Help with 15dB dip 24 - 40Hz

[stealth02r1]

I have been playing with REW in combination with a BFD off and on over the past year and a half and am constantly tweaking everything in my theater. With the latest release of 4.0 I have been at it again and more serious than ever. Today I actually went out and bought a ECM8000 mic to replace the digital RS meter I had been using. My room is a 23' x 15' x 8' dedicated theater and was constructed using a green glue sheetrock sandwich. I have an ACI Maestro sub and a couple of 4'x2'x4" bass traps in the front corner and 6 4'x2'x2" panels on the walls at the reflection points. My problem is that I have a NASTY 15db dip between 24 - 40Hz when sitting in any of the 4 chairs on the back riser. I am also getting 6dB of room gain at 20Hz.

My first attempt at EQing with REW and BFD got me to a fairly flat response following a 10dB house curve. That included using a fairly large BOOST around that 29Hz dip. Recently I have been attempting to get rid of that dip by upping the gain on the sub and aggresively cutting everything around it. Using this method I am able to get much better extension all of the way down to 15Hz.

In the following graph the purple line is the NON-EQ'd plot and the green line is with my latest BFD settings.



I should also mention that I have experimented with positioning of the sub and have tried every reasonable location near the right hand corner of the stage. I have not yet tried to move the sub to the rear of the room or to the left corner, nor have I attempted to elevate it. I did determine that I the sub behaved a bit better with the base plate removed and positioned on it's side. I used to have it positioned right up near the screen and facing in towards the left hand wall but just this week I have shoved it back into the corner at a 45 degree angle with the bass trap on top of it.

You can see pics of my setup here: http://picasaweb.google.com/mukilteo...ey=WnTeUAAc3Cc

I should note that the 29Hz dip is NOT present at all in the front row of seating, so with the current EQ settings there is a 5-10dB peak at 29Hz. I am not all that concerned with how the front row sounds though because I don't sit up there.

So I guess I am just looking for any suggestions on how to tame my dip. Should I keep my current settings arrived at by turning up the sub's gain and using aggressive filters to cut or should I go back to the old settings that used a single 9dB boost at 29Hz. BTW the position of the sub's gain knob is only at about 1:00 at it's current setting. I know these types of questions are asked repeatedly but...

Also if anyone has any other ideas on sub placement, or adding another sub or more bass trapping etc let me know. Thanks!

[allredp]

Hey David,
I wish I could answer your questions with any authority, but what I've gathered from haunting such posts like yours is that placement is the first line of attack, then room treatments, with electronic filters/EQ as the last and least desirable method of calibration. I'm sure some of the bass trap guys will chime in here soon for you with more experienced voices...
At any rate, I did want to compliment you on your impressive HT room! Great work and impressive equipment
Best to you in the hunt for audio perfection...
Phil

[Sonnie]

Actually acoustic panels are not always the first option if you are on a tight budget and the BFD is definitely not the last and/or least desirable method of calibration. It's hard to find and implement bass traps that will deal with sub 40Hz response, where the BFD can really be a great supplement.

Don't get me wrong, if you have the cash and the space, bass traps can be very useful, but it's definitely not always the first option.


The first thing I would ask is are you using a cal file with the ECM8000? You can use mine if you would like... it should work close enough for you.

Since you've boosted that low end... how is your sub reacting on the rather bass heavy movies like WOTW... any bottoming, strange sounds, etc.?

Have you experimented with the phase... 30, 60, 90, 120, 150, 180 degrees out of phase with measurements for each? Something to try anyway.

Is adding another sub a possibility?

Definitely a nice room...

[allredp]

Great point about the lower FQ, Sonnie! Thanks for correcting a Newbie

My over-concern about EQ'ing stems from Ed at SVS warning me against using the BFD for narrow issues in FQ. He suggests to correct only 1/3 trends or maybe 1/6 at most.

Lots to learn for me here!

Thanks again Sonnie, and good luck David (sweet looking room I must say!...)

[brucek]

I really don't know what more you can do. You're certainly on the right track in cutting instead of boosting. Boosting is a losing game.

If I look at the axial modes associated with the dimensions you gave, I see the length and width are producing a resonance at about 24Hz and 37Hz. This might say that you may not have a dip at 30Hz, but two peaks at ~24Hz and 37Hz. The best you can do then is move the listening position to get out of the peaks (as you've observed already in the front seats), but this of course is not an option. Your final curve ain't bad at all though...

Nice room.






brucek

[stealth02r1]

No I was not using a calibration file with the ECM but I will definitely try your file Sonnie.

