the other photos and you'll see that aside from hanging beads or a tapestry from the top of the hallway, there isn't much that can be done there without angering the landlords (parents).

I’d measure all three seats. That will provide more info and maybe even cause you to decide to live with the problem.

I will get to that as soon as my other Behringer cables show up, that way I don't have to be setting up and taking down my computer multiple more times . If my cables get here, then whatever measurements and sub locations we come up with can immediately be used to apply filters to my 2496 so I can just take the PC out there one more time.

How many mic locations are you measuring for the Audyssey MultEQ setup? You don't want to let the mic get too close to the back wall. Your ear location is as close as you want to get to the back wall in your case.

Well, this morning when I ran it (couldn't yesterday because it was VERY windy), I only used 3 positions, each on the 'ear' point of the sitting locations of the 3 cushions of the sofa. Speaking 2D, each position was equidistant laterally from the subs.

When I get this 'right' so to speak, and I know that I can move forward with final measurements, settings, locations, etc., I plan on taking 8 measurements from the entire couch area, including the coffee table.

The subs don't need to be aimed at your center seat. The low frequencies flow out from the subs in all directions. You could turn your subs 90 degrees and hardly know the difference (though it might take a blind a/b comparison to prove that). The mains definitely should be canted in and pointing directly at your center seat. This is the best angle for most speakers.

Yea...well...they look intimidating that way!

I’d go after more information to narrow down the problem, but I think if you want to solve it and don’t want to give up the fireplace or slider, you may end up thinking about moving the left sub to the left of the fireplace and move the right one a bit further to the right, maybe even to the corner near the kitchen (where the bookshelf is). The phase would need to be advanced a bit because they would be further from your seat, but you might lose that 91 Hz notch.

I can't give up the fireplace or even that side of the room. That's my dart board location - which makes it the third dart board location in that room. Parental landlords are getting ****** about the holes

I COULD however, move one of the sub to by the love seat. I don't actually have a shelf over there now (again, hopefully you saw the photos). In fact...I still have my old JBL subwoofer that certainly could augment frequencies?

I’m still concerned about the green trace not matching your RS measurements. Can you clarify that?

*shrug*

All I know is the original phase was 180 for both, and what I wrote down is exactly what was tested. And in fact, I'm dyslexic, sigh. Right meant Left in my chart and Left meant Right. (Corrected in this post)

LEFT subwoofer firing + mains:
91Hz 69
92Hz 67
93Hz 65
94Hz 64

RIGHT subwoofer firing + mains:
91Hz 66
92Hz 63
93Hz 63
94Hz 65

LEFT subwoofer (phase 180) + RIGHT subwoofer (phase 0) + mains:

91Hz 71
92Hz 70
93Hz 69
94Hz 68
95Hz 68
96Hz 70

Which fit into the values around them, as 90Hz was 69, 89Hz was 71 and 82Hz-88Hz were all 72.

Sorry about my writing problem.

Harrison

I appreciate each and every answer, particularly where I learn things while reading them.

[Trekari]

Ok, here goes. I know in my mind that these are significant.

http://www.hometheatershack.com/foru...html#post76177

(from a few replies up, my 4 graphs)

Green is where the JBL was in the recent photo on the left side of the TV taken from the couch seat on a tripod.

Red was a mirror image of the previous setup, but on the right side of the TV instead of Left.

Teal was in the corner with the port facing against my kitchen wall. (The corner of the Love seat/kitchen counter/front wall).

Gold was taken in that exact corner as Teal but with the port firing against the front wall.

As you can see, each of those pairs of locations (similar) have different areas of frequency problems. But the first two, each right next to the TV, have problems at 91Hz.

THOSE measurements however were taken with ONLY the subwoofer (JBL) active. These results today had/have problems while the mains are on too.

I dunno..I'm tired and thinking outside my boundaries of qualified subjects to try and troubleshoot.

[Wayne A. Pflughaupt]

Trekari wrote:

Because it's a chore to move my computer tower into the living room, then find the mouse, keyboard, monitor, etc to go along with it.

