CR Treatment and Tuning---The home stretch!!

[Stadank0]

Well,

Struggled with this a bit today but am getting closer to a working product..I think..

Got some minor leveling issues to sort with the lengths, but otherwise the unit seems doable.

20180119_160847.jpg

[Stadank0]

One down one to go!

20180120_141626.jpg

[Studio45]

Great work Frank!

Just curious, is there some insulation material behind the wood? or is it a blank space behind. Stuart's put a lot of details on this thing. Very very cool!

Cheers!

[Stadank0]

He had me put a small 1/2" thick bit insulation up against the posts, I used a shred of pink fluffy. Also a 1" piece of 703 against the back wall and 1/2 of the floor. The idea is that the box is an air spring for the resonator rather than an absorber. Because the tuning of a resonator is dependent on the volume of the air spring and size of the opening, you don't want to fill the space with insulation which would thereby change the math.

Should have the project completed by tomorrow's end. rain has been cramping my style a bit and the resonators took focus and attention to detail of course...will post more pics in a bit!

[Soundman2020]

is there some insulation material behind the wood? or is it a blank space behind.

There's a thin layer of light insulation right up against the rear of the "slats" (which are actually trimmed-down 4x4's), and there's also a thin layer of of 703 up against the back wall and on half of the "floor", but most of the cavity is empty space. This is Helmholtz based, so it has to be that way. It is tuned, but not too tightly. Reason: It's notoriously hard to tune Helmholtz devices accurately! They normally have very sharp tuning (high Q, very narrow bandwidth), and the issues they are designed to hit also have very high Q, so it's not a simple matter at all to line up the tuning with the problem. And if you miss, then it doesn't work. Slight inaccuracies in materials or construction can shift the tuning way off. Even having the wrong air temperature or air pressure can move the tuning far enough off that it no longer hits the target at all. So on a cool day, or a hot day, or when a weather system moves through, that would change the tuning of a high Q Helmholtz device (very tightly tuned).

In this case, I have tuned it with lower Q so it works less effectively, but across a broader region, regardless of inaccuracies or changes in the weather.

More accurately, I have "de-tuned" it slightly, so that even if the tuning is off, it will still hit the problem frequencies, albeit with lower effectiveness. In other words, the sharply tuned peak is lower, but broader: it covers a wider range of frequencies, but with lower intensity. The thin insulation on the rear wall and floor of the cavity does that. The even thinner insulation behind the slats is what damps the resonance, removing the energy from the room.

It's a common mistake to completely fill the cavity, thinking that "if some is good, then more is better". Wrong. "Some" is best, "More" is far worse. In fact, filling the cavity completely will pretty much kill the devices, so it won't work at all. To convince yourself of this, try making a classic "Helmholtz" resonator with a small empty glass coke bottle: Blow across the top of the bottle neck, to make it "sing": that's a classic demonstration of Helmholtz resonance. Now put a tiny little piece of something light and fluffy in the bottle neck, such as a small wisp of cotton wool that you have gently pulled apart and "fluffed out" with your fingers, and blow across the top again: the tone will have gone down a bit, and it will now be rather dull and muted: you have effectively damped the device, and it will absorb rather well at that frequency. Now fill the entire bottle cavity, all the way, top to bottom, with insulation (eg, pink fluffy, 703, etc). Stuff it completely full. Blow again. Nothing happens. There is no tone at all! It no longer resonates, not even slightly, because it is severely over-damped. If it doesn't resonate, then it isn't working... You killed it....

So that's the principle here: Yes, there is some thin insulation in there, carefully placed, to slightly de-tune the device so that it covers a broader range, but without killing it.


- Stuart -

[Stadank0]

And of course what he said...lol

[Soundman2020]

And of course what he said...lol

Sorry Frank! Looks like we posted at the same time! I hadn't seen yours when I posted mine...

Nice work, by the way! It looks pretty good!


- Stuart -

[Stadank0]

Thanks Stuart,

I'm feeling pretty confidant. I can easily access the boxes to tweak them right behind the cover if we need to. I was wondering, If we use veneers will they insert into the slots or go over the face?

[Studio45]

As always thanks for the explanation! I figured there had to be something behind those pretty wood slats.