Yes I have played with the variable phase control on the sub and have made many sweeps in REW to find the sweet spot - which currently is 0 - about right considering the sub is 1 foot behind the FR which accounts for the 1ms delay in from the BFD.

I should also mention that I have the mains set to small in the pre-pro and have experimented with XO settings from 50 to 80. I currently have it set to 60. I also have the ports plugged on my mains. Plugging the ports seemed to help blend the mains with the sub a bit better.

Bruce, my room is not exactly a rectangle - the bar area is octagonal and is 7'9" x 4'6" x 7'5". I don't know how to figure something like that into modal calculations. Also my main ceiling is actually 8'4" - is there some trick to calculating modes in an irregular shaped room like this?

Also, I have noticed that using these EQ settings with heavy cuts seems to improve the impulse response. I don't understand that...I will post up some graphs in a bit.

[brucek]

is there some trick to calculating modes in an irregular shaped room like this?

It sure isn't an exact science, so I feel if you take a look at the main three axial modes of length, width and height, it's about the best you can do. I always like to look at the first and second order to see if there is some unwanted reinforcement with those resonances. But, your room is finished, so how much does the information help.

with heavy cuts seems to improve the impulse response

Well, the boost filters meant you had to turn the input to the BFD down to accomodate the gain you were using. That reduces dynamic range and increases noise. I haven't been able to convince many people of the fact, but it doesn't make it any less true.

brucek

[stealth02r1]

Just for fun here is a TrueRTA sweep from each of my 7 seats. The red line is the main listening position. The difference from seat to seat is mind blowing...at least to me.

[Chrisbee]

brucek wrote:

Well, the boost filters meant you had to turn the input to the BFD down to accommodate the gain you were using. That reduces dynamic range and increases noise. I haven't been able to convince many people of the fact, but it doesn't make it any less true.

brucek

I wish I could agree with you. But to argue further would be 'jacking this thread.

Very posh HT! If you add another sub on the other side of the stage you may be able to tune that hump out and lift the trough by adjusting the two sub's spacing (or perhaps phase). A more central position with your existing subwoofer might be worth trying (as an experiment) just to see what happens to your response curve.

I strongly encourage you to test several more sub placement options as your next move. All it costs is your time, and like us, you do like to tinker. You have the tools needed to test.

The best location may end up being impractical for some reason, but maybe not. Find out.
Seems you've covered the front, now try the right rear corner of the room and then along the rear wall to either side of the entry door. A side wall location may even yield the best results across the most seats.

You may find your current primary seat is no longer the best choice. Keep an open mind and actually, that can work to one's advantage as the wife doesn't need to know why you are suddenly allowing her to sit in "your" seat.
Once you've zeroed in on final location for both sub and primary seat, create a filter set optimized for the primary seat. Then use the REW graph averaging feature to create more filter presets that offer the best compromise response for the rear row of seats, another for the front row and --with 10 presets to fill yet--how about another for a sold out screening.

BTW: I have a 4' wide 6' deep notch in my otherwise rectangular 20x12x8 foot dedicated room that has always had me wondering about it's effect on the room's response.

[Wayne A. Pflughaupt]

There’s no free lunch. Any outboard equalizing is going to increase S/N and reduce dynamic range. Boosting will increase noise from the equalizer (this is true of both analog and digital equalizers). Cutting will increase noise elsewhere in the signal chain. Take your pick.

The question comes down to, “Is it audible?” That’s something the end user has to determine for himself. Fortunately, most modern home electronics spec so well they have more than enough S/N and dynamic range to spare. More info here and here.

In this particular situation, we decreased the sub's S/N drastically by feeding it an extremely low signal, which required boosting its gain control substantially (a low signal/high gain scenario is a know recipe for increased noise). And, we increased output below 20 Hz, which can only exact a headroom penalty on the amp.

So, in this case, excessive cutting has cost us both S/N and headroom. To the sub amplifier anyway - I assume the BFD's S/N is intact via a maximized input signal. (Not sure what the advantage of that is if it sacrifices the downstream component. Perhaps someone can enlighten me.)

By the way, impulse response is an acoustical measurement. I doubt things like S/N or dynamic range, which are electrical measurements, have any bearing on it, but I’m sure John M. can speak to that better than I can.

Regards,
Wayne

[Wayne A. Pflughaupt]

stealth,

It’s obvious from your charts that you’re not going to be able to optimize response for all seats. As Don noted, try experimenting with placement to see of you can get a better “average.”

