How to get rid of an acoustic null in the room.

[Jimbroni]

A buddy borrowed me, a Behringer Ultra-Curve RTA. It came with a ECM8000 Behringer mike. I looked up the curve its pretty flat with about a 2-3 db boost at 2500 to 20K.

 

So I ran a pink noise signal through wavlab looped the signal, and checked the response of the RTA in several positions. At the mix position pink noise looked like it should proper slope and everything, but had a very significant dip at 125. Like nothing there at all dip. and smaller but noticeable dip at 400. I can see this becoming problematic in the future.

 

So I tried white noise. That was totally messed up. Instead of getting flat I got curve that gradually sloped upwards with increasing frequency.

 

My second position was the sofa bench behind the mix position. This corrected the white noise issue. However created a huge bump at like 60 and still had the null at 125 with the pink noise signal.

 

The room looks like (see attachment)

 

 

 

 

Sort of anyway. The small circles are the speakers and the big circle is the mix position

 

The 3 walls behind the speakers are hardwood with 12" fiberglass insulation boxes between the speakers and the wall. The 2 back walls are poured concrete and have 4x4 owens 703 panels. Used the mirror technique to place the panels to stop 1st reflections. The 3 corners away from the speakers have corner traps ceiling to floor. 2'x2'x33" triangle fiberglass covered with fabric. Small leather bench sofa in the corner furthest from the speakers was the second measuring position.

[Jimbroni]

It seems as though moving toward the back corner I can get more bass response enough to correct the white noise. However the dip at 125 with a pink noise signal doesn't seem to go away no matter where I go in the room.

 

Should I be alarmed or is more likely the measuring equipment, not having an adequate sampling window?

 

And if it is an issue any ideas you may have to help me get rid of the null at 125 would be helpful. Thanks.

[Jimbroni]

The dims of the room are

 

Front walls (speaker side) 8 ft x 8ft x 8 ft. Not exact just rough est thats pretty close.

Back walls 12 ft x 12 ft.

Ceiling 7 ft. With fiberglass insulation and covered by Anderson acoustic ceiling tiles.

 

Sorry to keep bumping but just want to give you guys as much info as I could as this stuff gets quite complex.

 

I'm thinking it maybe just that the control room is too small for the bass frequencies to cohere before they get to my ears.

 

Any ideas?

[Jimbroni]

Well as it turns out I just did a quick adjustment and I think I figured out what the deal is. Maybe.

 

I had my monitors set at sharp angle. Equilateral triangle with the monitors so that the triangle came together at my head. I turned the speakers out so that the triangle would come together behind my head about 1.5 to 2 feet. The nulls totally disapeered in the pink noise signal and the white noise signal flattened out in the mix position. Must have been a phase issue?

 

I still have problem in the rear corner where my clients would sit during play back. Its really bassy and appears to have some comb filtering issues.

 

Any ideas there? I was thinking of putting some bookshelves on the walls on both sides of the couch. Would that create more comb filtering or would it act as a diffuser that would scatter the lowend and prevent constructive interference in the back corner of the room?

[MrKnobs]

So I tried white noise. That was totally messed up. Instead of getting flat I got curve that gradually sloped upwards with increasing frequency..

 

Don't use white noise for measurement of this sort, always use pink.

 

Pink noise has equal energy per octave, white noise has equal energy per Hz, so white noise is much brighter than pink noise and not useful for room acoustics measurement.

 

Plus, if you crank it loud enough, you might blow your tweeters.

 

Terry D.

 

P.S. Can you post a graph of your room frequency response, after all the treatment you described is in place? Using pink noise and the measurement microphone where your head would be while listening? That would be helpful.

 

You're always going to have some degree of bass buildup along the back wall, unless you remove the back wall completely (hello, outdoors) or add so much absorption that you can barely fit in the room. Also, your room will never be completely flat but maybe it will be "good enough." This is my control room (showing modes) after adding a tube trap and regular LENRDS, but before I put in MegaLENRDs and RealTraps.

 

[Jimbroni]

Well my RTA is a stand alone outboard unit. No way to plot. However, I may be able to take photo. Like I said, just by adjusting the angle of the monitors slightly the curve seemed to flatten out.

[MrKnobs]

Well my RTA is a stand alone outboard unit. No way to plot. However, I may be able to take photo. Like I said, just by adjusting the angle of the monitors slightly the curve seemed to flatten out.

 

I understand.

 

I still use the image I posted, even though it's not the latest RTA of my room, because I don't have JBL SMAART anymore and I do have a handheld frequency analyzer. I suppose I could freeze the display of that and take a pic with a digital camera also.

 

Since you can't post an image, what are the frequencies you're having problems with, and what are the amplitudes of the humps and dips in the room? Using pink noise, measurement mike at your listening position, referenced to 1kHz at 0dB.

 

Terry D.

[Jimbroni]

The dip is at 125. Here is a pic of mix position.

 

And crap I realized I didn't fix the issue. I had messed with an eq to flatten out the response and didn't realize it was turned on so this is what I'm getting right now without the eq.