I fully understand how you're trying to "tame" a certain frequency range versus trying to kill-it completely. I can also imagine how complex it would be to figure out what to build in order to do it effectively without over killing it or lacking.

Great work! I can't wait to be at the "treatment" stage of my build.

Cheers!

[Soundman2020]

I was wondering, If we use veneers will they insert into the slots or go over the face?

They would have gone into the slots, but there's no need for that, as far as I can see. Your measurements are within spec, and in fact having the posts slightly narrower is working to our benefit: It is tuning them a little higher, which is good. To keep things simple for you, I approximated to the nearest 1/16", which turned out to tune them all a little low. I wasn't worried about it, as the detuning still had them covering the problem issues, but now they are closer than they would have been, so all is well. No laminated needed.

For example, there's an issue in the room at 131 Hz, and with the original dimensions the tuning for that would have been at about 128 Hz. Allowing for +/- 5%, that's fine (128/131 = 97.7%). But now it is tuned up to 129 Hz, which is 98.4%. So no problem. Of course, if it was NOT de-tuned, then it would have been a problem....

Just a word of caution here: These things are not magical "mode killers"! People tend to expect spectacular results from complex devices, but the effect won't be huge. Just a few dB, and a few dozen ms. It's the combination of all the treatment that does the trick, not one single device. One step at a time, with each step doing "something but not a lot" is how it works. Think of it this way: If you can do ten small things that each flatten out the response by just 1 dB and 50 ms, then in total you flattened it by 10 dB an 500 ms!

So don't get disappointed and discouraged when you see only small gains with each round of treatment... It's the end result that matters.

- Stuart -

[Stadank0]

of course. My expectations are well adjusted...

[Soundman2020]

I can also imagine how complex it would be to figure out what to build in order to do it effectively without over killing it or lacking.

There's also the issue of being careful to use the right equation! There's actually a wrong version of the Helmholtz equation circulating, and it has even been printed in several text books: It's a simple "typo" error, where there's a "plus" sign instead of a "multiplication" sign.... so it's sort of important to check that you even have the right equation, when designing slot based devices!

Wrong equation:

fo = 2160*sqrt(r/((d*1.2*D)+(r+w)))

Right equation:

fo = 2160*sqrt(r/((d*1.2*D)*(r+w)))




- Stuart -

[Stadank0]

Here's the Bass Trap... A lot of work...

Fortunately Stuart is pleased with the results... I'm sure he'll be commenting soon. Very encouraged!

20180125_151900.jpg

20180125_151839.jpg

20180125_151850.jpg

20180125_154514.jpg

20180124_102526.jpg

[Soundman2020]

Update on what all of Frank's hard work actually accomplished. It's rather nice, actually.

So here's some pairs of "before" and "after" graphs from acoustic measurements he did. The "Before" measurements were taken immediately prior to him starting to put in treatment, in the otherwise-empty room. The "after" measurements were taken yesterday afternoon, as soon as he had that rear-corner monster in place.

First, the waterfall plots, which show how the sound pressure levels decay over time, for every frequency. In this graph, the frequency axis runs across the page from left (lows) to right (highs), the intensity is shown in the vertical axis, up the page, and the time axis comes out of the page towards you. In this case, we are looking at just the bottom end of the spectrum, from 18Hz up to 500 Hz, because that's the most critical part. In fact, what's REALLY important here is the part under 200 Hz:

BEFORE:

Frank-REW-WF-20-500--131hz-highlighted--BEFORE.png

AFTER:

Frank-REW-WF-20-500--131hz-highlighted--AFTER.png

You can clearly see how the response has smoothed out very significantly, with all those huge mountain peaks very much flattened and rounded into hills. In a previous post, prior to Frank building this thing, I did mention that one of the tuning points was 131 Hz, so I marked that with the cursor in this graph (a white line overlaid on top of the "mountains". You can see that it is working very nicely, with both the intensity and the ringing greatly attenuated. But the biggest issue here is what happened down lower, towards the bottom end, which is always the hardest to treat: There were some pretty big modal issues at 50Hz, 59 Hz, 69 Hz and 94Hz (HUGE!) that have almost completely disappeared now, and the big one at 114 Hz is greatly reduced, as are several others.