In the end you might well end up having to EQ for your primary seat. It probably won’t be that big of a deal for the others in the room. After all, low bass in movies is primarily all about “boom,” and you don’t need laboratory accuracy for that.

It appears that applying “drastic cuts” did improve your below-20 Hz response, which is great. But of course (as I noted above) that came and the expense of reduced headroom for your sub’s amplifier, since you essentially electrically boosted those frequencies.

But that’s not necessarily a bad thing. It appears you sub has more than enough headroom to spare anyway. So, there’s really no reason not to boost that depression at 29 Hz if you need to. After all, you don’t indicate you suffered any ill effects boosting it previously.

Truth be told, people have been boosting depressed frequencies with the BFD for years, long before the advent of this Forum, and the only time I’ve ever seen anyone report bad results was when their subs ran out of headroom and/or their drivers bottomed out (i.e., the bad results manifest in the sub, not the BFD).

Now, if your sub has built-in limiting, boosting too much might limit its output during demanding scenes – i.e., reduced dynamics. I don’t think that will be the case with your sub, though – it appears to be highly capable.

As I’ve said many times before, any equalizing exacts a headroom penalty. You have to make sure you have plenty to spare going in.

Regards,
Wayne

[stealth02r1]

WOW. So I did some tweaking today. I moved my sub to the right REAR of my room and took some measurements. I got a huge 16dB gain from 33Hz on up. I had to keep turning the gain down on the sub amp due to clipping during measurements. Here is the graph.



Notice the dip is STILL there! I took measurements from about 6 different locations with the sub in the right rear including on the back wall next to the door. I also tried facing the driver in various directions. I played with a few filters but then gave up because I was cutting up to 25dB and still couldn't get any better than my original attempts with the sub up front. Besides that have the sub within 5 feet of one ear SUCKS! I know they say the human ear cannot locate bass below 80Hz but you sure as CAN detect SOUND PRESSURE. I could totally feel the sound pressure in my right ear more so then my left. With my eyes closed I could turn my head and totally locate where the bass was located because of the equalization of the pressure on my ears. Oh FYI my XO is set to 60.

Any way the sub found it's way back up onto the stage and this time I stuck it in the left corner - which I had never done in the year and a half since I finished my theater. Now we are talking! Check this out. Here is a graph that shows the sub non eq'd in both the left and right corners. Red is the right corner, blue is the left.



So a quick attempt at the BFD and I came up with this.



Rockin so far! Gotta get back to crankin the tunes...

[Wayne A. Pflughaupt]

That’s great news! You might need to turn up the sub a bit to get your house curve back. Are you still getting all the variation from one seat to another?

Regards,
Wayne

[stealth02r1]

I have to say I became mesmerized by this new sub placement and ended up listening to 2 channel stereo music for a good 4 hours last night. I still can't believe how much different everything sounds now - just by moving the sub from one side of the room to the other. Bass doesn't just sound tighter and more defined but there is a huge difference in tactile FEEL. The bass really hits you in the chest now and a lot more energy is being transmitted through the riser and seating. Everything just seems much more cohesive - the sub feels like it is much more part of the room. Actually it is more like the other way around - everything in the room now feels like an extension of the sub.

Here is a quick sweep from all 7 seating positions. Even though there are still major swings from seat to seat I think over all the results are much more consistent now. Especially from 27Hz and below and 60Hz and above.

[stealth02r1]

I'm trying to interpret my impulse response now. Is this peak at 1.83m the ceiling? I assume it is but I would expect the ceiling to come into play much later. Also besides this one peak how does this graph look? Does it look like I have all of the primary reflection points covered - except for the ceiliing?

[DrWho]

Hey Stealth,

It is easier to read the Energy Time Curve plots than the impulse response. Make sure you plot it as %FS and normalize it. And of course zoom in on the time-scale the same distance.

Nevertheless, you can see a time-arrival difference of 1.8ms. Sound travels roughly 1 foot every 1 ms, so it's not likely your ceiling (unless the sub is about 20" from it)

I would chalk it up as a side-wall reflection or the distance between your sub and mains (is the impulse just your sub or sub+mains?)

[stealth02r1]

Mike - You nailed it. The sub is exactly 1'8" behind the mains and that impulse response was taken with both sub and mains enabled. I dialed in a little bit of phase and that fixed that. It's funny, I don't think I have ever read anything about using impulse responses to help dial in distance/phase with the sub but it makes perfect sense. Thanks.

[DrWho]

You can use the impulse response (or the ETC) to show you where reflections are happening in your room too. It works wonders in tracking down acoustical issues.

