Using REW Waterfall to calibrate sub instead of FR graph

**oliverlim** · **04-25-06, 12:25 AM**

I was wondering if anyone actually tried to calibrate their sub using he waterfall decay instead of FR graph? What I meant was that looking at the waterfall, there are usually certain freq where the decay is much longer then the rest. And subjectively these decay seems to hang on in your room. So what can be done is find out the exact freq of the long decay and do a EQ setting to reduce it and remeasure till the decay at the level is close to the rest of your freq.

Oliver

**Sonnie** · **04-25-06, 10:23 AM**

Hey Oliver.... you slipped in here on us man. Good to see ya again. Welcome to the Shack!

Whew... waterfall decay is way over my head. Hillbilly's can't go there. Hopefully JohnM will be around soon... or maybe one of other tech guys can help ya with this one.

**oliverlim** · **04-25-06, 12:31 PM**

Hee. Good you see this up and it looks great! I just need more traps in my room as 5 units of 2x4x4" bass traps hardly made a dent in my low bass decay. The room just seems to store and slowly release the bass notes. Thats why I was thinking of trying another way of EQing while I am waiting for quotes on redoing my room.

Oliver

**JohnM** · **04-26-06, 09:55 PM**

Quote:

oliverlim wrote:

I was wondering if anyone actually tried to calibrate their sub using he waterfall decay instead of FR graph? What I meant was that looking at the waterfall, there are usually certain freq where the decay is much longer then the rest. And subjectively these decay seems to hang on in your room. So what can be done is find out the exact freq of the long decay and do a EQ setting to reduce it and remeasure till the decay at the level is close to the rest of your freq.

Ideally you should use both. Generally the peaks in the FR curve will have correspondingly long decay times in the waterfall, so fixing one fixes the other. However, the waterfall (or the spectral decay) curve can be better for identifying cases where what looks like one peak in the FR curve is actually due to two closely spaced modes, which become visible as separate peaks in the later slices of the waterfall or spectral decay.

**oliverlim** · **04-27-06, 05:05 AM**

I think I understand what you mean. Suppose for my case that means that I should be attaching with EQ the 4 long delay waterfall between 27hz to 40hz as shown here

oliver waterfall 500ms.jpg

instead of the 2 FR response from 37hz and 45hz shown here?

oliver fr.jpg

funny that for my case the waterfall problem areas and the Fr seems so different right? Or is this normal?

Oliver

**Sonnie** · **04-27-06, 09:55 AM**

Man that looks neat as grits... just wish I could understand it.

**bricor** · **04-27-06, 01:54 PM**

All of a sudden, I don't think my waterfall plot looks so bad...

**oliverlim** · **04-27-06, 07:40 PM**

Quote:

bricor wrote:

All of a sudden, I don't think my waterfall plot looks so bad...

****.....

Good thing about concrete rooms is that it pressurise the room easily and you can get a tactile feel. But the room boom can really irritate you without any eqing.... Sigh. I would kill to get a waterfall and FR like yours.....

Oliver

**oliverlim** · **04-27-06, 07:45 PM**

Quote:

Sonnie wrote:

Man that looks neat as grits... just wish I could understand it.

And I thought you are suppose to teach me

You know, not that I am picky, but that word just below your name on the left just does not seem right

Oliver

**jmprader** · **04-27-06, 08:16 PM**

Quote:

bricor wrote:

All of a sudden, I don't think my waterfall plot looks so bad...

Is this with/without eq and/or bass traps?

How big is that room, what's the construction and what were your settings when you ran this?

Inquiring minds want to know.

**Sonnie** · **04-27-06, 11:14 PM**

Hey John... welcome to the Shack! Good to see ya.

Quote:

oliverlim wrote:

And I thought you are suppose to teach me

You know, not that I am picky, but that word just below your name on the left just does not seem right

Hey... I'm here to learn too! I can teach you how to learn... just follow my lead... lol.

And what's wrong with my title... JohnM has that same title. I suppose you think it should Shack Hillbilly? I can change it. Give me some ideas and don't be rude.

I know what you are thinking...

**oliverlim** · **04-27-06, 11:54 PM**

Quote:

Sonnie wrote:

Hey... I'm here to learn too! I can teach you how to learn... just follow my lead... lol.

And what's wrong with my title... JohnM has that same title. I suppose you think it should Shack Hillbilly? I can change it. Give me some ideas and don't be rude.

I know what you are thinking...

What am I thinking?

And getting back on track, yes bricor is that waterfall with EQ already and after bass traps? Mine is with bass traps and after position optimizing but no EQ yet.

