Compact VPR experiment - looking for feedback on test setup

[refractioncat]

I'm planning to build some VPR/EPA absorbers in an attempt to improve seat-to-seat consistency in my living room in the 57-300Hz range, and I have an opportunity to do some comparisons along the lines of the Trapping Traps experiment. I'd like to get measurement data out of it that is as high-quality as possible.

I have 6 panels (120x60cm x 50mm) of Caruso Iso-Bond WLG035, 3 steel sheets of different thicknesses (100x50cm x 0.6, 0.8 & 1.0mm), and a few other supplies like some melamine foam and PET house insulation (very similar to fluffly fiberglass but less irritating).

The room I have available for this experiment is roughly 350cm long, 320cm wide, and 240cm high. From the door, the left side has a closet and the right side has a large window. There is some furniture present that I cannot realistically remove from the room.

For the hardware I'd be using a Wharfedale Diamond 12.1 speaker, Fosi ZA3 amplifier, and UMIK-1 microphone, all connected to a laptop. This means that realistically I'm limited to frequencies above 50 Hz, but this is fine because the range I'm mostly interested in is 57 Hz upwards.

I am considering two different setups for the experiment:

In the modal setup, I would be essentially copying Trapping Traps with a worse room with a fair amount of pre-existing absorption, and smaller absorbers (0.7m^2 instead of 2m^2).

The speaker would be on the floor level at position S. The microphone would be near the ceiling at position M. The absorbers could be placed straddling a corner at A1, flat in a corner at A2, or flat away from the corner at A3 or A4. The room would be measured empty, and with each absorber configuration, and the decay times can be evaluated, insofar as any meaningful difference can be detected. With some configurations I'd be able to use absorbers stacked or side-by-side to increase the area to 1.4m^2 or 2.1m^2.

Because of the issues with the modal setup, I've thought of an alternative using SBIR to isolate frequencies of interest. The speaker would be placed along the wall at varying distances in area S, the absorber would be at A to treat the null caused from the wall reflection, and the microphone would be at either M1 or M2. This requires separate measurements for each distance with each absorber (and the control condition), and then the frequency of the nulls and their depth could be evaluated. The frequency of interest can be controlled fairly precisely by the distance from the wall, and apart from the frequencies where the ceiling/sidewall reflection would interfere, it seems like at least the relative performance differences between absorbers could give some useful information.

My plan would be to mark the test positions with tape on the baseboard, and move the speaker between positions while keeping the absorber setup constant for each absorber. To check the reliability of this process, I'd take multiple control measurements where I move the speaker between the different positions to see how consistent the cancellation frequencies and depths are between different attempts to place the speaker in the exact same positions. The microphone would stay stationary through the measurement session; if I have to move the microphone, or it moves by accident, I'd treat that as a separate session, retake the control measurements, and distinguish between the sessions in the results.

The M2 mic position is more convenient and less likely to have to move between measurements. The M1 position seems slightly better geometrically, and allows moving the microphone outside the room if a greater speaker-mic distance is needed for lower frequencies and the reflections from the hallway don't make things too messy.

I can address the sidewall reflection somewhat by placing something absorptive in the way, but I don't have a practical way to deal with the ceiling reflection meaning that around 50ish Hz (if I calculated it right) would be inevitably confounded.

---

The absorber configurations of interest are:

Pure porous absorbers:
Iso-Bond (120x60cm): 5, 10, 15, 20, 25, 30cm
Melamine foam (100x50cm): 5, 10, 15, 20cm

VPRs (with Iso-Bond as the damping material, each with 0.6, 0.8 & 1.0mm plates):
5cm + steel
10cm + steel
15cm + steel

EPAs (steel of each thickness sandwiched between Iso-Bond):
5cm + steel + 5cm
10cm + steel + 5cm
5cm + steel + 10cm

In addition, I'm interested in trying out a couple of experimental solutions to see whether they work:

Bean bag chair bass trap:
Around 100x100x80cm, filled with polyester; turned out to be a lousy chair but maybe it could be an aesthetically inconspicuous superchunk.