My desk where all my computer cables go is not exactly accessible, so it truly is easier to measure manually. It might take more time, but it requires IMO less work.

Understandable...

Why not just get a long extension cable for the mic and run it to your computer in the other room?

Regards,
Wayne

[Trekari]

Do you know how many cable orders I've made in the past few weeks? hehe

I'd need an extension for both the ouput and the mic. I'm running out of dollars to buy things

I'll be moving the PC as soon as my second set of Behringer cables arrives from BJC...I don't think they got them shipped out today

[HClarkx]

32-36" .... that's 16" or 18" from the wall to your ears (and the RS meter). Or 32-36" round trip, right?

You are right, if the back wall is a problem only at 198 Hz (assumes 34" round trip) that's easier to absorb than 20-100 Hz. But, a quick check around 198 Hz will tell you if you have a problem there ... one Audyssey isn't fully eliminating.

You're right, measuring the ceiling reflection isn't very easy ... and maybe not worth doing since there is no good solution if that is the problem and you don't want to move furniture. Your 91 Hz dip is fairly narrow ... you aren't losing much.

You didn't by any chance have to change the RS meter scale to capture that one low reading at 91 Hz did you? I've seen RS meters that don't give the same reading when you compare the low end of one scale with the high end of the next lower scale.

I'm curious. You have a receiver with MultEQ, right? Why do you need additional sub filtering. Especially if your Audyssey MultEQ has ALFC, it doesn't get any better than that.

With your seating Audyssey recommends a minimum of six readings, one at each of the three head positions and three more about 3' forward of those (I use 2 feet forward). I'd stay within that area with any additional readings. Personally, I use the extra readings at MY seat (the center one) to give it more weight.

I'm not sure you have a 91 Hz notch problem. If you get relatively flat with each sub individually, and both are similarly phased and a similar distance from the center seat, the combined subs should be relatively flat as well.

I'm at a loss to explain why you get more SPL with two subs than with either sub individually when one is run out-of-phase. I'm wondering if the phase switch on one of them is mis-wired. Very strange. In any event, both subs should be set at 0 degrees if that gives each a smooth blend with the mains.

I take it you don't have adjustable phase, just 0 or 180. In your situation both should always be 0 even if you move them left or right some distance. And, without adjustable phase, you probably want to leave them close to the mains if that doesn't give rise to any more than the narrow dip at 91 Hz.

But, if the green plot which shows a >10db dip at 91 Hz isn't right and your individual sub readings are, then I think we are chasing a ghost here and Your system is working very very well.

Harrison

[HClarkx]

Trekari wrote:

Do you know how many cable orders I've made in the past few weeks? hehe

Hmmm. Might have been cheaper to buy a used laptop!

[Trekari]

32-36 round trip, yes.

Yes, I did have to change the sensitivity on the SPL meter.

I was thinking of using the Behringer (now that I've seen how Audyssey performs with a quality sub) to create a house curve and clean up any leftovers from Audyssey.

I will measure the individual subwoofers as soon as I get the motivation to go through the sine waves twice again, or my cables get here and I'll move the PC for a day of measurements.


Here is a corrected RS values graph

I also measured (corrected values here too):
11Hz 66dB
12Hz --not measured
13Hz 65dB
14Hz 69.3dB



Extra weighting for YOUR seat eh? How many people in your house know you did that?

[HClarkx]

Tekari,

I'm not that familiar with the Behringer. It will be interesting to see what you can accomplish with it.

If you just want the Behringer to help flatten the bass response, my instinct is to attack any problems you can with it, but then do the Audyssey room correction calibration with it in the circuit. The Behringer contribution won't be defeated by Audyssey. All that will happen is that any problems that Audyssey wasn't able to fully fix alone will probably become fixable by Audyssey because you've reduced them with the Behringer. Room treatments can have the same benefit (though room treatments to fix problems below about 200 Hz tend to be huge and very intrusive on room decor).

Doing a final setup with the Behringer in the circuit will also allow Audyssey to correct any time domain anomalies introduced by the Behringer filters. I'm assuming the Behringer unit provides just frequency response correction and thus will introduce some time domain distortion characteristic of such devices.

