method to identify room modes

[huguetguy]

Hi all...been reading & researching for a year & a half now, and just recently finished the basement so I've started calibrating my subs...i'm in the process of trying to identify my room modes so that I can place the subwoofers in best locations for minimal eq necessary with the BFD and to eliminate nulls at the seating positions...

this is my room layout (home theater/media room in lower left, or south west corner):
basement layout.gif

The media room is closed off from the rest of the basement, with framed drywall walls all around. the south and west walls have concrete block behind the framed walls, the walls are all filled with insulation. concrete floor with carpet.

I've run the spreadsheets to identify what room modes I theoretically should encounter, and had some ideas to combat those modes with 2 svs 25-31 pci's...but what I'm finding is that some modes are showing up, and some aren't, and some are coming out of nowhere...

These are the predicted room axial modes:
room modes predicted.gif

And this is the process I've followed so far, please correct me if I'm misunderstanding something...
1. Put 1 subwoofer in southeast corner, put spl meter close to floor in north east corner (near 2 closed doors)
2. Ran 15-200 hz sweep with REW

I see easily recognizable 19, 38, 57, and 77 hz modes/peaks for sure…but the puzzling thing is that there’s a 31-32 hz NULL in the corner…and it seems to remain in other measurements also when I move the mike out to seating positions too…

I will try to post REW graphs soon, but for now wanted to see if I’m doing something wrong, or if I’m misguided…how can you get a null in a corner? That’s my main question right now…

[atledreier]

Images don't show for me...

[jerome]

I don't see the images you posted.

It's important to remember that you may have to move your subs from the corners if you want to get the best frequency response. Corners increase bass but may nor fit your room/setup.
Did you try to change the phase on your subs in order to correct the NULL areas ?

Also, you should only be interested in measuring around your sweet spot. Will you seat in the north east corner ?

Another trick may be to test with open/closed doors. You may be surprised to see how much difference a door can make ......

[huguetguy]

jerome wrote:

I don't see the images you posted.

It's important to remember that you may have to move your subs from the corners if you want to get the best frequency response. Corners increase bass but may nor fit your room/setup.
Did you try to change the phase on your subs in order to correct the NULL areas ?

Also, you should only be interested in measuring around your sweet spot. Will you seat in the north east corner ?

Another trick may be to test with open/closed doors. You may be surprised to see how much difference a door can make ......

I realize all that...I don't plan on leaving them in the corners...it's just to start, I put one in the corner and ran it by itself to find out what the room modes are, and am running into this weird null...

if I can figure out the room modes, I can calculate locations that will neutralize the modes/nulls, and then put the mic in the seating areas and tweak from there...

just wanted to see if anyone knows how you can get nulls in a corner...

[huguetguy]

Here is my graph of the corners measurement...
corners.jpg
I ran this with REW & the sub calibrated to 80 hz baseline, using the REW pink noise output.

to me, it sure looks like the axial length 19, 38 & 57 hz modes are all showing, along with the axial width 44 hz mode. But what are the sharp valleys in there - they sure look like nulls to me, around 56, 77 (shouldn't that be a peak instead, from the axial length modes?!).

Is it possible to get nulls in a corner? Do you think that's what those are, those sharp valleys, or are those just what it's supposed to be at, and the modes are so pronounced that the parts in between appear to be valleys...

Based on these mode measurements, I had wanted to put the subs at the 1/3 & 2/3 positions along the length wall, based on a suggestion from Ed Mullen, but when I did that and mic'ed the corner again, it didn't really eliminate the 19 hz bulge, and I still got a 31-32hz null, coming out of nowhere...will have to redo that measurement and post the graph to show you.

[atledreier]

you still have floor and ceiling nulls, at the very least.

[huguetguy]

but would I get those with the mic at the floor, in the corner? I mean, I know there's a 1st harmonic null halfway up in the air, and then a 2nd harmonic at the 1/3 & 2/3 heights...but aren't the strong modes located always in the corners, closest to the room boundaries?

[clubfoot]

You have to remember that the predicted modes are for an ideal rectangular room, from your sketch your room isn't square in one corner! Looks like a corner fireplace?!
Forget about the modes for now, just place your mic in you seating location and take a measurement of your sub only and see if it tracks the target curve.