Oliver

**jmprader** · **04-28-06, 12:49 PM**

Quote:

oliverlim wrote:

...yes bricor is that waterfall with EQ already and after bass traps? Mine is with bass traps and after position optimizing but no EQ yet.

Oliver

Bricor:

And are we close-miked or seating position? What's the gate on your plot, default or something else (hey, I am a few steps past my technical limits here, so be gentle if the question is phrased improperly)?

**bricor** · **04-28-06, 09:58 PM**

Hey Guys,
Sorry for not replying sooner, been out of town the past 2 days.
My room is rougly 16x17.5x10. 2 of the walls meet the ceiling at 90 degree angles and the other 2 are vaulted into the ceiling. There's also a hallway off the front right corner of the room. Sub is an infinite baffle with 4 15" ascendant audio aresenal's mounted in the ceiling out in front of the screen a little bit.
Measurements were taken with a behringer ecm-8000 at the seating position (my seating position). I'm also using an SMS-1 to smooth the response. It was pretty good before but the SMS allowed me to get it even better. And I still have some tweaking to do.
I have 8 1" acoustic panels around the screen and 3 on one side wall. The other side wall has 3 windows which are covered with drapes. I also have 4 4" thick panels that I made (2'x4' with 6lb density fiberglass). One straddles the back corner and the other 3 are hung flat in other spots.
Here's a full range sweep of the sub and left front and some pics of the room.

**Sonnie** · **04-28-06, 11:05 PM**

Holy snootin' goots Brian... did you do that sub design yourself? DIY IB?

You oughta share that with us in the Gallery | Installation area... tell us all about how ya designed it and such.

**media2368** · **04-29-06, 08:20 AM**

Quote:

oliverlim wrote:

Hee. Good you see this up and it looks great! I just need more traps in my room as 5 units of 2x4x4" bass traps hardly made a dent in my low bass decay. The room just seems to store and slowly release the bass notes. Thats why I was thinking of trying another way of EQing while I am waiting for quotes on redoing my room.

Oliver

Hi Oliver,

Nice to meet you here! You have post a interesting topic.

Mine showing us how/where you place your bass trap?

Will re-EQ your both SUB solve the decay problem?

**bricor** · **04-29-06, 08:51 AM**

Sonnie,
I'll start a new thread and post more info and pics of the construction. I got "hooked" into it reading the cult of the infinitely baffled forum and decided to give it a try. I love it.
I've done 4 others for people since I did mine as well. Terrific performance for the money.

**oliverlim** · **04-29-06, 10:22 AM**

Quote:

media2368 wrote:

Hi Oliver,

Nice to meet you here! You have post a interesting topic.

Mine showing us how/where you place your bass trap?

Will re-EQ your both SUB solve the decay problem?

Here there

Good to see you on this forum.

I have 5 traps, 2'x4'x4". 2 in the front corners, 2 on each side beside the front speakers and 1 in the back corner wall.

EQ cannot solve room deacy. It can help with FR by playing with phase and reducing/adding on each sub to balance it out. But waterfall, sadly no. Yes you can reduce the FR and the waterfall looks to be lower, but the moment the volume reaches back to that same level, the waterfall will be back to that same decay level. At least thats what I found out.

Oliver

**JohnM** · **04-30-06, 05:02 AM**

Oliver, that would be because your correction filters are not well matched to the modes they are correcting. When the match is good the decay is fixed also. Try putting a filter in for your 45Hz peak, for example (at whatever the exact frequecy is) with a small amount of cut (say -3dB) and try varying the BW setting and observing the effect on the waterfall. You probably need a bit more range vertically, with the bottom of the graph around 45dB or so. Also use a longer gate time for your FR graph, 600ms or more, for better frequency resolution.

**oliverlim** · **05-01-06, 01:37 PM**

Quote:

JohnM wrote:

Oliver, that would be because your correction filters are not well matched to the modes they are correcting. When the match is good the decay is fixed also. Try putting a filter in for your 45Hz peak, for example (at whatever the exact frequecy is) with a small amount of cut (say -3dB) and try varying the BW setting and observing the effect on the waterfall. You probably need a bit more range vertically, with the bottom of the graph around 45dB or so. Also use a longer gate time for your FR graph, 600ms or more, for better frequency resolution.

JohnM,

What I dont understand is how Eqing for a freq can reduce the waterfall. What EQ does is to reduce the level of that freq right? So if you play it back to that same level after the EQ, the decay should look the same as it was? I will play around with this again to see if I can optimise it better and report again.

