Room voodoo

[jmh]

I have a region of frequencies around 100hz which can resonate many decibels (~30db) higher in regions of the room. I was experimenting with playing a 60 second 20-300hz sin sweep. I recorded the room with a ribbon with the null orientated toward the speaker. I got a volume bump at around 50 - then a giant spike centered ~100 then more mild (and low concern) bumps and dips in volume to 300.

I'm wondering how many of us have worked on this and with what measure of success?

It seems to be dump a ton of money into pro treatment or DIY - which in some cases seems to be bolt stuff on the walls and ceiling until the room geometry is transformed...

[John Ludlow]

Hi John - I think the course of action, and the degree of success you can expect per dollar, depend a lot upon how much of the room you want/need to control at a time, and how thoroughly you need to control it. In my space, I want to be able to generally avoid the biggest problems everywhere, but I accept that the room is going to contribute some character to the recording.

Just the same, I want it as flat as possible at my listening position - the one spot where I mix. My friend and I built twelve 4'x2'x4" frames that house acoustic fiberglass that has been covered in rayon batting (for safety). These can be hung from the wall, or mic stands, or tied to wheeled chair backs and placed where needed for the situation to avoid the worst room effects - in particular: managing low notes.

But if I point a mic at a monitor from a few feet away and record pink noise coming out of the monitor - it's not the same when I run that recording through a frequency analyzer. That is - it's no longer pink noise. It's improved a lot from the room without treatment, but the room is still in the recording. The room has modified it. That makes mixing problematic.

So, I also use a VST software product that comes with a reference microphone to cause the mix to be flat in my mixing listening position. In its setup, it produces sound while it listens to the result with the mic. Then it analyzes the difference between what it played and what it heard and adjusts frequencies throughout the spectrum such that the result is flat when it gets to the mixing position.

It does a much better job in a room that has its worst problems partially solved (bass traps). In my case I still had one very low note that is amplified disproportionately - but it's really low. And everything else is really pretty flat. There's a whole series of posts in here from several years ago that reference screen prints of frequency analyzer output that prove my case.

You probably will never get a perfect room - only an improved one. But you can make it near perfect in the one spot where you need to be able to listen accurately. For the rest, even an improved room still makes a huge difference. There's ingenious other stuff you can employ besides traps. There's a whole other thing you can do by causing simple signals to be defused and broken up and become more complex that make the space sound bigger than it actually is. And there are sites where you can define the geometry of your space and position of your monitors and they will suggest placement of various of their trap and diffuser products within that definition to result in a 'good' room.

I think one way to approach it is to recognize ahead of time that it will never be perfect everywhere, and then start smallish and add to that until you are willing to live with the result. Fix the bass first. Corners and places where surfaces meet. The wall opposite your monitors. I tend to be technically bent, like you. So, I like the comfort and accuracy of frequency analyzers where you can compare the change when you add something new to just listening. But, at the end of the day, listening is what you're either going to be satisfied with - or else not. So, do a lot of that too.

[jmh]

It is kind of funny that I am pretty tolerant to mediocre audio conditions. But when I was a little kid a song playing on an AM radio - even transistor, could be the best thing in the world. Bass; who needs it?

The other thing is if a great song and performance are captured in non optimal situations whether mechanical or reproducing, the goodness trumps many obstacles and I can love it.

On the other hand, I find bad conditions unbearable - the roar of kids in a HS cafeteria - or a bar with bad acoustics.


Edit:
Yikes! I just had a posting error and lost a few paragraphs where I had a couple of realizations. I guess they were not meant to be at the moment.

[John Ludlow]

A guy that works at Radford University developed a method of building light weight cages for acoustic fiberglass (also known as '703'). He took lots of pictures as he did it and published on a web site at Radford. He told me that a lot of people over the years had seen that site and used it to build their own - and then written him to thank him, as I had. One of those people was our Angie. She documented that somewhat when she built her studio many years ago. I remembered that and contacted her, and she led me to the the original site. This is it:

https://www.radford.edu/~shelm/acous...ass-traps.html

Unfortunately, that page is now blocked. Not gone, apparently, but not available either - at least, not today. I have no idea why after all this time. But I put it here for posterity in case it someday becomes available again.