[stealth02r1]

Well I'm stumped. Changing the phase on the sub did NOT eliminate that 1.9ms peek after all. I have spent three days now trying to track down the cause of that peek but I can't find it. I don't know how accurate REW is regarding these impulse responses because I get a different FS reading at that 1.9ms peek with each sweep. All other peeks seem to measure fairly consistently except for this one. That peak is still there with the sub turned off and I have tried to rule out any vibrations - even trying to hold the mic in my hand. One thing that I did notice is that my leather chairs are the cause of most of the peeks I see. Throwing blankets over them really cuts many of the peeks. However the 1.9 peek does not seem to be caused by a reflection off of any of the chairs because blankets don't seem to be affecting it. I have also measured the impulse response of the FR and FL speakers individually and while the 1.9ms peak seems stronger when only the FR speaker is used I can't seem to kill it. I even took a bunch of impulse response measurements while holding a 2'x4'x2" acoustic panel right next to the FR speaker attempting to block any possible reflections - I even held it right in front of the speaker and STILL get that 1.9ms peak. Am I incorrect in assuming that holding a panel in front of a speaker like that SHOULD have had some affect?

And one other thing. Every once in a while my impulse response graph gets hosed and the impulse actually starts on -2ms instead of 0. It looks to me like sometimes the 1.9ms peak is actually the BIGGEST peak and the graph sets that peak to 0ms.

Somebody please help - this is killing me. Here is a typical response...

[Chrisbee]

Would you have worried about the impulse response if you couldn't see it?

Could you actually hear any difference with the leather furniture covered in blankets?

[JohnM]

Which seat are you measuring from? If it is in the back row, the reflection may be from the wall behind you. If not, try moving one of the speakers to see which surface the reflection is coming from - e.g. if you move the speaker away from the sidewall towards the midline of the room by a foot, does the peak move to a later time? If so, it is probably a sidewall reflection. If you stand the speaker on something to raise it by a foot or two and that makes the delay longer it is probably from the floor, if shorter it is probably the ceiling (but the time is so short it is more likely to be a much closer surface).

[DrWho]

stealth02r1 wrote:

Well I'm stumped. Changing the phase on the sub did NOT eliminate that 1.9ms peek after all.

Changing the phase shouldn't affect it either...

I have spent three days now trying to track down the cause of that peek but I can't find it. I don't know how accurate REW is regarding these impulse responses because I get a different FS reading at that 1.9ms peek with each sweep. All other peeks seem to measure fairly consistently except for this one. That peak is still there with the sub turned off and I have tried to rule out any vibrations - even trying to hold the mic in my hand.

REW is plotting relative time, not absolute time. It makes measuring easier because you don't have to enter the distance of the speakers from the mic to ensure the phase response makes sense. An absolute time measurement would let us know exactly when the reflection is arriving. Right now, REW just calls the first largest peak time = 0. The phase at that time will pretty much be random.

One thing that I did notice is that my leather chairs are the cause of most of the peeks I see. Throwing blankets over them really cuts many of the peeks. However the 1.9 peek does not seem to be caused by a reflection off of any of the chairs because blankets don't seem to be affecting it.

I don't doubt that you're getting reflections off your chairs - just keep in mind that the ability of a "treatment" to absorb reflections is going to be dependent on the frequency content of the reflection. In other words - claiming a peak doesn't get fixed by a blanket just means the problem frequency is under 10kHz.

I have also measured the impulse response of the FR and FL speakers individually and while the 1.9ms peak seems stronger when only the FR speaker is used I can't seem to kill it. I even took a bunch of impulse response measurements while holding a 2'x4'x2" acoustic panel right next to the FR speaker attempting to block any possible reflections - I even held it right in front of the speaker and STILL get that 1.9ms peak. Am I incorrect in assuming that holding a panel in front of a speaker like that SHOULD have had some affect?

Maybe.

The frequency content of the reflection determines which kinds of treatments might be able to remove it. Also keep in mind that some frequencies are going to bounce off your panel.

But also - it's quite possible that you're measuring a reflection inside the cabinet of your speaker. What mains are you running?

And one other thing. Every once in a while my impulse response graph gets hosed and the impulse actually starts on -2ms instead of 0. It looks to me like sometimes the 1.9ms peak is actually the BIGGEST peak and the graph sets that peak to 0ms.

As mentioned earlier - REW is just showing relative time so it's nothing to worry about.

If the peak is occurring without the subwoofer, then we can ignore the subwoofer for now. In your pics, the mains look to be near the side walls and near the rear wall...if moving your mains around doesn't affect this impulse, then we can fairly confidently claim that the "problem" is with your speakers. Are they rear ported?

