John Sayers' Recording Studio Design Forum Archive

Thanks for the help I've received here in the past. I'll try to cut the small talk and be succinct here, because there are a lot of details.

The Room:
My studio/control room is above my garage. It is L-16', W, 13'(walls are splayed 5 deg, so this is average), ceiling is about a 6:12 pitch, with the ridge going across the room width wise. Ceiling starts at about 8' high and is about 12 at it's peak. There is a 4' flat spot to the ceiling running parallel to the peak. I had attempted to plot out the room modes, and it seemed favourable.

Floor is laminate, on 2 layers of particle board, on 6LB carpet underlay, on 3/4" plywood, on 2 X 10 joists, 16" O.C. The bottoms of the joists are open to the garage.

There is a room within a room. All interior walls are:2 layers of 1/2" drywall on resonant channel, on 2 X 4 studs. The ceiling drywall is against the resiliant channel which is directly on the roof rafters,

Treatments, etc:
-All cloth absorbers are made with 1X4 frames around 3" Roxol AFB.
-Corner absorbers have another 3" of AFB stuffed behind them.
-Ceiling absorbers are of the same design, and are hung from the ceiling from chains. They are about 4" from the ceiling
-Most of the panels are 2' wide, except the vertical corners, which are only 16" wide.
-Wall cloth absorbers have a piece of wood on top that fits into a wooden 'rail' that goes along the side and back walls. I can place a panel, move it, try different positions, etc. All wall panels are hung so they are about 3" from the wall against which they rest.
-Back wall plywood jobs are a bastardization of Ethan's "build a better bass trap" design. It is more or less that design, except that the plywood is installed in an arc, theoretically to diffuse the sound somewhat. Two have 1/8" plywood and two have 1/4"plywood. The boxes are completely sealed.
-speakers are soffitt mounted in angled front walls, about 54" to the cone of the woofer.
-at the top and bottom of each angled front wall is a space, about 18" high, about 2' deep, stuffed full of AFB. In the pictures they can be seen as having green cloth over them.
-Other than the aforementioned absorbers, there are two cabinets inset into the angled walls, and the rest of it is drywalled,

The Test:
I downloaded and used Ethan's test tone CD, set up my 4050, in omni mode, at mix position, about 54" from the floor, and about 6' back from the front wall. I used my Yorkville PA speakers just in front of the soffit mounted ones, because they go much lower than my wharfedale studio monitors, and because they are more suited to putting out decent energy in those lower ranges. I did from 40 to 210Hz

I left the room and the mic recorded the signal playing through the loudspeakers. Then I normalized the track so that the loudest part of it would be 0dB. I played back the signal and measured its strength with the meters in my recording software. I understand this is not the usual method, but my only SPL meter is in my phone and may not be accurate. This seemed to do the trick and now I have a permanent record of the test, as well as a visual waveform showing the wild variations in strength of recorded signal.

The Verdict:
The room is an abomination! It's all over the place. With the highest peak, 65Hz, at 0dB on the meters and it will rattle your fillings loose. The lowest point is 80Hz at -26dB. It's barely audible.

NOTE: I did some experimentation with other things. I took 4 wall panels off the walls and placed them on the floors in the corners to hopefully add some absorbing in the corners. I simultaneously added an unopened bag of insulation along the back wall at the floor. The effect was negligible.

I tried the test with both rear doors wide open, thinking if things were piling up there in a bad way it would make a big difference. Again, the effect was negligible. This leads me to believe that the rear of the studio is not the problem. If it was, I would expect big changes with another 40 or so square feet of open door at the back corners of the room. I think the problem is occurring "before" it gets to the back wall.

The Questions
So? Any recommendations as to a starting point?
Are my wide band absorbers too wimpy, to trap the bass?
Do I need to add more density all the corners?
Is the room just too blasted small and can't be fixed?
My open door test seems to indicate the back of the room is not the problem, so what possible causes are happening at the front 1/3rd of it?