[Jimbroni]

The peak on either side of the valley jump around abit, but hover between -60 to -50. The peak on the left seeming to be a little lower within that range. The crevice at 125 is consistent at -70 to -75.

[Jimbroni]

I understand. referenced to 1kHz at 0dB. Terry D.

 

Are you saying my response should be 0 db at 1k if so I'm gonna need some ear plugs that'll be loud. Or are you saying the input/signal generator should be 0 db at 1K. or are you just asking if my mike is calibrated?

[Jimbroni]

Well I calibrated the mike to the spec in the manual and got the 1 K level up higher to 36 dB. Its a digital unit the max is 0 dB, so I'm totally confused by the reference comment. If you could explain when you get a moment.

 

Thanks.

[MrKnobs]

 

Well I calibrated the mike to the spec in the manual and got the 1 K level up higher to 36 dB. Its a digital unit the max is 0 dB, so I'm totally confused by the reference comment. If you could explain when you get a moment. Thanks.

 

 

It's not really necessary to do that, it's just customary to adjust the reference mike gain to set the 1kHz response to 0dB. I see that your device reads out absolute SPL so disregard.

 

All that really matters is the difference between the reading at 1kHz and your peaks and dips. The use of 1kHz as a standard is arbitrary, but you must have some base reference to relate peaks and dips to. You say your 1kHz is reading 36dB (that's very low, either due to quiet playback of the pink noise or a low setting of gain on your reference mike), so what are the frequencies of your peaks and dips (can't read the scales in your pic).

 

Just trying to ascertain how severe your room modes remain.

 

Terry D.

 

P.S. You say your problem dip is around 125 Hz. 1130fps/125Hz = 9.04 ft, which is consistent with your (3?) room dimensions of 8 ft or so. Where you have the width, length, and height of the room the same like this causes some strong cancellations and reinforcements. The answer is always the same: more/better bass trapping. I can't read your vertical scale, so I don't know the magnitude of that notch. Is it 15 dB?

[TimmyP]

A dip at 125 would tend to indicate that a speaker is 27" from a room boundary. Check http://www.padrick.net/LiveSound/CancellationMode.htm and better yet http://100db.com/misc/boundary.xls

[Ethan Winer]

 

A dip at 125 would tend to indicate that a speaker is 27" from a room boundary.

 

 

Right, here's mine:

 

http://www.realtraps.com/sbirlbir.htm

 

--Ethan

[MrKnobs]

 

A dip at 125 would tend to indicate that a speaker is 27" from a room boundary.

 

 

More likely 9 ft. He said (unless I'm mistaken), his room is 8' x 8' x 7'. That will mostly likely explain the 60Hz and 125 Hz modes - especially if his ceiling is actually 8ft with some ineffective (at bass frequencies) material dropping it down to 7 ft. At least that's how I understood his description.

 

In the real world, measured room modes don't often exactly match the calculated ones - they're just a rough estimate. Also, if you have a 125 Hz problem from a reflective barrier 27 ft away, you have a much less severe problem than one from 8-9 ft away.

 

Also, we've yet to see the graph, which is reality.

 

Terry D.

[Jimbroni]

Everyone,

 

Thanks very much for the responses.

 

I'm gonna take a measurement from my speakers to the wall closest/behind them(attack wall). But 27" sounds about right.

 

Knobs, if you look at the cheesy drawing I did with google sketchup. It shows the room is a five sided room. The 3 - 8 ft walls are behind the speakers with two 12 ft walls that converge to a corner behind the listening position. The ceiling ht is 7 ft. This makes it a bit more complicated as I have two horizontal boundaries for each speaker. Which could explain the small dip I was seeing at 400 hz? Between the boundaries and the speakers are placed Insulation boxes. Broadband not really traps, just R38 insulation packed into 4x4x7x10" boxes with polyester batting and burlap covering.

 

My hope was that this would suffice for my attack wall to reduce bass anomalies. Compared to what I had the frequency response looked like a dinosaur mouth it seems to have worked but I wanna get rid of this dip.

 

Anyway thanks alot, I'm going to review the links posted and see if I can put a graph together.

[Jimbroni]

Handy little program Ethan. As it turns out Timmy was right boundaries are roughly 27 inches from the wall. Now what? I suppose if the move the speakers away from the wall that will only shift the null upwards and visa versa. I have listening spot at the 38% rule and feel like I'm already pretty close to the near fields. I'm skeptical of pushing them closer to me.

[Ethan Winer]

Handy little program Ethan. As it turns out Timmy was right boundaries are roughly 27 inches from the wall.

 

Yes, 27 inches is the quarter wavelength distance for 125.6 Hz.

 

Now what?

 

Bass traps?

 

I suppose if the move the speakers away from the wall that will only shift the null upwards and visa versa.

 

As explained in the article that goes with my Frequency-Distance calculator, peaks and nulls are caused by two different but related issues. One is the distance between the sound source (loudspeaker) and a boundary, and the other is the distance between the listener and a boundary. Multiplied times all six of the room's boundaries. (Quite a mess, eh?)

 

There is no speaker or listener location that will give a perfectly flat response in a small room.

