John Sayers' Recording Studio Design Forum Archive

So after looking at the results of these measurements and nothing better to do until I hear back from Stuart, I have a theory...

There is now a significant dip at around 177hz that was not there before. Of course there are improvements as well so this is not something that I'm worried about, but more interested in understanding.

I understand that as a broadband absorber, insulation absorbs some amplitude and also affects phase and decay as well(on a broad spectrum). If the null was not there before the insulation was installed, then I theorize that phase shift at that frequency with regards to SBIR is causing the null.

I noticed no appreciable difference between the covers on and off save the null being slightly worse with the covers on. This of course adds the 1" of duct liner. I further theorize the the duct liner would further exacerbate the phase shift and increase the null with regards to SBIR.

Can't wait to see what techniques Stuart has up his sleeve in this game of whack a mole!

Before: After:

If the null was not there before the insulation was installed, then I theorize that phase shift at that frequency with regards to SBIR is causing the null.

It might have been there before, just masked by NOT having all that absorption around: In other words, the absorption trimmed the long grass way down, so now you can see a few rocks and stones laying down on the ground that were there all along, but not visible before. Or you cut down some tall trees, and now you can see the shorter ones that were hidden behind them.

I think we might be looking at an issue that was there before, but moved up the spectrum slightly and is now much more visible because you "trimmed the jungle". If you take a close look at the spectrograms, you'll see that there was a vague hole around 165 Hz in the untreated room, which moved up to about 167 with the rear trap, then 171 with the ceiling absorption, and now 174 with the side wall absorption..

It's definitely phase-related, though. You can't see it on the straight phase plot, but you sure can on the minimum phase and excess phase plots!:
No time to go into what those two graphs actually mean (minimum phase and excess phase are a bit complex to explain), but in general they are a lot more useful than the plain phase data, and here you can clearly see some major stuff going on at the same point where that dip is happening. So it's pretty clear that a reflection of some type is responsible for this.

So . . . I'm wondering if that is a floor bounce, perhaps, or maybe from the front walls... Quick check on that: Put a thick pad of insulation on the floor, between the speakers and the mic, like this: Lay down a full panel of 4" thick 703 in front of each speaker, as best you can fit it in, then throw some pink fluffy on top of that, and do a REW test. It will be interesting to see which way that null moves (if it moves at all), and how the amplitude and reflections change.

Next point:

Your graphs: when you look at your REW graphs, apply some smoothing so you can see what's really important. If you are looking at the low end, then apply 1/48 octave smoothing, or maaaybe 1/24: Anything higher will hide modal issues. If you are looking at the mid range, then apply something like 1/24 or 1/12, and if you are looking at the high end, then 1/6 or maybe even 1/3 would be more appropriate, depending on what you are looking for. As rule of thumb, when I'm hunting for "needles in a haystack", I use 1/48 for lows (or maybe no smoothing at all), 1/24 for mid and 1/12 for highs. If I'm looking more for "How will the human ear perceive this?", then I'd go more like 1/24 for lows, 1/12 for mids, and 1/6 for highs. Of course, the specs for control room acoustics usually define things in very low res, such as 1/3 across the entire spectrum, but that does not show the details. And as the saying goes, the devil is in the details!

If you follow the above "rules of thumb", you'll see a much cleaner graph that is far more useful.

In addition, you'll probably want to zoom in a lot more, so you can better see those details where "the devil is"! For looking at the low end, I normally set the horizontal scale to cover 18Hz - 500 Hz, and the vertical to cover 40 dB - 110 dB. That's where nearly all of your major issues will be. For some issues, I go even tighter, and use 18Hz - 200 Hz with 50db - 100 dB. You get to see all the ugly stuff more clearly like that, and everything below about 200 Hz is critical for bass response.

On the other hand, if you take a quick look at the second graph you posted above, in the gold color, you'll notice that your unsmoothed frequency response is already within +/- 10 dB of flat, across the full spectrum (except for that biggie at 170 Hz)! That already puts the room well within acceptable response for home studios: +/-10 dB is pretty good. +/-6 is really good, +/-3 is high-end pro-level, and better than that is magic! So you already have "pretty good" frequency response, unsmoothed, but we are only just getting started!

OK, enough rant, on to practicals: As expected, all those hundreds of square feet of pure absorption sucked out all the life from the room: it's dead now. That's good, actually, since it took it out fairly evenly across the spectrum:
So we are starting from a good position, where the decay is nicely even across all frequencies, just too low over all. And all we have to do now, is to add some life back in to the room with carefully crafted reflections. OK, that sounds way too easy... the concept is easy, but the implementation... not so much!