Thanks for any thoughts. I'll try to upload those photos now.

Cheers
Keith

Bump.

I would suggest testing it with REW. You also need a lot of additional info and analysis to figure out what is going on in that room, not just basic frequency response. For example, a waterfall plot reveals how the room behaves in the time domain as well as the frequency domain, simultaneously


- Stuart -

Sounds like the mic/LP is in a null for @80Hz. Try moving the mic back and/or forward and to the sides a couple of inches. Also try moving it up and down.

The best way would be to generate just the problem frequency and then move the mic around while monitoring the recording level. Also put your head in the same place as the mic (if you don't physically hear the attenuation at this frequency, then you probably have an issue with the mic).

If this doesn't highlight the problem, it could be the speaker/s (you should really only test with one speaker at a time) is sitting in a null for @80Hz, so try moving the speaker a few inches (again in all 3 dimensions).

I doubt this has anything to do with your treatments, it sounds like a classic standing wave problem, and has more to do with room modes.

Hi gents. I thought I should do a bit more testing so I could give a more educated response:

Yes, there is definitely a gargantuan null, and as we see below, it increases and decreases in intensity and frequency as we move around the room.

I would also guess that it's the result of bad room modes, which sucks because I paid quite a bit of attention to that, arriving at my dimensions after much calculating. I guess it's the curse of a too-small room maybe, huh?

The spike, and null, are both moving targets. They move up and down as one moves around the room. To test this I tried the mic in different positions.

If I move the mic CLOSER from 80" from the front wall, to 56":
-the spike becomes somewhat less pronounced, and moves DOWN from 63Hz to 60Hz.
-The null too becomes less, and moves DOWN from 80Hz to 68Hz

If I RAISE the mic from 54" up to 80" (just out of curiosity):
-The spike moves UP from 63Hz to 68
-The null moves UP from 80Hz to 113Hz, and becomes VERY pronounced.

If I move the mic further away from the front wall, from 80" to 136":
-The spike moves UP from 63Hz to 72
The null moves UP from 80Hz to 86Hz

So I have more information now, I suppose. As we move closer to the wall, the spike, and null both seem to move down in frequency. Move further away from the front wall and the bad spots move up in frequency.

So now that these are the cards I've been dealt:

What might be a solution to address the bass buildup from about 55-75Hz?
Is it practical to build deep Helmholz resonators for the corners that are tuned to address those frequencies?
Or should I get a hold of some very heavy insulation(I believe OC705 is 6LB per cubic foot, no?) and build larger and thicker broadbands for the corners?

What, if anything, can be done about the huge nulls in the room, or are they addressed at the same time as the bass buildup?

Stuart, can an average guy with average equipment do a waterfall plot? and will it give me markedly different information from what I already have? I know I have huge issues at certain frequencies, and I now know those issues rise and fall in frequency as I move around the room, so do I not have to do first aid on the patient to get him breathing and his heart pumping first before I can worry about a diet and exercise program to tune him up properly?

Thanks all.
Keith

I would also guess that it's the result of bad room modes, which sucks because I paid quite a bit of attention to that, arriving at my dimensions after much calculating

Maybe there's a misconception here: ALL rooms have modes! No matter what ratio you choose, there will still be modes, and you cannot get rid of them: they are a simple consequence of having walls around the room. What changes as you adjust your ratio is the spread of the modes: how close they are to each other on the spectrum. A good ratio gets you an even spread, with the mode at roughly regular intervals, starting far apart at low frequencies and getting closer and closer as you go higher. Bad ratios get you modes that are not spread evenly: Some are right on top of each other, a the very same frequency, while there are also big gaps between some pairs, and tiny gaps between others. But the modes are always there. You cannot get rid of them unless you remove the walls from your room!