Another very interesting one, is the RT60 graph. It shows the reverberation time for many small frequency ranges (just one third of an octave wide) across most of the spectrum. It's not really technically accurate to call the low end response "RT60", since there's no statistical reverberant field in small rooms for low frequencies, but most people still call it that. More accurate would be something like "energy decay times", but RT60 is fine.

BEFORE:

Frank-REW-RT-20-20k-BEFORE.png

AFTER:

Frank-REW-RT-20-20k-AFTER.png

Those cover almost the entire spectrum, from about 35 Hz to 11 kHz, and you can clearly see how the rear corner devices are having a major effect on smoothing things out. Before the treatment, the overall RT60 time for the room was around 900 - 1100 milliseconds: Now it is neatly and smoothly under control, at around 400ms. Also note that the high end is still there, not killed, and it's looking good. We are aiming for something around 250ms for this room. Maybe a little more.

Next up; the spectrograms. These show the same data as the waterfall plots, but in a different representation. 2D, not 3D. Here the frequency axis is still across the page from left to right (lows on the left, increasing to the right), the intensity is shown in different colors, and the time axis runs up the page: higher peaks mean longer ringing, slower decay.


BEFORE:

Frank-REW-SP-20-500--131hz-highlighted--BEFORE.png

AFTER:

Frank-REW-SP-20-500--131hz-highlighted--AFTER.png

The improvement is pretty spectacular, actually! Each of the long thin spikes in the "before" image is a room mode. You can see how narrow and sharp they are: very thin frequency bands, very intense. In other words "high Q".

And you can see how all of that has flattened out and smoothed over very nicely, with the monster at the back of the room. The SBIR hole that was there before at around 110 Hz is gone, and overall it is a lot smoother. Once again, I highlighted the 131 Hz issue, so you can see that the tuning is working out fine. The monster is eating the modes!

And finally, the graph that everybody loves to look at first but actually doesn't tell you a lot: frequency response curve. All it shows is the sound intensity for each frequency. Not very useful, but here it is anyway:


BEFORE:

Frank-REW-FR-20-500--131hz-highlighted--BEFORE.png

AFTER:

Frank-REW-FR-20-500--131hz-highlighted--AFTER.png

The smoothing is pretty clear here to. Once again, look at the 131 Hz point, to note that the device is doing it's job, and you can clearly see how the peaks at those other problematic modal frequencies I mentioned before (50Hz, 59 Hz, 69 Hz and 94Hz) have all come down. The results at 59 and 69 are very satisfying, as that low down is very hard to treat. That's the deep bass region, always a tough one, but it's coming under control rather nicely. The change at 94 Hz is also rather substantial.


OK, one more: this is a comparison of "before" and "after" on one single graph, so you can see the difference more easily:

Frank-REW-FR-20-200-COMPARE--BEFORE-vs-AFTER.png

That one is zoomed in to show just the most important part: 18Hz to 200 Hz. Purple is "before", green is "after".

The overall result is better than I had hoped for, so I'm pleased with it. I think the graphs speak for themselves!

Consider that this is a corner control room, and also consider that the ONLY treatment in the room at this point is the rear corner. We still have the ceilings and walls to play with, plus the cloud, and the soffits. So there's plenty of room for playing still.

The rear corner bass trapping usually gives the most spectacular improvements, as you can see here, and subsequent treatment accomplishes less and less with each round, but I'm confident that there's enough room to get things rather nice for Frank. The ceiling is up next: Frank is starting to work on that right now. That should smooth out some of the other spikes that didn't change much. They didn't change because the are associated with the vertical axis of the room (ceiling and floor), not the walls. So treating the ceiling should take a bite out of those.

After that we'll be doing the cloud, and the soffits. Not sure which one will come first...

STAY TUNED!


- Stuart -

[Studio45]

That's once again great work guys! A lot of work for sure Frank but you did great man looks awesome!

So Stuart you design these bass traps/treatment once you've made a proper analysis of the test results correct? In other words you can't fully predict what's required when it comes to treatment until you know how sound gets thrown around the room?

I'm amazed at the level of details that go into this sort of thing..I love it!