Sub measurements with graph....suggestions? (1 Viewer)

[Sonnie Parker]

Crossover is 80hz. Results are for sub only. Mains off.
Values are compensated with corrections. (RS SPL Meter)
Sub is shoved up in the front corner of a 4300 cubic feet room with a large 5 foot X 9 foot opening into another room with approximately 6000 cubic feet. (I should probably install some french doors in this opening but it wouldn't pass the WAF.)
I have already set up a couple of presets on my BFD to smooth out this response but am always open to other suggestions that I might input into another preset. Then I can switch between the presets and determine what sounds best to me. I'd like to end up with 3-4 presets total.
I believe I would favor a house curve rather than flat. I believe this will help me when I get to the 50hz dip as well.
I have had some strange experiences with this 50hz dip. I tried boosting it up about 8-10db and got rid of a good bit of it. But, when I was running my test tones with the boost, that 50hz tone sounded really strange, as if the amp was struggling. The tone did not sound smooth like the other tones. It sounded too loud althought the reading only increased 8-10db. I suppose the headroom was sucked up, even though I was playing only the one tone. This really doesn't make sense to me. I might would understand it if I were playing several octaves of tones but this is just 1/6 octave playing. I tried +6db and got the same. When listening I noticed the same thing, so I notched the boost back down to +3db and it sounded much better. It didn't sound like I was missing anything, but I still have a huge dip at that range. It actually made me wonder if I was getting the right reading at 50hz. But I've always measured this dip in that range, even when I ran the sweep on Spectra Plus.
At any rate, I'm open for suggestions and am curious as to how others might set up their BFD if they had this same response and set up (obviously without the benefit of a listen).

 

[Wayne A. Pflughaupt]

Sonnie,
Sorry to take so long to get to this – those other threads are keeping me pretty busy (I know you know which ones!).

I have had some strange experiences with this 50hz dip. I tried boosting it up about 8-10db and got rid of a good bit of it. But, when I was running my test tones with the boost, that 50hz tone sounded really strange, as if the amp was struggling. The tone did not sound smooth like the other tones. It sounded too loud althought the reading only increased 8-10db. I tried +6db and got the same. When listening I noticed the same thing, so I notched the boost back down to +3db and it sounded much better.

You may be stressing the amp’s headroom, but not necessarily. The change from +10dB to +6dB should have made a difference. Did you try just reducing the sub’s volume?
I’ve noticed that sine waves can be strange (assuming that is what you are using). For instance, there are a few in my room that sound shockingly loud, but measure normal – i.e., in line with everything else. I never had that with filtered pink noise tones.
It’s getting late; I’ll dig in to your readings tomorrow and let you know something then.
Regards,
Wayne A. Pflughaupt

 

[Sonnie Parker]

Thanks Wayne, no big hurry, I plan on setting up a couple of filters on Wednesday (my day off).
You have been pretty busy lately. And helpful as usual.
I suppose I should mention that I have one wall full of windows and french doors to the outside. Probably one of the worst setups for home theater you could be faced with. It sounds respectable though. I'm mainly a movie buff so it makes it a little easier....I don't have to be so precise for music listening.
If you have time, you can get an idea of what my room looks like here:
Pictures
But warning, I haven't mastered imaging on the web too well yet, nor my digital camera. I plan on having some better photos up soon. You can still get a good idea of what my room looks like though. The htpic4 will show you the large opening and htpic5 will show you part of the windows. In htpic6 you see another opening which I included in as part of the main room.
I did reduce the subs volume but then it didn't sound loud enough on movies. The drop to +6db did have a small effect but the drop to +3db sounded better, on the test tone.
I really probably wasn't fair on stating that when listening it sounded better because I only watched part of a movie. The tone sounded better but maybe I should watch a couple of movies to make absolutely sure.
Any maybe I should try out some filtered pink noise tones.

 

[Patrick Sun]

Room suckouts are pretty tough to fix with simply boosting those frequencies because you are fighting audio physics, and that's why the goosing that trouble spot frequency range taxes the amp because "it's giving it all she's got, captain!" but you just aren't getting the acoustical energy for the input power applied at your listening position.