So what is a mode anyway? It's just a frequency where the wave happens to fit perfectly into one dimension of the room, and is directly related to the wavelength, obviously. Change the room dimensions and you change the frequency where the mode appears. That's all. So a wave that fits perfectly between 2 walls, for example. bounces back and forth between those walls, and has its peaks and nulls in the exact same positions in the room on each bounce. The wave is said to be a "standing wave, because it appears to be "standing still" in the room (which is a misnomer, really, as the wave still moves like any other: just the peaks and dips "stand")

For the same reason, as the speaker continues to produce that tone for several cycles, each cycle adds more energy to the wave, exactly in phase with the energy that is already there. So that specific tone gets to increase in loudness very fast, whereas a slightly different tone would not, because for that "non-modal" wave, even though it still bounces around the room, its peaks and valleys occur in different places on each bounce, so the speaker is not always adding energy to the same point, in phase.

Then when the speaker stops making that tone, the energy carries on bouncing around for a long time, since it reinforces itself on each bounce, and only dies away gradually. There's nothing to make it stop, so it carries on.

The purpose of bass trap treatment is to sort of tame the modes a bit, by damping their resonance. Think of it this way: the bass trap absorbs a bit of energy from the wave on each bounce, so it can't stick around for as long as it would without the trapping. The bass trap does not "cure" the mode: it just damps it. The mode is still there. Even a resonant trap, tuned to a specific room mode, such as maybe a panel trap or Helmholtz resonator, does not cure modes: Once again, it just takes some of the energy away on each bounce, taming the mode.

OK, it isn't really that simple, but I've found that this a good way to think of it: it puts a really complex issue into simple terms that you can get your head around.

So, rooms always have modes, and you cannot destroy a mode. That's the key issue here. You did nothing wrong in choosing a good ratio: you made sure that the modes are spread evenly, which is good. The issue now is that one of the modes is causing a problem, not that you messed up with your ratio. A different ratio would have most likely given you a different problem at a different frequency, nothing else. So don't blame this on a bad ratio. Blame on it on a problematic mode.

So, if you have this specific modal issue at 63 Hz (or whatever) then maybe the best plan is to build a resonant trap tuned to that frequency. Once again, it will not "kill" the mode: the mode will still be there. It will just extract energy from the mode, so that it is no longer a problem.

But it still might not be a modal problem you are seeing...

More later on the rest of your post: Gotta run urgently now...

- Stuart -

As we move closer to the wall, the spike, and null both seem to move down in frequency. Move further away from the front wall and the bad spots move up in frequency.

I think you might be looking at different modes, not the same one! You might also be looking at comb filtering, or other reflection-related issues.

OK, I just noticed that your walls and ceiling are all splayed, so you cannot predict the modal behavior, and any calculation you did on ratios is pointless. Ratios are only valid for rectangular rooms. Splay a wall, and all modes related to that wall are shifted in frequency in ways which are very hard to predict. Splay two walls and it gets even harder. Angle the ceiling as well, and it gets harder still. The only mode you can predict accurately now are your axial modes for the front and rear walls. Your other axial modes don't really exist any more as such, and all of your tangential modes, and all of your oblique modes are now in the realm of guesswork. (Yes, they CAN still be predicted, but you need FEM/FEA to do so, and that ain't easy...)

So instead of trying to predict, just measure! Predicting tells you what theory says the room should be doing. Measuring tells you what it is ACTUALLY doing, which is most likely different from what theory says anyway!

HOWEVER; you didn't splay much: about 10° or so, it seems. So a theoretical analysis for a rectangular room can give you a vague idea. It turns out that a rectangular room with the same dimensions as your averages, would have modes at 66, 70, 71, 79, 87...

You say you have problems at 63, 68, 72, 78, 86. Notice something? I reckon you are not seeing a change in frequency as you move around the room: Rather, you are looking at different modes, in different parts of the room. Some of those are tangential, some are oblique, so the peaks and nulls will show up in different locations. The frequency is not changing as you move around: Its a different mode in each place!