His basic idea is to cover the 703 with rayon batting using spray glue. Then to produce lightweight cages to put them in out of drywall corner bead and pop rivets (a whole lot of pop rivets...). Then, finally, cover that in burlap, which is acoustically transparent. I bought my 703 here:

https://www.gikacoustics.com/product...raps-supplies/

...because it was the cheapest place I could find at the time. Keep in mind that those are 4'x2'x2". So, if you want 4" (and, for bass, you probably do...) you have to spray glue 2 - 2" sheets together.

At the time, I just couldn't believe how expensive finished bass traps were for what they are. Building these traps seemed easy enough and it's a lot cheaper. However, it turns out that it's much more time intensive than it at least looked like it would be to me. If you decide to build your own - set aside a lot of time for sweat equity. To put it into perspective, every trap has 8 corners; each corner joins 3 surfaces of corner bead; each joint requires 2 pop rivets. Each rivet requires a drill hole. So, 8x3x2 = 48 rivets per trap frame, times 12 traps = 576 rivets and drill holes. It's very time consuming. Those guys earn their money more than I thought they did. Still, if you don't have money to throw at it, and you're handy, this is a way to go. If money is no object, you can save yourself a lot of time and a medium amount of frustration by just buying them professionally finished. In my case, when they became acoustically functional, I decided to rest and finish them later. Still waiting... But they function great.

Something else I found out is that 703 does not produce airborne glass filaments the way that thermal fiberglass does - or at least not nearly as much. In fact, at least one studio that has a video on Youtube got rid of their (heavy wooden) frames altogether and now uses naked 703. I don't really know whether that's dangerous or not and there is an ugly factor. But, at the time, I handled the 703 a lot, and it was summer in my garage, and I never itched at all from handling it. YMMV.

I did the search through past archives for that test I did in 2019 of pink in my 703-treated room after also running it through the Sonarworks Reference 4 VST for my mixing listening position.

https://imgur.com/Bnr8d90

What you're seeing in the three frames is how Reference 4 claims it did on the left, the original pink on the top right via a frequency analyzer (Voxengo Span) and a recording of that same pink using the reference mic (that comes with Ref 4) at my listening position pointed at one monitor on the bottom right. Both frequency analyzed results have been smoothed to a third of an octave. I also added a red line to make the variation easier to see in both of the pink windows (on the right...). Except for that big hump at 38 Hz - that's pretty good, I think. And I routinely low pass that out.

[jmh]

Now to restore my deleted portion...

The problem I've encountered is with a few songs where sustained g# (i.e. 2nd bass string first fret) is inordinately loud. This has been with 2 very good bass players on different basses. These high notes can be knocked down with a multiband compressor. But on the songs where this occurs, I have some muddiness pop in here or there that I have not been able to conquer.

The tracks were probably recorded unamplified with a live drummer in the room, so there was additional energy in the air to set things into motion imposing room resonance on electric basses - which I would barely believe - except for abundant evidence - and now finding several mentions of this on line.

I have had an epiphany in recent days - and that was realizing that the problem would be sporadically introduced by various other tracks that were miced in the room. An overdubed guitar or vocal - you think nothing of a slight bleed from headphones or even moderate speaker monitoring knowing that it will be insignificant to the rest of the mix. But when one frequency rings out 30db louder, it is significant - and to add insult, it isn't in tune with the song - but the room. So you get this detuned interaction between the real bass track and a note that you would never think to be a problem - because the overall track may sound excellent.