Multi-plate VPR/EPA:
Stacking the thinner steel plates to see whether they can achieve a better performance in low frequencies due to having the pistonic effect of a thicker plate but the lower bending stiffness of a thinner one. I can use a very large mousepad (roughly 40x90cm) to decouple the plates a bit if needed.

Edge-on absorption:
Placing Iso-Bond absorbers on the wall in line with the sound waves, instead of face-on, to see what happens; or in other words, 120x10cm and 120x20cm panels with 60cm thickness and different airflow resistance behaviours due to the material's anisotropy.

---

To get the necessary disclaimer out of the way, I've read+watched my Sauro and understand that this is very much a "I changed something and then something happened" experiment. The 120x60cm VPR is a particularly convenient size for people to make because, at least in the UK, 100x50cm steels are easily available on e.g. ebay, and I haven't found much data beyond "these are the before and after measurements from my room", so I figured I'd try contributing something back to the commons.

I'd appreciate any feedback/criticism about the experiment design and setup, and opinions on whether I should go with the modal setup or the SBIR setup, or even both. Any ideas on additional tests that you'd like to see performed, that can be performed with the supplies, gear and space I have, are also welcome.

 

[refractioncat]

Update: I found that the SBIR setup was unable to isolate the reflection frequencies the way I wanted, with the strong nulls remaining constant regardless of speaker positioning, and the variable null being rather dirty and weak.

In the modal setup, the decay times in the room, even stripped of as many objects as realistically possible, were still quite low and even using 3 panels simultaneously I could not get clear differences. The closets, window and creaky lightweight floor provided too much low-frequency absorption compared to the absorbers.

One thing I did discover was that with one side of the (sliding) closet door open, the room had one of the cleanest low-frequency response curves I've seen in my own measurements.

As a result, I'm probably going to just do what everyone else does, and test absorbers in-situ and give before-and-after measurements showing that X treatment had Y effect in Z position.

 

[kemmler3D]

Sorry I missed this thread when it was first
Posted, but I'll be following with interest. I haven't built any but I've been interested in VPR style traps ever since I came across some threads on gear space.

My main thoughts reading this are 1) I'm not sure if you will see the most benefit if the frequency range is >50hz. I think they can be effective in that range but IMO their real edge is in the ability to go lower. 2) I'm not sure if they are big enough? In the other before/ after tests I've seen they used steel sheets closer to 240 x 120cm. Maybe smaller is better for higher frequency absorption though?

That said, if you have the time to test multiple configurations of foam / steel it would be totally new info to me and as far as I know the internet as a whole. So thanks for kicking this off!

 

[refractioncat]

Sorry I missed this thread when it was first
Posted, but I'll be following with interest. I haven't built any but I've been interested in VPR style traps ever since I came across some threads on gear space.

My main thoughts reading this are 1) I'm not sure if you will see the most benefit if the frequency range is >50hz. I think they can be effective in that range but IMO their real edge is in the ability to go lower.

The reason for resorting to VPR-style absorbers is that I need them to be very compact. I'm ultimately trying to absorb some SBIR/early reflection nulls in my living room:

The room is roughly 470 x 300 x 240cm, with masonry walls. The windows have thick curtains (the right side curtain is partially double-layered for extra absorption) roughly 20 cm away from the glass. As a multi-use room the freedom in positioning is very limited, and the treatments also must be compact enough not to interfere with other uses. The convenient position for the speakers has them exactly 38% of the room's width from the sidewalls, forming a narrower triangle than the usually recommended 60 degrees but REW room simulator suggests that this is a surprisingly decent positioning insofar as frequency response is concerned.

The speakers and listeners are placed as close to their respective walls as possible, to push the wall reflections into frequencies high enough to treat easily (around 150-300 Hz). In addition to the obvious cancellations from the front and rear walls (and ceiling, which I'd ideally do something with), there is the 57 Hz axial mode which takes heavy EQ to get under control.

