Help Me Make Sense of Sub/Mains Integration Graph!

[zilbenz]

Upon occasion of removing most of the Auralex bass blockers from listening/family/guest/kid room today leaving one in each of the top 4 corners (too hardcore of a look for a multipurpose room) and nothing else, I proceeded to rerun REW.

Before today, I was experimenting with 2 full columns of bass blockers and a 4X8 panel of Auralex sound absorber right behind my sofa (against the wall, the only possible place in this room).

I ran REW using Wayne's hard-knee house curve, obtained/applied between 3-4 filters, re-measured sub and mains together hoping for a smooth integration, and obtained the graph below.

Unsure of what to make of what I've got here, could somebody please "translate" this graph for me? I would appreciate any tips or advice as to what I've done accomplished. As usual, it does not seems that the time I've invested in reading forum postings has prepared me to answer my own questions. Sorry!

Thank you in advance.

Chris

[zilbenz]

And this was the sub only, again using Wayne's hard knee house curve. Thank you.

[brucek]

Looks OK to me. I might remove the peak at 70Hz....

brucek

[zilbenz]

brucek

I retried the measurement data this time using brucek house curve and obtained the graph below. The only undesirable result is that I must use a boost filter out of three as detailed below:

Freq Gain BW/Oct

56.37 9.0 0.111
48.53 -8.0 0.500
73.46 -9.5 0.500

I pretty much manually adjusted the filters REW recommended until I got the curve posted here.

1) Is adding a boosting filter a "no no"? Is it the wrong way to do it?

2) What other ways to reduce or avoid a room node at around 56 Hz because that caused the most trouble in manually setting the filters? I'm down to 1 desirable location for the sub.

3) You recommended removing the peak at 70 Hz. REW didn't see that as a problem or recommend a filter. Is it because of the particular house curve I'm using? I use 20 - 80 for Find Peaks and didn't get any after 60 Hz.

Thank you.

[brucek]

It looks pretty good now. I would like to see the waterfall too, using the same vertical and horizontal axis and set to LOG with ~ (x=1, y=100, z=150).

Generally gain filters are something to avoid, but you are using gain in combination with two cuts to shape an overall filter. You're not boosting the 9dB overall, so it's OK.

REW has very strict rules it follows to identify room modes and suggest filters. Many times you need to add extra filters to taste....

brucek

[zilbenz]

brucek wrote:

It looks pretty good now. I would like to see the waterfall too, using the same vertical and horizontal axis and set to LOG with ~ (x=1, y=100, z=150).

Generally gain filters are something to avoid, but you are using gain in combination with two cuts to shape an overall filter. You're not boosting the 9dB overall, so it's OK.

REW has very strict rules it follows to identify room modes and suggest filters. Many times you need to add extra filters to taste....

brucek

Thanks for the reply. Here's the waterfall graph. Please comment. Thanks again.

Chris

[brucek]

Can you set the display to LOG instead of LIN.

Set the horizontal axis from 15hz-200Hz.

Set the time span to 600ms.

Then repost the graph

brucek

[zilbenz]

Sorry for the ignorance. It's my first time posting waterfalls. Or comprehending their meanings

Hope this is the one you want to see. Thanks.

[brucek]

waterfalls. Or comprehending their meanings

Well, if you move the waterfall slice from 30 back to 1, you'll see your frequency response graph at time zero. Since you have a 600 msec time span selected, then when you move the slice to 2, you'll see what is happening 20msec later after the sweep, and so on as you move the slider. The sound persists in the time domain. Keep moving out to slice 30 at 600 msec. I've seen some resonances that are still persisting at a high SPL level after 500 msec. That's over a half second from the time you hear the REW measurement sweep.

What you want is a nice drop at all frequencies down to below the standard room noise ~(45dB-55dB) in an appropriate amount of time. You have to be careful interpreting some of the low noise tails, as they may be room noise from refrigerators, furnaces etc. But we're basically watching for large resonances ringing out in time that persist in the room. They make the sound mushy (for lack of a better word).

Overall, I think you have a pretty good response.....

brucek

[zilbenz]

Thank you, brucek. Familiar Sade DVD never sounded better. Now I know why I couldn't stand the bass lines in a couple of songs in light of the -8.0 and -9.5 reductions

But I know it's only a short time until the next ... tweaking! Long live the Shack!

[Wayne A. Pflughaupt]

I’ll agree, looks good. Waterfall especially looks great.

I retried the measurement data this time using brucek house curve and obtained the graph below. The only undesirable result is that I must use a boost filter out of three as detailed below:

Freq Gain BW/Oct

56.37 9.0 0.111
48.53 -8.0 0.500
73.46 -9.5 0.500

I pretty much manually adjusted the filters REW recommended until I got the curve posted here.

1) Is adding a boosting filter a "no no"? Is it the wrong way to do it?

What’s happening is that the 48 Hz and 73 Hz filters are overlapping each other. The boosted 56 Hz filter is only 1/10-octave, which is really narrow. You might try switching it in and out with some familiar music; there’s a good chance it’s doing nothing audible.

Also, you might want to try dialing in the hard-knee curve again...

Regards,
Wayne

[zilbenz]

Wayne A. Pflughaupt wrote:

I’ll agree, looks good. Waterfall especially looks great.

Thank you, Wayne. That was immediately obvious upon pressing play! It's gratifying to hear immediate difference for the effort.

What’s happening is that the 48 Hz and 73 Hz filters are overlapping each other. The boosted 56 Hz filter is only 1/10-octave, which is really narrow. You might try switching it in and out with some familiar music; there’s a good chance it’s doing nothing audible.

I will give it that a try but surely it wasn't a piece of cake getting those filters with 3 variables each to produce the curve as presented.

When I read your various posts about manually tweaking the filters, I didn't think much of such because I was barely able to comprehend the automatic filters and how they worked. Not any more.

The nasty dip at about 56 Hz caused the most difficulty in getting to curve to flatten out. On top of that, I'd also read brucek's comment somehere that trying add boost to lift the curve is not possible. So I just wanted to give it a shot and then post it here for learning.

Right now the room is mostly untreated which probably results in room resonance at 500 ms (see brucek's comment above).

During the whole time, I was conscientiously trying to spot and avoid "useless filters" but ended up with the boost filter at BW of only 0.111. But that was the only way I got it to work as you see here.

Also, you might want to try dialing in the hard-knee curve again...

At the risk complicating matters due to lack of knowledge and time, I simply loaded yours and brucek's house curve and played with the filters.

House curve: I did the 30 and 80 Hz sine wave test and my 24 X 15 X 8 room seemed to indicate a difference of only 4 or 5 dB. Does that sound correct? From what I've read, house curves tend to be greater than 4 dB. Am I wrong?

For that reason, I didn't want to spend the extra effort to dial in the hard knee house curve as you laid out in your sticky posts until I'm able to confirm that here.

Thank you.

Chris

[Wayne A. Pflughaupt]

House curve: I did the 30 and 80 Hz sine wave test and my 24 X 15 X 8 room seemed to indicate a difference of only 4 or 5 dB. Does that sound correct? From what I've read, house curves tend to be greater than 4 dB. Am I wrong?

If you did the makeshift test I outlined and that’s what you got, then it is what it is. Every room is different. Naturally, your ears should be the final determinate, using program material. You may find the curve needs to be steeper, or perhaps reduced even more.

Regards,
Wayne