[stealth02r1]

Would you have worried about the impulse response if you couldn't see it?

No.

Could you actually hear any difference with the leather furniture covered in blankets?

Nope. But I do now recline all 6 of my other chairs when I am in there listening to music. The sound is much more enveloping if the two chair backs next to me are not blocking any reflected sound.

Which seat are you measuring from?

Center seat - back row. It is 4.6 feet to the back wall.

What mains are you running?

Focal JM Labs Cobalt 826S. They are front ported but I have a foam plug stuffed in there.

[stealth02r1]

So I have spent another week of endless tweaking and have probably taken well over a hundred measurements - yeah I am kind of obsessed with this at the moment. I finally bit the bullet and decided to experiment with the positioning of the mains - which I was reluctant to do at first because it took me so bloody long to get them aimed, spiked and leveled. Anyway so I turned off the sub and enabled only the right speaker and moved it around taking measurements with it near the front wall, near the side wall nearer the center of the stage and down off the stage on the floor. The only thing that seemed to affect the 1.9ms impulse spike was moving it down off the stage. Doing that increased the delay by about a millisecond. The stage is 10 inches tall. So that led me to conclude that this spike must be the ceiling. And now I think get it. The tweeters are 4'4" from the ceiling, the microphone is placed about 4'5" feet from the ceiling and the mains are 15'3" away - so the sound only has to travel about 17' when bouncing of the ceiling which accounts for the 1.9ms delay.

Anyway I took the opportunity to reposition and realign my mains. I listened to them in a number of different locations and toe angles and ended up pretty near my original placement. Once I figured out where I liked them I actually dug out the owners manual and found that the manufacturer suggests placing them so that the center of the woofer is 22" from the side wall and 39" from the front wall. My initial placement had them about 21" from the side and 37" from the front walls. Well moving them just a few inches to the mfg recommended distances in combination with slightly LESS toe-in not only helped add width and DEPTH to the soundstage - but moving them forward just 2 more inches even measurably smoothed out the midrange response in my graphs.

Another interesting thing I discovered as I was tracking down a rattle in my main entry door was that there was a very slight gap in the weather stripping near the top of the door - just small enough to see light in the next room. Well after building up the weather stripping to fix that gap I immediately saw an almost 3dB gain in a narrow dip that I was getting right about 21Hz. I couldn't believe that such a tiny gap could be the cause of 3dB so I ran a few more sweeps and recorded the same. Then I went over and cracked the door open just a hair. Bingo! The dip was back!

So of course I then went around the room looking for and plugging all gaps I could find. Even though I built the room with double drywall with green glue and used acoustical sealant to seal gaps, I still had 4 1.5 inch penetrations for cabling that I never plugged because I didn't use wall jacks. I also had some big holes at the toe kicks leading to my bar cabinets and a 1 inch gap between the fridge and the counter top. After plugging all of those holes I saw slight but positive responses in my graphs. I'm thinking that the cabinets were actually functioning as a Helmholtz resonator and I could probably actually use them as such if only I could figure out how to tune them to a specific frequency.

The other thing I learned is that the positioning and location of my bass traps make a substantial difference - to the measurements at least. I used to have my 2'x4'x4" GIK traps at a 45 angle in my front corners. Well I actually experimented with placement and learned that when they were both at 45, they were absorbing too much between 100Hz and 250Hz I think it was. I actually ended up locating the left one on top of my sub but flat against the front wall. Any angle and it was working too well. The right one seemed to work best angled in the corner but with an angle less than 45 and more against the front wall.

Oh yeah. One other interesting thing to note. I have a Bluegears b-Inspirer sound card that features DTS and Dolby Digital real-time encoding and digital outputs. I have the sound card hooked up to my pre-pro via optical and analog and I can run REW sweeps and take measurements using either the analog or digital outputs. When running sweeps via optical I get a FULL 3dB increase from about 100Hz on down below 20Hz. I am guessing that the DAC in my pre-pro gets the credit for the extra gain here. All of my sources sure sound better using digital connections and therefore this DAC.

I am pretty happy with the results so far. I feel I have now optimized the location for the sub, mains, bass traps, 60Hz XO, and phase and am getting decent enough measurements with only two filters on the BFD. I know if I move the XO up to 80 I would be able to cut the peeks at 61Hz and 73Hz but every time I play with the XO I end up liking it set to 60 the best.

The latest graph.



Next I think I am going to build some panels to kill the ceiling echo.