The absorbers in positions #1 and #2 have 10 cm of space available, and it would be nice but not absolutely necessary if they could have some effect down to 57 Hz. According to Sauro's measurements OC 706 seems like it would be optimal, staying effective down to 80 Hz, but I don't have a good place to work with fiberglass, and 706 is not really available in Europe anyway. Everything else seems to taper down in effectiveness somewhere in the 100-200 Hz range when used in 10cm thickness, which I'm hoping to extend lower with the steel.

The absorber in position #3, to bring some symmetry to the side reflections in the low frequencies, can only be 5 cm thick before it gets in the way of the desk. The left side corners also have room for bass trapping if needed to balance out the right side (the door is always open, and the curtains+window+gap seem decently absorptive even in the lows).

At certain listening positions a 80 Hz null shows up, and the side walls are the best geometric candidates for its cause. I know of nothing porous that can deal with 80 Hz at 5cm thickness, so a VPR is the only real candidate. Even if I can't affect the null itself, cleaning up the side reflections from the low frequencies while having a reflective surface for higher frequencies seems possibly optimal; I suspect that the comb filtering from side reflections isn't that bad in wavelengths short enough to present a meaningful phase difference to different ears, and the difference in distance+speaker directivity mean those reflections start around -12 dB even without absorption. If a 5cm thick VPR is too thin to do anything useful, I'd probably rather leave that spot bare than lose the high reflections with a 5cm porous absorber while retaining the low cancellation (I just don't understand why 5cm for first reflections is so popular).

2) I'm not sure if they are big enough? In the other before/ after tests I've seen they used steel sheets closer to 240 x 120cm. Maybe smaller is better for higher frequency absorption though?

That said, if you have the time to test multiple configurations of foam / steel it would be totally new info to me and as far as I know the internet as a whole. So thanks for kicking this off!

The VPR pistonic frequency is theoretically size-agnostic, while the equation for bending mode frequency contains the square root of (thickness^3 divided by areal density, i.e. thickness^ 3 / (thickness * density)), which simplifies down to just thickness; or in other words, a panel half the thickness has the bending modes of one with dimensions twice as long. Thus a 100x50cm x 0.6mm panel should, in theory, have bending modes similar to those of a 167x83cm x 1mm panel. Since I'm trying to address higher frequencies to begin with, this seems okay. Also, having the steel surface out helps avoid excessive absorption in the highs, as the room is already pretty damped in that range with carpet and curtains.

Given that my test room turned out to be excessively structurally absorbent, I'll have to experiment in situ in the living room. I'll report back what I can on the design variations and their relative effectiveness.

 

[kemmler3D]

Thanks for the detailed reply, makes total sense.

I guess my next question is whether you've looked into BAD panels for this? From what I've seen they can supposedly be effective a bit below 100hz even at moderate thickness... for reasons I don't claim to understand it seems the low frequency absorption goes up a little when the diffuser front is added to absorptive materials.

 

[refractioncat]

The effect of the 1" BAD vs. 1" porous seems similar to the GIK range limiter (which itself is, iirc, a thin piece of HDF or similar between the fabric and absorber) and I wouldn't be surprised if the enhanced absorption of the BAD, range limiter and VPRs all operated on somewhat similar principles (as they all have a more or less non-porous plate over semi-rigid porous absorber material). However, compared to the 4" Absorbor panel, the 4" BAD seems to perform slightly better at 160 Hz and slightly worse at 100 Hz according to RPG's graphs, and the 4" Broadsorbor eats both their lunches.

I'm also not convinced binary amplitude diffusion is what I'd want for the surfaces I'm treating, especially right behind the speakers and listeners.