We need some more data to see which way to go here. So first test (after you do the "insulation on the floor" test above, and remove that insulation from the room again!); if you have a couple of full sheets of plywood, OSB, MDF or something similar on hand, lean those up against the side walls, symmetrically (same on left and right). No nee to attach them: just lean up temporarily. Thickness not too important: it's surface area we are after. So try putting one full panel on each side of the room, do a REW test, then two full panels on each side, and another REW test. That will give us some data points to work with, in figuring out how much total reflective area you need to get things looking nice again. Don't expect the response to be even, though! It's gonna mess up that nice smooth decay curve, but that's fine. If you don't have full panels, then use whatever large scraps and off-cuts you have (hopefully nice long ones that go almost all the way up the wall), trying to keep the surface area balanced between the walls. You could even use a couple of old doors, or any other large, flat, rigid surface that you happen to have on hand: counter top, table top, deck planks, whatever.

So there's some things you can play with today, to help figure out what is going on, and which way to go to fix it.

I wont be available for the rest of today and tomorrow, as there's a yearly planning meeting I have to attend (and it's also the weekend!), but I'll check the data as soon as you sent it to me.

If you wanted more things to do to keep you busy, after the above tests (remove the wood panels from the room first...) you could do 40-50% coverage of the insulation with that 1mil plastic, but do it in horizontal strips above and below ear height, and not going all the wall to the floor and ceiling. In other words, leave a strip of bare insulation at the top and bottom of the walls, and another bare strip about 18" - 20" wide centered at ear height (48" above the floor). Do yet another REW test once that is in. That's not going to make a big change here, but it will help with some of the lost life, at the high end.

- Stuart -

I'm on it! will report back soon with results...

Things are definitely progressing in a hopeful way. I started this project with home studio ambitions but those have changed after 3 years and lots of money invested. My hope is to get to a professional level result so that I can monetize the space. I know that was not the scope of the original design, but I'm hopeful that we can get within range based on how things are shaping up....

Thanks for the graph lesson by the way. Useful for sure . Was posting those quickly to show the 177 null but will definitely use the guidelines going forward...

My hope is to get to a professional level result so that I can monetize the space. I know that was not the scope of the original design, but I'm hopeful that we can get within range based on how things are shaping up....

I kind of figured it would go that way! It often does... However, it might have been easier if I would have known in advance! But we can still get it good, I reckon...

- Stuart -

i knooooooowww...if only!! I never would have guessed back then, I can tell you that.


In all reality, I still believe that the quality of work done in this space will speak for itself. I have no doubt that outstanding engineering work gets done is worse spaces that is totally passable and respected in the real world.

There are several local studios (excluding Universal Audio of course) that have produced hundreds of projects on a professional level that were not done to this level of detail. I'm certainly not trying to be a full blown LA studio or the like..That would be unrealistic.

forgot to mention,

I think I'm gonna hold off on adding the 1 mil plastic until we have a reflection scheme because I may have to modify the panel frames to suit. I will put it in ultimately.

OK,

task complete! now its off to go do this:


I highly recommend this sport!!!

Soundman2020 wrote: Basically there's a small superchunk right at the back, floor to ceiling, as an anchor for the other things, with a tuned slotted poly-cylindrical device at the top and bottom, an un-tuned one in the middle, hangers in between, and un-tuned slats across the front. I won't go into all the details, but this bunch of devices is a total of about 39" deep (far tip of corner to front face), and does a bunch of things in different places across pretty much the entire spectrum, but mostly in the bass. It will hopefully do some nice things to the modal response, SBIR, and overall low end, while at the same time NOT killing the high end too much. It's designed to be balanced: it targets some specific things while actively NOT targeting other things.

Hey,

Nice design. Can you clarify some things in the system for me?

1. MIDDLE PART
As I understand it - this is a diffuser of sorts, is that correct? Do the slats in the middle section match the placement of the slats in the top and bottom resonators? Is it placed specifically at ear level to diffuse or reflect back into the room?

2. TOP AND BOOTOM
Are these resonators identical? I assume using a superchunk of 703 in those corners would be much less effective, correct?