 

I'm skeptical of pushing them closer to me.

 

You'll never know where is flattest until you try, while measuring the LF response at each proposed location. Measuring is the only way to find the best place. Or maybe I should say the least compromised place.

 

--Ethan

[Jimbroni]

Moved the speakers out from the wall. The dip is now at 80 Hz.

[Jimbroni]

Then I moved the listening position closer to the speakers.

The dip is still at 80 but it seem to level things out a bit. Though I'm not sure about the peak at 200Hz.

 

Well at least I've learned how to use my camera better.

[Jimbroni]

The thing has been bothering me is that pink noise would normally have a slope that decreases with frequency. All my measurements seem to be relatively flat except at my peaks and nulls. Am I missing something?

[MrKnobs]

The thing has been bothering me is that pink noise would normally have a slope that decreases with frequency. All my measurements seem to be relatively flat except at my peaks and nulls. Am I missing something?

 

Pink noise is same energy per octave (or 1/3 octave), white noise is same energy per Hz.

 

You are displaying the response by 1/3 octave, so you get a flat line for perfect response.

 

Terry D.

 

P.S. My bad on Tim's post, I read his suggestion as 27' instead of 27", for some reason. Need more diopters in my reading glasses or less Grand Marnier in my orange juice.

[Jimbroni]

Thanks. Yeah I should have realized I was lookin at 1/3 octave bands.

 

Anyway here is a better image of where I started 27" from the walls So you can see why I was concerned. I've made improvements but now I'm worried I have too much of a peak in the 180 to 320 range. Which you can see a 5 dB boost in that range in my last two curves. What to do? What to do?

[MrKnobs]

 

Thanks. Yeah I should have realized I was lookin at 1/3 octave bands. Anyway here is a better image of where I started 27" from the walls So you can see why I was concerned. I've made improvements but now I'm worried I have too much of a peak in the 180 to 320 range. Which you can see a 5 dB boost in that range in my last two curves. What to do? What to do?

 

 

That's a pretty severe null you have there at 125 Hz. I have to agree with Ethan; the solution is bass trapping, and 125 Hz isn't too hard to mop up, compared to lower frequencies.

 

First off, I notice that overall the bass output of your monitors is down about 5dB not counting the null area around 125 Hz. That's probably not due to the room effects. What monitors are you using? How big is the woofer? Are you applying any EQ or room compensation either at the speaker or, if unpowered, at the amplifier? Bringing up the overall bass would make your 125 Hz null less severe.

 

If my room had your graph, I'd either put in about 4 MegaLENRDs or build a tube trap tuned to 125 Hz. A fat tube trap (or several tube traps) would suck that 125 Hz problem right out, but since tube traps have a rather sharp "Q" they might not cover that whole gap you're showing.

 

So I'd go with broadband absorption. You already have some, right? What are you using? Can you move it a little farther from the wall or corner? If you have foam, how thick is it and where is it? If you use foam wedges, it needs to be seriously thick, maybe 2ft sides like a MegaLENRD.

 

Also, forgive me if you've already covered this, but is one of your monitors maybe in a corner? Or, worse yet, a corner near the ceiling?

 

Terry D.

 

EDIT: I reviewed your original post. I see one speaker is in a corner, and, if I understand correctly, you have fiberglass boxes behind your speakers. I don't think that's a good idea. Are they just hanging or difficult to remove? If the former, remove them temporarily and run the room analysis again. See if you don't have more bass overall, especially if you have powered monitors and have a switch on the back set for back wall compensation.

[Jimbroni]

I have two corner traps 24" x 24" x 33" x 7' permanent mounting. 1 is in the back corner. 1 is in the right corner. The left corner has a sliding glass door on one wall so I couldn't install one there. Though I have some lernd corner traps coming so that corner will be taken care of as well.

 

The fiberglass boxes - wooden frames 1" x 10" 's. The boxes are 4'x6' x 10" deep. They are mobile. The construction is R38 Fiberglass insulation - Chicken wire - 2" polyester - Black burlap. Kraft paper is still on the fiberglass. I have the paper towards the wall and not the speakers.

 

Yes all these treatments are broadband.

 

The monitors are Event PS 6. Powered monitors and no back wall compensation that I know of.

 

The speakers are not really in a hard corner. My drawing wasn't perfect to scale. The Attack wall which is 3 - 8 ft walls have a roughly 175 deg angle between each of them. So corners aren't that sharp.

 

I can move the boxes out and check. I admit I'm skeptical of that. I done measurements in the room in the past without treatments and found the curve to be very sawtoothed. Though it wouldn't hurt to check again. Are you thinking the treatments are reducing the overall bass response. Question where to put more traps if I do decide to do so. I have all the corners taken care of already except 1. So Along the ceiling-wall intersection maybe? or out in space? Near or away from the speakers?

 

However, I've found that by moving the listening position closer I get more bass response.

 

My question however is do you think I'm better using the curve with the dip at 125 which is a really flat curve except the severe dip. Or am I better off with curve where I moved the speakers away from the wall, and now have a less severe dip at 80 but with a spectrum that is not bad but not as flat.