If you tried the sub in different locations (if possible), then you might fix the suckout problem, and may create some humps that are easier to fix.

 

[gomez_a]

Sonnie,

i'm trying to post my chart as well, but everything I try fails,I posted the link to my free web host site (angelfire), but the site gives a message that it doesn't allow remote linking. The site works because when I click on the link outside this forum, all the data clearly comes up.

I tried to cut and paste into this windo and that doesnt work either - anyhelp would be appreciated

 

[Wayne A. Pflughaupt]

Sonnie,

I’m with Patrick: This is going to be a tough one to smooth out. Is there another corner available – preferable one with maximum uninterrupted wall length in both directions?

If not, we’ll just do the best we can.

Regards,

Wayne A. Pflughaupt

 

[Sonnie Parker]

Wayne ask:

I’m with Patrick: This is going to be a tough one to smooth out. Is there another corner available – preferable one with maximum uninterrupted wall length in both directions?

Good to finally run across you here Patrick.

Fellows, I'm afraid I'm stuck with the corner I'm in now. All my walls are pretty much interrupted. I will say that I don't believe the dip has bothered me that much so far. I don't realize it if I'm missing anything although I suppose I am. I still want to get a hold of some filtered pink noise and see what it sounds like on the boost.

Do you think I might reduce my crossover point to 50 or 60hz and see if this might help some. I'm just not so sure my PSB 6T's can handle it that well, of course then again, I don't listen at really loud levels to begin with. Moderate, but not as loud as most probably.

I figure if I go with a house curve and start the roll off at 28-32hz that might cause the dip to not be so bad maybe.

I will be all over this tomorrow. No chance of me doing anything outside since the weather is going to be very chilly for here. It will be an all day insider.

 

[Sonnie Parker]

gomez....
E-mail me your chart... are you using the excel workbook?
If so, e-mail me your entire sub test file and I'll put it up for you here and the fellows can look at it. I'll even stick it up on a webpage for ya so you can do whatever you wish with it.
Send it to [email protected] don't use the e-mail here because I'm about to head home....the e-mail link here is to work and I won't be back here til Thursday.

 

[Wayne A. Pflughaupt]

Okay, Sonnie, I think the following adjustments will give you much (read audibly) improved results.
Starting at the upper end, set a filter for 1/3 octave bandwidth, centered on 71Hz, and cut 6dB.
The outer parameter of the filter will be 63Hz and 80Hz. However, it is common for a filter to “spill over”, or affect an area a little wider than it’s “supposed” to. I’m counting on this effect to make needed adjustments to 63Hz and 80Hz. If you find the BFD’s outer parameters are rigid (i.e., no change at 63Hz and 80Hz), then tweak the filter a little wider – in addition to the –6dB at 71Hz, we need a 2dB cut at 63Hz and 80Hz.
Next let’s tackle the hump at lower end. I think a filter set for 1/2-octave, centered at 34Hz and cut 4dB will do nicely there.
That leaves us with that pesky situation at 50Hz. As Patrick noted, room nulls can be a problem. However, room-induced nulls are usually pretty narrow, typically 1/6-octave or so. Your problem is fully an octave wide, Sonnie. In my opinion this is too wide for null. A null may be aggravating it, but I think much of the problem is the response characteristics of the sub.
I’ve already made it know what I think of equalizers and headroom. For those who missed it, you might want to check my third post at this thread(
http://www.hometheaterforum.com/htfo...threadid=52092), but this excerpt pretty much summarizes:

But irrespective of that, I personally think the headroom thing is overblown. While it is true that EQing can cause the amp to run of headroom at any frequency where too much boost is applied, all that means is distortion. However, distortion is largely inaudible at low frequencies. With a few notable exceptions, it is not unusual for the sub driver itself to introduce distortion at levels of 10-15%, even when given a clean signal. So, any clipping from the amp generally will not usually add to the distortion that is already there.

That said, Sonnie, I think you might be the exception, for two reasons:

• The null is extremely deep – about 20dB.
• Considering your tremendously large area (two huge rooms that open to each other, totaling 10,300 cubic ft.) I think you are seriously “under-woofed.”

