Recording room initial REW measurements

[garymacf]

Hello. I’m an audio book narrator who has recently moved out of a cramped booth situation and into a a larger recording space, a small building separate from the house. The space originally didn’t sound very good, with a booming, echoey quality. I’ve worked on the acoustics, and recently used REW to measure the room. I’m hoping for some assistance in interpreting the results of the .mdat file I’ve included.

The space is 15 ft. 4” long x 9 ft. 4” wide x 7 ft. 7” high (or 112” x 184” x 91.5”). I’ve got a mixture of 12 16”x48”x3” deep traps filled with Roxul; 5 24”x48” 2-inch deep ATS panels, and two triangular ATS bass traps on the front wall. Three of the 16”x48” traps are hung from the ceiling as a cloud above the mic position. A piece of 6 ft. x 6 ft. carpeting is on the tiled floor.

For testing I used a Dayton EMM-6 with its calibration file, and for an interface I used a Steinberg UR28M for which I produced a loopback calibration file. I did tests at 7 reference mic positions, and the distance from the front wall, side walls, etc. are noted on each test in the REW file. For playing the sweep I used Yamaha HS50M studio monitors on temporary 15” stands above the desk.

The room response while I’m narrating is more important to me than how the room sounds during playback. I’ve read (and re-read) the Room Measuring Primer posted recently by DanDan. That was very helpful. But REW produces a lot of information, and I’m not exactly clear on what I’m looking at. Overall I want to gain a better understanding of how the entire room is behaving, and what I might need to do next.

Thank you in advance for any suggestions for next steps.

 

[DanDan]

Hmmm. Those measurements are of those two combined source positions only. They do not describe the room or the response of a single source, e.g. voice.
A Modal Map of the Room could be very useful. This would tell you where the nulls and peaks and dead bodies are. A voice may benefit from a little help from a low Mode for instance. Or if it is sounding honky, your are probably suffering a higher modal peak. Place the two speakers on the floor in two corners. This will light up all your modes. Place the measurement mic in any opposing corner, up at the ceiling if possible. All of the modes will congregate here.
You should see your modes in the Waterfall plot. A modal calculator will confirm which is which and where you would expect to find the peaks and nulls. https://www.hunecke.de/en/calculators/room-eigenmodes.html The sine generator will confirm and highlight these. Turn on frequency follows cursor and slide onto a Waterfall ridge. The room will take off. Note the pattern in that calculator and walk the room to confirm the peak null pattern. To find the best sounding recording position, place one loudspeaker where you mouth may be, the mic as it would be.

 

[garymacf]

Many thanks for the detailed reply, DanDan. I'll start with making the Modal Map. With the speakers on the floor in the two corners at the front of the room, is it okay that one would be partially blocked by the equipment desk? Also, for testing the recording position, with one speaker placed at the potential mouth position, where would I place the measurement mic in the room?

 

[DanDan]

You are welcome. The speaker locations are not critical, LF waves are very long. They won't take any notice of your desk. Presuming you want to know the actual recording response the microphone would face the speaker. As it would your mouth.

 

[garymacf]

I took another whack at the recording room measurements following DanDan's suggestions. I've attached two .mdat files. The first file is testing for the room modes. The monitors are on the floor in the right-hand and left-hand front corners of the room; the Dayton EMM-6 reference mic is located at the rear RH and then LH corners of the room at the ceiling position (approx. 2 " from the ceiling); the mic is pointing straight up.

The second .mdat file was made with the EMM-6 positioned where my recording mic is normally positioned, approximately 38" from the front wall, and 26" below the cloud panels above the recording position. I held the monitor at my mouth position; the monitor was pointed at the EMM-6 capsule and was about 8 inches away from the mic.

My next step is to investigate the modal calculator DanDan referred to in an earlier post. Thanks again for any suggestions on additional room treatment.

 

[Bruce Black]

First, look at your decay plots. Unfill them so they show as traces, and set the slice interval at 60 ms. Leave the initial 0ms./direct sound trace, and then add each other trace one at a time. This will show you what your room is doing to the sound over the course of ~1/2 second. This will reveal how the sound changes over time, and may reveal things you can't find with an RTA or similar.

Next, look at the waterfall. You'll see that there are some "ridges" that run down the face of it. These are your room modes, what I call resonances. If you build some Helmholtz resonators at these frequencies, you'll smooth out your low end. I've attached a PDF of a talk I've given at several AES conventions, plus a design for a simple, easy to build Helmholtz resonator.

The PDF also explains why bass traps are pretty much a waste of time.

I also see some significant low frequency dips. These may be caused by phantom resonators, so make sure all your doors, windows, and other room penetrations and holes are well sealed up.

 

[DanDan]

Interesting. Your corner modes tests are quite different from each other. I am glad you went the extra mile, tried both. You could try to see if there is anything in common sticking out in the Waterfalls. On a quick look I see 60Hz in the Waterfalls, and 125Hz in the RT60. But overall nothing troubles me here. Nice tight recording room. I have quite a collection here of Altiverb IRs of world famous Pro Recording Rooms. You would be surprised how wild many of them are. The vocal response seems a bit hot around 1K. Worth experimenting with different locations.
Bruce, I can't really let that statement stand. Bass Traps are almost invariably beneficial. One of the world's best acousticians, sadly RIP exclusively used 2' fibre behind his diffusive surfaces. That takes space but so does any LF treatment, except perhaps VPRs and the BBC whole wall version of that. It is rarely stated how much area of resonant is required and there seems to be an implicit suggest of 'not much compared to other treatments'. e.g. The miracle cure caused by one Tub Trap in the Master Handbook of Acoustics. This is utterly faked. Couldn't happen.
Sound waves travel by well established Wave Motion. I think it erroneous to suggest that this changes with frequency. There are other reasons why HF exhibits more specular reflection.