[huguetguy]

yup, it's a corner fireplace. how does that affect modes?

i know, i know, i could skip all this and go straight to testing my 6 seating positions...but the problem with that is that I wouldn't have solidified my understand of room modes and how sub placement affects them. I am trying to confirm whether placing a mic in the corner will work to show what room modes are excited for each sub location that I test. I guess I'd like someone who understands the scientifc theory to confirm or deny that for me...

I'll get to the seating locations in a bit, but first I want to understand the room interactions at their most basic level before getting into seating locations (because the couch can be adjusted, and the subs can be moved depending on what I find).

[clubfoot]

The accuracy of the predicted room modes is based on rectangular room dimensions, once the room deviates from a rectangular shape out goes the accuracy! It takes allot more math for oddly shaped rooms, which is why I suggested going straight for the seating location

[Wayne A. Pflughaupt]

Here is my graph of the corners measurement...

Is that with both subs in the two corners? That’s some pretty weird-looking response. Almost looks like some kind of comb filter happening. Have you tried taking a reading of each separately. It would help for comparison’s sake. As brucek eloquently put it in a recent thread, the only real reason to separate subs would be to achieve a combined response that you can’t get from each separately.

Regards,
Wayne

[huguetguy]

that was with 1 sub in the southeast corner, and the mic in the north east corner. I have lots of graphs now, I will post something later maybe wednesday after christmas...it's quite frustrating trying to achieve flat response over 8 seating positions...there's so many possible sub locations & phase combinations that it's mind boggling...just trial & error at this point...

more to come later...

[jerome]

huguetguy wrote:

it's quite frustrating trying to achieve flat response over 8 seating positions...there's so many possible sub locations & phase combinations that it's mind boggling...just trial & error at this point...

As I have experienced, it's not an easy task to achieve a satisfying response curve over 8 seats. Usually, you will have to decide what is the most important seating area (typically your sweetspot) and try to compromise for the other seats.

As for the best way to achieve a good calibration, the try & error method works. For a given sub position, you will quickly find what is the best phase/delay adjustment using real-time measurements in REW.
In my opinion, it's a very difficult task (if not impossible) to find the best theoretical location and phase in most rooms. You always have furniture, windows, doors, a non-rectangular room, ... that will make the theory fade away. I know how it feel sometimes: Your current curve is not looking good at all but you should only focus on the seating positions, not the corners.

Good luck !

[huguetguy]

Well I have found 2 locations that seem to give pretty good response over the 8 seats. I'm putting one sub along the bottom wall, in the middle. The other sub is going along the east wall, about 38" off the corner.

I have measured, and remeasured...I loved the response I was getting with both subs by themselves, but when I added in the main speakers, it threw things off a little...so I adjusted them slightly...

This is what I'm getting now, pre BFD. The charts will start at the leftmost seat on the couch, and go right. 8 seats total...

Left 4th seat out:
left 4.jpg

left 3rd seat out:
left 3.jpg

left 2nd seat out:
left 2.jpg

left middle seat:
left middle.jpg

right middle seat:
right middle.jpg

right 2nd seat out:
right 2.jpg

right 3rd seat out:
right 3.jpg

right 4th seat out:
right 4.jpg

Can you guess which seat I'll be sitting in?

Next I'll guess I'll measure each sub separately at all seating locations, try my hand at some filters on each sub, and then remeasure together.

Oh, another thing. The eastmost sub is set at 130 degrees phase while the one in the middle is set at 0 degrees phase. the couch is pulled 13" off the wall.

[huguetguy]

here's a graph showing specific null frequencies I found when I had just 1 sub in the corner, measuring along the length of the room (the longest dimension)...they occur at specific locations that just happened to be in bad seating locations.

length nulls.jpg

here are the height nulls I found, when the sub was on the floor, and the mic was at 51, 57, and 58" from the ceiling.

height nulls.jpg

the width nulls didn't really line up with the theoretical predicted nulls at locations expected, these occured at 96 & 99" from the south wall, along the east wall. the sub was at the southeast corner.

width nulls.jpg


I tried putting the subs at the 1/3 & 2/3 locations along the length at first, it didn't seem to help much. Then I tried a sub in the corner and one in the middle of the length to combat the huge 19 hz hump. Then I tried one in the middle of the east wall, and the other middle of the bottom...didn't like that.

I might also try 1 sub at 96/99" along the east wall, and the other in the middle of the length to see if that gives even flatter response...I had to pull the east sub away from the corner, and the couch away from the wall or there was a major hump around 44 hz...

It's really hard to explain all the different combinations that I've measured, but I thought I'd try to explain a little in case others were wondering...