Thanks
Oliver

**JohnM** · **05-01-06, 02:47 PM**

Quote:

oliverlim wrote:

What I dont understand is how Eqing for a freq can reduce the waterfall. What EQ does is to reduce the level of that freq right? So if you play it back to that same level after the EQ, the decay should look the same as it was?

Yes, and no

It is often thought that an equaliser only affects frequency response and does not affect the time domain. This is completely wrong, everything you do that affects the frequency response will have a corresponding time domain effect, the two are inextricably linked. Parametric EQ filters have a resonant tail similar to the resonant tails of room modes, except that (when they have loss rather than gain) they decay much faster. Both the room mode and the EQ filter are second order systems, when the loss and bandwidth of the filter match the gain and bandwidth of the mode the mode's resonance is cancelled out, with an effect that can be clearly seen on the waterfall. The matching needs to be good however, with centre frequency within 1/60th of an octave of the mode's resonant frequency and bandwidth within about 1/30th of an octave of the mode's bandwidth.

**Wayne A. Pflughaupt** · **05-01-06, 08:06 PM**

Quote:

Sonnie wrote:

Man that looks neat as grits... just wish I could understand it.

It’s not that hard, Sonnie – let me ‘splain it to you.

The “waterfall” term might be what’s getting you. Actually what waterfall charts show is signal decay. They call them “waterfall” charts merely because that’s what it resembles. To get a handle on it, let’s take a look at a chart that might be a little easier to read than the ones posted above:

Notice that there are some similarities between this chart and the regular one-dimensional charts we’re more used to looking at. For instance, you see a graph for dB level (vertical markings), and frequency (horizontal markings - although this one could use finer resolution). If you focus on just the top line of the plot, you can see that it reads just like a regular one-dimensional response chart. For instance, we see a couple of peaks between 128 and 164 Hz. And a deep valley between 92 and 128 Hz. Make sense so far?

Now – that top-line plot is the baseline for the rest of the “waterfall” readings, which are the ones that look like they’re “moving towards you.” The readings that “move towards you” are showing how long it takes that particular frequency to fade away. This is known as “decay.” Decay is caused by reverberation – i.e, the signal bouncing back and forth off the walls and around the room. Decay is also caused by room modes (a.k.a “standing waves”), which are a build up of energy at certain frequencies, caused by the room’s dimensions reacting to the wavelengths of those frequencies.

Now, to be entirely meaningful, the “moving towards you” part of the graph showing signal decay should have some graduated markings in milliseconds (i.e., all the horizontal lines really close together). This one just says “800 ms,” so I guess that means 800 ms is the line that’s furthest “forward.” (However, that seems at bit ridiculous. Keeping in mind that sound travels roughly a foot per millisecond, 800 milliseconds would be essentially 800 feet. I have a hard time believing that decay times in even the worse residential rooms is almost the time/distance of a football field! Maybe it’s supposed to be 80 ms, not 800.)

Also, notice that (for the most part) there is a distinct correlation between signal amplitude and decay time. In other words, the peaks generally take a long time to fade away, while the valleys fade out pretty quickly.

It has much to do with a measurement standard known as “RT60,” where “RT” is “reverberation time” and “60” is “60 dB.” Basically, RT60 is the time it takes for a signal to reduce to 1/1000 of its original level after the signal source stops. A 1/1000 reduction in signal level is 60 dB.

Now, note on the chart I posted, and the ones Oliver and Brian posted earlier, that the lowest SPL reading is 55-60 dB. This is the point where audio is essentially reduced to ambient levels – i.e., the room’s noise floor. This is highly relevant. Obviously we don’t really care about the actual RT60 of signal that is only 70 dB to begin with. We’re not worried about how long it takes to get down to 10dB! Once it drops below the ambient noise floor, its essentially “outta there,” for audio playback purposes. So, the time it takes a signal to drop to 55-60 dB is what we’re primarily concerned with.

In case you haven’t already figured it out, it’s important to establish that waterfall charts and the delay times they show are entirely amplitude driven. On other words, all delay times across the board will rise and fall as you raise or lower the volume level, in relation to the 55-60 dB ambient level. Make sense? In other words - turn the level up high enough, and when that valley between 92-128 Hz hits 95 dB, its decay time will look as bad as the peaks on both sides of it do now (unless of course it's a true null). Make sense?

I’m sure you’ve followed some of those threads on some of the other Forums debating whether or not an equalizer can reduce the long decay times you see with those response peaks. The answer is &

Advertisement