The next issue is that I'm set up to mix in this same room, so I may perceive the problem squared as in another environment a slight increase or even few moments of room resonance might be passable - but the room now gets a second shot at amplifying the sound. The ones generated on miced tracks might even pop worse because they are in tune with the room resonance instead of the bass' note that would be a near frequency.

Anyway this has caused endless hours of wasted time - but I guess its okay, I tend to learn the hard way.

I've been doing my own investigation and thinking of building something like this:
Adsorber
I have a few spots where it could be built into existing structure - if any of those spots turns out to be a promising location.

Some of the things I have been reading support your suggestion that diaphragm style traps might be better bought than built.

I have also found that there are several types of natural insulating materials some made from hemp, straw, wood, etc. If I do something home brewed, I may attempt it with one of those materials - Let it rain Sweet Mary Jane...

Sounds like the makings of a hit.

[cgrafx]

1. don’t track live amps when you don’t have to. bass can always be tracked direct. in almost all cercumstances there is nothing that a bass amp adds that can not be done in the box after tracking.

2. use, headphones as a check on the mix, not just the room speakers.

3. definitely worth spending some money on room acoustics.

[jmh]

Philip,

This is what blows me away; it is happening with conditions 1 & 2; unamplified and with headphones.

Overdubing electric guitar where a miced amp is so much louder than the room is the scenario where I tend to go without headphones.

The original bass tracks was often a precision bass (keys & guitar) played silently, direct with the live drummer in the same room. A vocalist in the next room (usually expected to be redone - although with a few keepers - which presented it's own challenges).

[John Ludlow]

John, that's weird stuff. If I'm understanding you correctly, an interesting test might be to record the bass direct (no amp/no open mic) while someone plays drums. And then methodically fret bass notes, but do not pluck them, while deadening in between. If you're right about the drums resonating with the bass, you should be able to see it in the recording. But 30 dB seems like a lot for a solid body instrument, I think...

Still, if that is the problem, then you might be able to solve it just by placing repositionable absorption between them - or even a hard barrier that will bounce some of the low frequencies at a non-right angle to walls. That's all configurable in realtime by recording with a frequency analyzer on while you move the trap/barrier. You'll be able to see the changes on the screen as you move it. If it's the room, you might have to fix the room.

But making the effort to improve the room is overall the best thing you can do generally. It really will make a difference you can hear including things you don't even notice now.

[jmh]

>>John, that's weird stuff. If I'm understanding you correctly, an interesting test might be to record the bass direct (no amp/no open mic) while someone plays drums. And then methodically fret bass notes, but do not pluck them, while deadening in between. If you're right about the drums resonating with the bass, you should be able to see it in the recording. But 30 dB seems like a lot for a solid body instrument, I think...

The impact on the bass is maybe 6dB and it takes a few moments for it to develop. I just mentioned the drummer because it's loud pulses of energy and it may have an impact in setting things into motion. Nevertheless, your experiment is interesting and I may give that a whirl maybe I can enlist my wife to do a drum solo.

The 30dB is the mic listening to the room resonance. Here is a link to a picture of the waveform:

SinSweep_16to300_bottom_listening_mic_top.jpg

I got a volume bump at around 50 - then a giant spike centered ~100 then more mild (and low concern) bumps and dips in volume to 300.

I already know there are several resonant frequencies clustered around 100hz which makes sense since I have some similar dimensions. The room ceiling follows the roof rafters (6 on 12) and flat topped on collar ties (11'10") then there are paired 2x12 (parallel to the ridge) that are offset about a foot into either side the room, so there are a lot of funny corners. It seems it will be a challenge to wrap my head around it - but I'm trying.

If I play a steady 102hz signal at what would be a tame level for playback, in some spots you might feel you need earmuffs and a few steps away, it will be quite mild. I have been using waves emo generator which has 1hz increments - but want to look with more detail.

[John Ludlow]

I can't tell the scale - but it looks very dramatic. Welcome to interesting times.

Actually, from your description I would have thought you'd have an advantage in non-parallel walls. But the graphic says it all.