3. GENERAL RESONAT QUESTION
When you designed these tuned resonatos - does the box have a wall inside it to match the curvature of the front or back of the slats? Or is it just a corner box? What im trying to wrap my head around is if the shape of the enclosure is a specific design element or not. In this case, the distance between the back of the resonator and the slats gets less the further you go from the middle. Does this have an effect of variating the resonant frequency? My understanding is that the priciple of this box is that air resonates at a frequency corresponding to the volume of the cavity and the slat with, depth and space between them, regardless of the variating distance between the slat and the back wall. Is that correct? I hope I explained everythign correctly..

4. MEASUREMENT
After you measured the room and saw the problem you need to fix, you based the design on the measurements you took in the listeners position, correct? If i was to measure in the corners themselves and design a resonator based on those measurements - would it make sense acoustically or would it be stupid?

Thanks in advance guys, I hope Im not asking for too much

Cheers,
D

I highly recommend this sport!!!

I hope you made it through that without too many broken bones!

Anyway, I analyzed your data, and it seems you only had one plywood panel on hand, which you moved around the room? Or more likely, I didn't explain very clearly..... That's enough to get some useful conclusions, but it would be better if you could do additional tests, with two panels (one on each rear wall), then again with four panels (two on each rear wall). That will give me the data I need to confirm the plan I have in mind.

OK, on to the test results:

First, the floor bounce theory for the 175 Hz thing is partially confirmed. Here's the frequency response curves for before and after you put the insulation on the floor: Gold colr is before, turquoise color is after. It's also interesting to note the overall smoothing effect of that insulation pad, on the lower mid range...


You can clearly see that the 175 Hz dip has disappeared, but there's a new, smaller one at 160 Hz. That suggest that perhaps the dip is still there, but just moved down the scale a bit, which is strange... that sort of implies that the room got bigger! I'll wrack my brains a bit to figure it out.... In theory, it COULD be because the insulation slows down the speed of sound, and changes the path length like that, but the shift seems too big for that.... Hmmmmm...

On the other hand, the IR response graph shows. crystal clear, that you nailed the reflection with that insulation:

Here's the "before" and "after" plots for that. Look at 4ms....
Very clear difference there. The insulation completely wiped out that reflection, which indicates that it really is gone, and the 160 Hz dip is something else..

Conflicting data... I need to put my thinking cap on....

But either way, clearly it's not an option to have thick stacks of 703 on the floor, so we'll have to do the treatment for that on the front side of the desk. Have you decided on a desk yet? If not, I'd suggest to make it as small as possible, and as low profile as possible. Avoid large flat surfaces, if you can.


The good news is that the single plywood panel brought up the decay times from around 100 ms to around 145 ms, as expected:
The bad news is that it did so MORE for the lower mid range, than it did for the highs... also as expected, unfortunately. So you will need some mix of treatment there, with a larger area aimed at the highs, and a smaller are aimed at the mids.... So it looks like you earned yourself a pair of non-tuned slat walls!

Your data also allowed me to identify the reflections and where they are coming from, which is great, as it makes it easier to avoid problems in the treatment, and easier to target what needs to be targeted. For what it's worth, here's the mug shot of all the guilty suspects:
So that's where we are right now. If you can get more data, then I'll design you some more treatment to put in place!

- Stuart -

I sent the new mdat file!

So the (plywood test center) is 1 sheet on each wall and the last one un-named is the 2 sheets on each wall...

I've been mountain biking for about 17 years now...The trick is to get good at it before you go and do what I'm doing in the photo ..I've definitely had a few injuries along the way. I climb about 8000 ft a week. There's no better way to stay young. How else do you think I managed to build this monstrosity at age 47 by myself?... lol I loves me some mountain biking!

I hope Im not asking for too much

Well, Frank is the paying customer here, so it's up to him how much of the design he wants to release! I don't mind going into it a bit more, but he's paying for this specific design, and is very kindly sharing some of that in public for others to see. I don't like sharing anything that I design specifically for an individual customer, for his particular room, unless he gives me the OK to do that.

So it's your call, Frank!

- Stuart -

So the (plywood test center) is 1 sheet on each wall and the last one un-named is the 2 sheets on each wall...

That was fast!

I just took a quick look at the data, and I'm liking what I see...

Give me a while to analyze in more detail, and I'll get back to you. Do you have other plastic on hand, thicker than the 1mil? 3mil by any chance? 6mil?

- Stuart -

Soundman2020 wrote:

I hope Im not asking for too much

Well, Frank is the paying customer here, so it's up to him how much of the design he wants to release! I don't mind going into it a bit more, but he's paying for this specific design, and is very kindly sharing some of that in public for others to see. I don't like sharing anything that I design specifically for an individual customer, for his particular room, unless he gives me the OK to do that.