I'd go for superchunks in several corners, and if that's not enough then panel traps on the side walls, front wall and rear wall, tuned to the most likely candidate for each wall, and an angled hard-backed cloud on the ceiling.

Is it practical to build deep Helmholz resonators for the corners that are tuned to address those frequencies?

Wellll.... you could, yes, but that would be hard to tune right. Panel traps are easier.

Or should I get a hold of some very heavy insulation(I believe OC705 is 6LB per cubic foot, no?) and build larger and thicker broadbands for the corners?

That would make it WORSE! Contrary to popular belief, you do NOT want denser absorption for bass. What you actually want, is LESS dense. Heavier stuff is better for highs, lighter stuff better for lows. 703 will work fine. In addition to the panel traps, I'd build superchunks for the rear corners.

What, if anything, can be done about the huge nulls in the room, or are they addressed at the same time as the bass buildup?

Yup! A null is just the opposite end of a peak located somewhere else in the room, so if you damp the peak, you also damped the null... Sounds logical when you think about it that way: they are just two facets of the exact same wave, and waves are just energy in motion. A null is just a peak, inverted, but it still represents energy associated with a wave. So taking out some energy from the wave means that you lowered the peak and also raised the null at the same time. So putting in bass traps will deal with both issues, since it is actually the same issue, just seen from the opposite side.

Stuart, can an average guy with average equipment do a waterfall plot?

Yup! That's why I suggest you download REW. It does it for you.

and will it give me markedly different information from what I already have?

Oh yes! BIG time! REW gives you a whole slew of very useful data, but even just the waterfall plot is a hugely valuable tool, since it shows how your room is behaving over time, not just what frequencies are giving problems. It is the certain indicator of where your modal issues really are, which you cannot know from just looking at frequencies. Frequency response curves show you the peak values that the room attains for each frequency with a steady-state tone. It does not show you how that changes over time once the tone is gone, and that's what matters. A peak on a response curve could be a mode, or a comb filtering artifact, or a reflection of some kind, or several other things. But when you look on the waterfall plot, you see exactly how the room is releasing the energy stored in modes over time, and that is a major clue as to your modes. Modes store energy and release it slowly. Comb filtering artifacts do not.

I know I have huge issues at certain frequencies, and I now know those issues rise and fall in frequency as I move around the room, so do I not have to do first aid on the patient to get him breathing and his heart pumping first before I can worry about a diet and exercise program to tune him up properly?

Naah... Just download REW, do the acoustic equivalent of a CAT scan on him, and post the results...


- Stuart -

As Stuart correctly reckons, you are not seeing a change in frequency but in fact the different modes. It requires a certain amount of energy input to "trigger" the modes so what your absorber treatments are doing in effect is attenuating the energy available to drive the modes. Unfortunately, absorber treatments are broadband in nature, so this reduces your overall frequency responce. This means that a tuned trap might very well be in order.

One thing I mentioned before which should not be overlooked is the positioning of your speakers. Just as you've experienced with the mic sitting in an antinode (null), think for a minute what would happen if your speaker were sitting in a null. You could drive the speaker as hard as you want, but it just can't (apparently) produce that specific frequency. Similarly if it is at a node, the apparent loudness of the specific frequency will be extremely dominant.

So I suggest you also move the position of the speakers just as you repositioned the mic. I know this complicates the exercise as now you have more than one variable - but that is unfortunately the reality. Nobody said this would be easy did they? This is another reason for testing with just one speaker at a time.

Okay. Thanks for the starting point.

A question...or two.

I have a pair of 36" wide doors at the very back corners, on either side of the room. Oddly when I opened them wide and did the experiments, there was almost no measurable change. That makes me wonder if I should bother to put superchunks or whatever back there.

Onward...

You saw no measureable change because your problem (the one you're seeing at the moment) is probably not a modal problem based on your rooms length, but rather the width or height. Once this is dealt with and you do further testing you'll see the effect of treatment at the rear.