Thus I imagine you are already pushing your sub/amp combo to the max, so you might indeed have a headroom issue.
In light of that, I make the following recommendation with some trepidation:
Set a filter for 2/3-octave bandwidth, centered on 45Hz, and boost 10dB. Notice I shifted the filter off the 50Hz center. I’m basically writing that off as a lost cause (even though this filter should boost the center about 5-6dB). However, this filter, coupled with the other two, will reduce the problem from 1-octave wide to 1/2-octave – a significant improvement.
Again, the caution: If you find problems using this 3rd filter, you will have to reduce the boost as much as needed. Or get a more powerful amp.
If you can pull off all three filters, Sonnie, you will have a built-in house curve of 8dB from 80Hz to 32Hz. You might find this ideal for you room, given its size. If it seems to be too severe, a couple of 1/6-octave filters at 32Hz and 22Hz cut 2dB will reduce the slope to 6dB.
Good luck, and let us know how it goes.
Of course Sonnie, if you got better results with your own settings, I guess should consider retiring and turn this gig over to you .
Regards,
Wayne A. Pflughaupt

 

[Sonnie Parker]

This should be interesting Wayne. I wish my wife and daughter were away for awhile tonight and I'd go set it up and take a listen.
Under-woofed.......hmmmmm! You mean I might need 2 of those expensive HGS-15's?:frowning: Ouch! Or 3? Triple ouch! You would think at 1200 watts nominal and 3000 watts peak I'd have enough, but then again, my open area is quite large. Heck, I haven't turned the volume on the sub past 3 dots yet.
I did read your post that you summarized above on the other thread. As a result of your commments about this I have started re-wording my BFD page somewhat. I believe instead of advising someone not to boost frequencies because they may be sacrificing headroom, I would rather explain to them what you stated. Then they can listen and hear for themselves if they like it. I think they can notice if they are pushing their amp too hard, or at least I hope they can.
Look, if you retire on the account of me, I'm gonna send you some get well pills because you will have to be sick. lol No way I can handle this gig. That's a funny one Wayne.
Now, it's quite possible I may like the sound from my own settings but that would be because of what I'm accustomed to hearing or because I don't know any better.
Believe it or not, I haven't listened to many home subs in my lifetime. There are none to be found in these parts. I can't say that I have ever heard a properly setup sub in a home, so I don't really know what it is suppose to sound like. As my profile says...I live in the boondocks, always have....I am truly a country hillbilly redneck. I wear overalls and dip snuff. It's all very true. I look out my window and see cows. There are no dealers with any kind of respectable home theater stuff even remotely close to me (not that it would matter because it's not a home environment). The folks living in the counties I live and work in would pure laugh if I suggested they purchase a $2000 sub. Most folks around here go to Wal-Mart and purchase the all-in-one systems and are completely happy. They laugh at me all the way to the bank. Yeah, my friends love my HT setup and believe it sounds great, wish they had one, but ain't gonna spend the money.
I'll let ya know how it turns out.
Thanks for taking the time to work on this. You are most kind.

 

[Wayne A. Pflughaupt]

Sonnie,

The folks living in the counties I live and work in would pure laugh if I suggested they purchase a $2000 sub. Most folks around here go to Wal-Mart and purchase the all-in-one systems and are completely happy. They laugh at me all the way to the bank. Yeah, my friends love my HT setup and believe it sounds great, wish they had one, but ain't gonna spend the money.

Believe it or not, there is no shortage of those types here in the big city.

Regards,

Wayne A. Pflughaupt

 

[Sonnie Parker]

Okay, Here's what I did:
I got everything (environment,etc.) setup as I did when I made my measurements for the first graph in this thread. I realize now I had forgot to shut my bedroom door previously so I left it open. I was wondering why that dip was about 3 db lower than what it had been once before. And two of the peaks were a few db less than previously. Anyway I ran a pre-filtered measurement and actually came up with the same response. I was expecting to be off a db or so here and there, but not so.
I input the 3 filters you suggested and took more measurements and this is what I came up with:

There doesn't appear to be much difference in the response lines here. I thought there would be. Reckon why this is?
Then, I shut the bedroom door and made an input level adjustment (increase from +3 to +5 on my pre/pro) and turned the volume on the sub down a tad. I felt my level had been a little low.
I took raw pre-filtered measurements again and got a little different response. The peaks rose a db or two and the dip rose as well by about 3db. Amazing what one shut door will do. Then I made some adjustments (8 filters) and came up with this:

I watched about the first 5-10 minutes of Swordfish (throught the first explosion of the girl and the 2 swat team fellows that get blown to pieces...ooh) as follows:
BFD off...
BFD set to preset 6....
BFD set to preset 7....
I really couldn't tell much difference between off and preset 6, but I could tell a slight difference in preset 7. It was a tad softer, the explosion didn't hit quite as hard (I'm sure attributed to the 31.5hz peak being cut). But they all sounded good. I should probably watch a few movies in each setting, but I'll probably forget from one to the next what it sounded like.
There has to be a rhyme and reason to this.

 

[Wayne A. Pflughaupt]

I didn’t use the graph - I find 5dB readings too coarse. I put the corrected readings on graph paper and adjusted the relevant, specific readings for the effect a filter would have. Makes it tough, though, when the parameters change (open vs. closed doors ), etc.
I assume the magenta line is the EQ’d response? How does a 4dB cut at 34Hz end up as a 2dB boost...? Similar with the 71Hz filter – a 6dB cut shows up as a mere 1-2dB reduction. What’s up with that?
At least the 45Hz filter worked as it was supposed to, closing much of the 1-octave valley.
Nevertheless, it looks like you’ve done a great job on your own, congratulations! I’m ready to give you the gig.

Watched about the first 5-10 minutes of Swordfish. I really couldn't tell much difference between off and preset 6, but I could tell a slight difference in preset 7. It was a tad softer, the explosion didn't hit quite as hard (I'm sure attributed to the 31.5hz peak being cut).

It is better to use music. Movies really aren’t the best way to judge subtle EQ changes, since explosions are merely random, broad-band low frequency energy. Music has individual bass notes, which will better reveal EQ changes. If it sounds good with music, it will sound good with movies – the reverse is not necessarily true. The only caveat is that music often requires lower sub volume.
Regards,
Wayne A. Pflughaupt

 

[Sonnie Parker]

What amplitude readings would you suggest? I felt 5db increments causes the graph to look more proportional and not so crowed.

I didn't do you wrong on the parameters, I kept them the same on that graph. Yes, I did change them for my graph but that would not effect the first one.

Wayne is puzzled like me and ask:

How does a 4dB cut at 34Hz end up as a 2dB boost...? Similar with the 71Hz filter – a 6dB cut shows up as a mere 1-2dB reduction. What’s up with that?

You got me. I wondered the same thing and the only logical answer I could come up with was the 2/3 octave 10db boost at 45hz overran the cuts.

I wonder Wayne if we could see on a graph what a 2/3 octave 10db boost at 45hz would look like if we started with a flat line? If we started at a flat line of 80db we know that at the center (45hz) the line would be at 90db, right? What would the line be at 40hz and 50hz, then 36hz and 56hz, and so on back down to the where the boost would level off to 80db again. Can you determine this? I know you can.

Look, there's no way I can eyeball someone's graph or numbers and just figure it out. It took me 14 measurements to get mine right. I use much narrower bandwidths because everytime I've tried the wider ones I get screwy results.

Do you think it might be my test tones? I used the NCH generator and burnt a CD. Is there a filtered pink noise generator?

 

[Wayne A. Pflughaupt]

Sonnie,

Look, there's no way I can eyeball someone's graph or numbers and just figure it out. It took me 14 measurements to get mine right. I use much narrower bandwidths because everytime I've tried the wider ones I get screwy results.

Wide filters are effective for wide problems, if you can determine how wide to set them. Of course, with all the filters you have to watch out for the “overspill.” Sometimes it can help you, sometimes you absolutely don’t want it.
With my recommendations you may have noticed I only addressed the worse problems, not fine-tuned it to the nth degree. Addressing the worse problems gives immediate and audible improvements. Tweaking everything to the last dB of perfection “looks” great on the graphs, but you have to have a pretty sensitive ear to hear those subtle changes.
Regards,
Wayne A. Pflughaupt

 

[Sonnie Parker]

I appreciate you taking the time to work on this with me. I believe this is really helping me understand this stuff much better. After your last post I did go back and double check to make sure I didn't accidentally input something wrong and it looked right. Of all things I hope I know how to input the settings for the filters...lol....but anyone can make an input mistake.