That said, I'm not sure the Behringer will be very helpful at achieving a flat response. Does it have the resolution to bring up your very narrow 90-95 Hz notch without putting shoulders on either side of it? Also, if the Audyssey version you have has ALFC, it has 512 terms in its Finite Impulse Response filter all crowded into the sub region. That provides very high resolution correction of both frequency and time domain anomalies. That's hard to beat.

You mention doing a house curve with the Behringer unit. In that case you would let Audyssey do the room correction and then apply Behringer afterward to tailor the bass to your personal liking. You would do this at the risk of degrading the even time-response provided by Audyssey. I've found that the bass tone control in my pre-pro does that very well and shapes the bass response so that it provides a correction for the ear's weaker response to lower frequencies when the volume is reduced, e.g., from 105 db to 80 db. Hopefully the DSP-based bass tone control is not introducing time-domain distortions. If it's well done it shouldn't.

Cool stuff. You really need that laptop!

Harrison

[brucek]

That said, I'm not sure the Behringer will be very helpful at achieving a flat response. Does it have the resolution to bring up your very narrow 90-95 Hz notch without putting shoulders on either side of it?

The behringer equalizer is definitely resolute enough to flatten and remove any modal resonance peak at the point of measurement. The filters effects are felt both in the frequency and time domain. The filter(s) address both the increased amplitude of the modal resonance and the longer decay time resulting in an elimination of the effects. REW is used to measure and design the filters to be used to flatten the subwoofers response, and many member here use the Behringer in conjunction with receiver/processor auto EQ systems.

These auto EQ systems are usually quite effective above the room modes ~100Hz. The auto EQ system is initially set up for the best response and then the REW is used to flatten the crossover region and remove resonant peaks in the low frequencies.

brucek

[Trekari]

Brucek,

It is my understanding that whether I have filters engaged or not, that the Behringer adds 1ms processing time to the signal?

Therefore it shouldn't matter that I have no active filters if I run Audyssey with the Behringer in the loop, then start charting my frequency responses of the main speakers and subs in various combinations?

I plan on charting the main speakers by themselves at various crossover points until I get a nice flat graph, then charting the subs (individually) until I get nice response graphs from them at various xover points, then combining the two and making my house curve.

Hopefully this sounds like an appropriate procedure.

[HClarkx]

brucek wrote:

These auto EQ systems are usually quite effective above the room modes ~100Hz. The auto EQ system is initially set up for the best response and then the REW is used to flatten the crossover region and remove resonant peaks in the low frequencies.

brucek

I've seen your thoughts on this but haven't seen anything proving that the room decay is fully offset by the filter time response (i.e., that the two match). Is there some mechanical-electrical duality between a room and the filters that ensures a match? Also, can you refer me to tests that confirm the time components added by the filters correct room decay characteristics?
But assuming that all works as advertised, why do you suggest the auto EQ systems in the receivers are beneficial only above ~100 Hz?

I did have to move one of my subs to get rid of a 15 db null but once I had my combined subs within about +/- 5 db of flat, Audyssey brought me down to within a small fraction of a db when calibrating just one seat and within about +/- 1 db when calibrating three seats (i.e., no more than about 2 db difference between seats at any frequency). The later result obviously benefited from all three seats having nearly the same raw frequency response. This was the case down to 20 Hz (actually, the maximum 1 db deviation from flat occurred right at 20 Hz).

My situation wasn't particularly challenging, but I got excellent frequency response room correction down to 20 Hz (without Audyssey's ALFC). I can only assume that the time domain correction was equally well done and that Audyssey's claims of decay time improvement also are effective. I don't have a way to test that.

I've had the Sound EQ since they came out and have gone through different subwoofers and somewhat different room arrangements always with impressive results.

Does REW provide room decay info? Maybe I should give it a try.

Harrison

[Trekari]

REW does indeed provide a waterfall, decay, and impulse measurements in addition to the basic frequency graph you're accustomed to seeing.