 

[Bruce Black]

I have the deepest respect for Everest, however I've found a few errors, of a sort, in his book ( I find fourth edition to be the most comprehensive. But I've found that the best LF response comes from treating the peaks and leaving the rest alone. With bass traps, you pull everything down, the good stuff with the bad. Resonators tuned to the peaks just bring down the problems.

Everest also refers to flexing walls as providing some much needed LF absorption. The problem is that while they absorb some lows, it's nearly impossible to predict what it will be. Also, it takes time for the acoustic energy to overcome inertia, so the absorption is delayed, usually around 120 ms. in my testing. This means that absorption won't really show up in an RTA (most often the only piece of acoustical test equipment in nearly every studio), so unless you have another piece of test gear that also shows TIME (like the decay plots in REW), you won't see this and be able to deal with it.

Anyway, we can agree to disagree on the pressure/wave business. The pressure/particle motion stuff is actually out of Everest's book. In my many years of consulting and designing for some major league players, treating for pressure below 300 Hz. always works, and quite well. Treating for particle motion in the low end with resistive absorbers just doesn't cut it.

 

[DanDan]

I am not sure what you understand by the term Bass Traps really. Let us assume for now, fibre trapping though. The benefit of broadband absorption is that it addresses nulls, be they SBIR or resonant, as well as the peaks. Plasterboard absorption has been well documented. This concept of acoustic inertia seems as unique as your (mis) take on the Wave formula. Please do not try to explain RTA etc. etc. by your interpretations. There are agreed understandings amongst people who take this seriously and are qualified to do so. Resonators are notoriously difficult to get right. They need to resonate but not too much. Furthermore predicted tuning does not happen in practice. Room load, particularly at boundaries and corners, changes the resonant frequency. Again you accuse other professionals who do not use your HH techniques of 'not cutting it'. That sort of self aggrandising by uttering lies will not be tolerated here.
Here is an example of a real studio treated using fibrous absorption only for LF.

 

[garymacf]

Bruce Black, thanks for your input. I’ve taken a quick spin through the .pdf you sent, and the diagram of the Heimholz resonator you provided looks like something even I could build with my limited shop skills.

I have to admit you’re the first person I’ve run across in the studio design world who doesn’t recommend bass traps as initial treatment for a room. The people I’ve been following who design voiceover spaces (George Whittam, Tim Tippets, John Brandt, and others) all make use of acoustic panels, although they might differ on the actual construction (703, soft insulation, rockwool, etc.). But I intend to go through your .pdf and get a better handle on your approach.

Also, thanks for the suggestions on looking at the decay pots and resetting the slice interval. I’m still a newbie with REW, and I appreciate every hint I can get on how to best use this powerful program.

Finally, you are correct that the low frequency dips in my initial plots might be caused by windows. The building I’m using has four small windows. I used makeshift covers (blankets) on them when I did my initial tests, but I didn’t really expect those to stop much sound bouncing around. I’m in the process of building removable window plugs, using a design suggested by John Brandt in his Facebook group devoted to voiceover and booth building. I’m building frames of 3/8” MDF and filling them with two layers of 5/8” drywall, then one layer of R-13 glasswool. It will be interesting to see if this affects those low frequency dips you mentioned.

 

[garymacf]

DanDan, thanks for taking a look at my initial plots. I will give the Waterfalls a closer look. As I mentioned in my post to Bruce Black, I’m in the process of building window plugs which may have an effect on the low frequencies.

I suspect the reason my corner modes are quite different is that in the LH rear corner of the room, I have a nearly floor to ceiling bookshelf. I’ve arranged the books on the shelf in varying heights, and even turned quite a few of the books sideways, in an attempt at a poor man’s reflector to keep some life in the room after having added all of the acoustic panels. (I picked that idea up from a studio designer doing a YouTube video, but I’ll be darned if I can find the video.)

I just started recording in the room last week, and thus far I’m pretty happy with the sound. But I will try some different recording positions and see if I can deal with the vocal response around 1K.

Thanks again for your help.

 

[DanDan]

You are welcome. Windows allow LF to pass through. It may be counterintuitive but this is often very desirable. Nulls are caused by destructive reflections. At LF these reflections are caused by solid massive sealed structures. A concrete bunker being the worst. Conversely a wooden shed with doors and windows will tend to have great even bass response without long modes are nulls. Newell refers to such leaky light boundaries spaces as 'Limp Bag' This is why the sound in cars etc. is often remarkably good. The thin steel and glass are great bass traps. No reflections, no nulls no booms.
Books or indeed any irregular shapes are bound to break up standing waves and specular reflections to some extent. But if you really want some lively areas in the room for drums, strings, winds, there is a simple cheap way to go. Plywood sheets angled against the wall. Take a look at johnlsayers wall traps and scatterers.

 

[gullfo]

couple of thoughts if not too late - use thin curtains on windows if you want to retain the light + reducing slap echoes. here's a mode map and "safe" positions. consider that a room can be too dead - so slotted panels of thin ply can reinforce the higher frequencies you need to make the room comfortable and still remove unwanted LF and low-mid build ups.

 

[DanDan]

He gullfo, how nice to see you again. Yup ply, can be angled too.

 

[mike w]

Could someone explain the purpose of looking at the REW decay plots at 60 ms intervals as Bruce Black suggested? Yes, there is a difference in decay WRT time but what does that mean and how do you act on that info?

 

[DanDan]

No idea, but Reverb measurements, even when hardware based, have a ' quality factor' If the decay is uneven, it suggests interference from Modes or even a secondary decay, as in an adjacent room.
I always have two catch phrases at the back of my mind Maximum Bass at All Frequencies, Jay Bless and seasonally.... Deep and Crisp and Even.