Tweaking everything to the last dB of perfection “looks” great on the graphs, but you have to have a pretty sensitive ear to hear those subtle changes.

I tend to agree with you here. + or - 3db shouldn't have a significant impact should it? Doesn't it take about 10db for most of us to notice a change? Or did I dream that up?

I see now how 2db increments can be a little better to be precise with. Didn't seem to effect the smoothing though.
It is my understanding though that bandwidth is actually measured from where the curve crosses + or - 3db (depending on if it's a boost or cut) and not from the flat line. I can't say it in words so here's an image (isn't a picture a 1000 words sometimes...lol :

 

[Wayne A. Pflughaupt]

Sonnie,

I see now how 2db increments can be a little better to be precise with. Didn't seem to effect the smoothing though.

Notice the individual “boxes” of the 5dB chart are nearly square, while on the 2dB chart they are “squashed” rectangles. Make the latter like the former, and you will see a significant change is the curve’s appearance.
Wow, the gain numbers at 36Hz, 40Hz, etc. are just as I predicted! Or did you do that on purpose to make me look good?
Sonnie, I owe you and a lot of other people here an apology. It seems I have been laboring for some time under a false presumption. After contemplating the discrepancies between my recommendations and your real-world results, I dug out a manual for an equalizer that shows response graphs of 1/3-octave filters. I was surprised and chagrined at what I saw.
I’ve been presuming that a filter’s bandwidth is as wide as it “named” - that is, a 1/3-octave filter is a total of 1/3-octave wide, etc. I mean, it makes sense, doesn’t it?
But it seems this is not the case! Judging from said graph, a 1/3-octave filter extends 1/3 octave in each direction from center! So, a 1/3 octave actually affects an area 2/3-octave wide.
Now it starts to make sense: Remember I noted that from your chart that it looked like the 45Hz, 2/3-octave filter was affecting an area 1 1/2-octave wide? That was pretty much right! The filter was functioning 2/3 octave in each direction from center - a total of 1 1/3-octave affected bandwidth.
Problem is, I’ve been making recommendations to a lot of people here, giving them erroneous information on how to set their equalizers. :b :b :b
So now that I know what I’m doing, go back and try my recommendations again, but cut the bandwidth in half.
Regards,
Wayne A. Pflughaupt
P.S. I can send you a j-peg of the graph if you like.

 

[Sonnie Parker]

Here's what the 10db boost would look like with the squares fairly square:

Using the bandwidth example chart in my last post, and a Q factor vs. Octave chart, it appears this peak on this graph has a Q factor of 2.65 which translates to approximately 35/60. So you are not off much at all on the 2/3 octave (40/60). That's using the +3db points of the peak. If the peak stretchs out to 80/60 (1 1/3 octave) then the beginning and ending points at flat should be very close....wouldn't that be 30hz and 60hz?

Now it starts to make sense: Remember I noted that from your chart that it looked like the 45Hz, 2/3-octave filter was affecting an area 1 1/2-octave wide? That was pretty much right! The filter was functioning 2/3 octave in each direction from center - a total of 1 1/3-octave affected bandwidth.

So this explains the overspill we were getting and we are not crazy after all. lol
You know my intentions were not to make you look bad or wrong because I didn't know this myself until you just explained it. I really haven't fully understood that bandwidth example chart. I guess that's why I always used narrower bandwidths though. I would go to figuring with the calculations and come up with a wider bandwidth as my setting. I'd plug it into the BFD settings and it wouldn't do right. I found myself taking a 1/3 octave bandwidth down to 1/6 octave or even 1/8 or 1/10 sometimes to make it do right. I was always looking at the point of the peak or dip as being on the flat line instead of at + or - 3db. Now I know why I tossed the calculations out the door everytime.
Of course then again, not many peaks and dips are that perfect bell curve so it still makes using the calculations complicated at times.
Please let me know if I'm wrong in my thinking here. I don't want to go about thinking I've got this halfway figured out and only have it about 10% figured out.