So it's your call, Frank!

- Stuart -

Yes, that is perfectly understandable

Lets wait for frank to chip into this

Yes,

I have some 6 mill and can get 3 mil easy...so whatever you like..Feeling excited again! can't wait! off to go ride bike!! lol

I sent the new mdat file!

I have taken a better look, and it's nice. Well, it's ugly, actually, but it's what I expected, so that's nice! It's more like: "The Good, the Bad, and the Ugly"!

The good part is this: That's the "RT-60" graph, or more accurately: decay times. As you can see, the overall decay time has come up rather nicely now, to roughly where we want it: around 230 ms. I'd like a little more, maybe 250, but there's plenty of room for doing that still. So that's the good part.

Not so good (but expected) is that it isn't flat and even. The decay times are too low in the low end, and too high in the high end, making the room slightly "boomy", and a bit less "bright", somewhat "dull" in fact. So we need to tweak the treatment more to the direction of the high end: reflecting more highs, less lows. Easy to say, not so easy to do....

Then there's the ugly part: All of those reflections are highly specular! There's a dozen new and strong reflections coming at you from all over the place, produced by those panels on the back wall. Each spike on that graph is a reflection, and if you check the horizontal scale, you'll see that I'm concentrating on the Haas time window: all of that is within the first 20ms or so. The most important part, psycho-acoustically. The part where reflections will mess up your ability to determine directionality and frequency.

This is why it is soooooo important to look at ALL the data that REW gives you, and to do it intelligently! For example, I prepared a couple of graphs here by very careful selection of parameters, to show that Frank's room is already world class!

The frequency response is within +/- 5 dB of flat, which is really good for a home studio, and up there with the pro studios: Ain't that amazing! And just look at the incredible waterfall plot! Hard to beat that!...

Take that along with the nearly flat RT60 response from above: And you are done, Frank! Your room is obviously top notch. So stop building and start mixing!


Except, of course, that those are all highly smoothed, carefully selected graphs, and I did NOT show the ones that matter most, and cant be faked so easily: Spectrogram, and the most important of all: Impulse Response. Those reveal another story...

So that's my "rant of the day" for anyone following this thread: be wary when you see REW graphs that look really amazing: Check the scaling, check the smoothing, and check ALL the data. By selective posting, you can make things look a hell of a lot better than they are. That's one of the main reasons why I shared ALL the graphs on Studio Three Productions and will do the same here: I'll use the highest resolution and the best parameters that show up the problems, rather than trying to hide them. That's the honest thing to do.

It turns out that even with those honest criteria, Frank's room is starting to look good...

[ RANT mode = OFF ]

OK, what to do now: Take the wood panels out, get some of your 6 mil plastic, and put a strip about 9 inches wide, horizontally across the top and bottom of both walls, leaving a gap of about about a foot of bare insulation, above the top one and below the bottom one. Do a REW test like that (L, R, and LR) and send me the data.

Then leave another gap of about 6 inches bare insulation, and put a strip of your 1mil plastic about 4 inches wide across the both walls. Do another REW test (L, R, LR), send data.

So at that point, starting at the floor and going up: you'll have about 12" of bare insulation horizontally across both walls, 9 inches of 6 mil plastic, 6 inches of bare insulation, and 4 inches of 1 mil plastic. Then a large empty gap across the middle of the room, followed by the same but in reverse order: 4 inches of 1mil, 6 inch gap, 9 inches of 6 mil, then 12 inches of gap up at the ceiling.

CAUTION! DO NOT STRETCH THE PLASTIC TAUGHT! Leave it slightly "baggy". We don't want it acting as a membrane as well as a reflector...

That should start to return some of the top end to the room. It might sound a little "zingy", but don't worry: there will be wood slats going over some of that. Well, in reality, your cloth will go on first, then the wood slats. So from the inside going out, you will have : mineral wool insulation --- plastic --- fabric --- slats. Nothing else. But don't put any fabric or slats on yet! We might need to adjust the plastic, based on the REW data you send me from the two "plastic" tests, and I still have to figure out slat sizes, also based on the REW data...

(Just to clarify: in some places there will be slats that are over plastic on the other side of the cloth, and in some places there will be ONLY plastic on the other side of the cloth,, with no slats, and in some places just bare insulation on the other side of the cloth).

- Stuart -