 

[refractioncat]

I guess "diagnosis before treatment" wins again. I was not expecting the room, without any treatment, to look like this:

I thought masonry walls would have terrible boominess but this is remarkably well-behaved. However, the immediate reflections from the front and rear wall, plus a ferocious flutter echo between them, still necessitate treatment:

Placing a 5cm + 1.0mm + 5cm EPA behind the listener, a 5cm + 0.8mm + 5cm EPA behind the right speaker, and a 5cm+ 0.6mm VPR on the side wall along with 5cm CIB behind the left speaker (thickness and orientation limited by window+curtain) seemed to be the most effective combination of the ones I had time to test. The spectrogram shows some activity down to ~72 Hz but the effects are quite subtle, apart from the unfortunate diminishment of higher frequencies as an unavoidable side effect of the elimination of the flutter echo and reflections:

The frequency response is now relatively dominated by the ceiling and floor reflections where the speaker's directivity error influences the timbre in the 1k-4k range:

It's really hard to distinguish the effects of the different versions of the panels, and I'll need to do another run of measurements to try to tease some actual effect out of the data now that I have a bit of an idea how REW works. And figuring out if there's something useful I could do with the modes without totally killing the decay times at higher frequencies. The VPRs and EPAs are doing something in the lows, especially when straddling corners, but unfortunately straddling corners tends to make the frequency response worse by eliminating reflections that are useful for filling in dips from elsewhere.

 

[refractioncat]

I have some test results showing isolated effects. These were performed with a microphone in the same position, 50mm+1.0mm+50mm EPA (E10 in measurement notes) behind the listener, 50mm+0.8mm+50mm EPA (E8) behind the right speaker, and using the remaining materials to test varying locations for either a 50mm+0.6mm+50mm EPA (E6) or a 100mm+0.6mm VPR.

It's hard to see much in the waterfall/spectrogram view, but from SPL tabs one can observe changes in the frequency response of up to 1 dB at 58 Hz and 3 dB at 72 Hz from a single 120cm*60cm panel. The 100mm VPR seems a bit more effective in the lowest frequencies than the 50+50mm EPA. Both are significantly more effective when straddling a corner, even with the space behind them entirely open above:

(red: control, blue: VPR straddling)

Overall this seems to suggest that even the compact VPR design works to some degree, but a larger number of panels is needed to treat the room effectively. A lot of the time an individual change with a single panel would improve something while making something else worse. The 100mm+0.6mm VPR is the lightest, still has some effect under 100 Hz, and provides a reflective surface for higher frequencies so it seems viable if I wanted to make significantly more panels to try to treat the room more comprehensively. Unfortunately I had sealed the 1.0mm and 0.8mm steels inside EPAs by this point, so no comparative tests of steel thicknesses for 100mm VPRs were made. In the end I put the 0.8mm behind the listeners, the 0.6mm behind the right speaker, and the 1.0mm in the left rear corner under the desk, along with a melamine foam box with PET insulation stuffing.

Another discovery I made was that placing a thick melamine foam absorber on the left wall (1m*1m size, roughly 30cm thick of loosely stacked panels leaning against the wall) made a significant effect on the 58 Hz front-to-back mode, reducing it by almost 2 dB:

(red: control, green: side wall absorber, mic position in these measurements is different)

 

[refractioncat]

Also, in terms of DIY friendliness this is one of the easiest and most benign combinations I can think of:

The Iso-Bond is extremely pleasant to handle, and looks fine even without a covering. It doesn't need cutting in this design, but if you do cut it you only get some polyester fibres which aren't that different from what synthetic clothes and furnishings produce anyway. In comparison, melamine foam has a bit of an unpleasant feel on bare skin and creates a dust when cutting.

Double-sided tape can work for adhering the Iso-Bond sheets to each other and the steel (I recommend using a large quantity of wide tape, covering a good fraction of the area to compensate for the inability to adhere the tape optimally), so VOC-emitting glues are not necessary either, depending on how you mount the panels (I used some glue in mine). In Finland a material similar to Iso-Bond has the highest M1 emission class for building materials, meaning that even sensitive people should be unlikely to have issues with it.

Wrapping the panels in fabric all around the edges (according to Camira's specifications, the Era 170 seems best for the purpose as it improves low frequency sound absorption of 100mm melamine foam in impedance tube tests and is slightly reflective in high frequencies; in comparison the popular Cara fabric actually performs fairly poorly in Camira's tests, and the highest performers are wool fabrics that cost a lot more than Era 170, not to be confused with the regular Era fabric) would improve safety by ensuring that if the steel's adhesive fails, the fabric can intercept it before it falls on something important or expensive. Attaching the fabric to the absorber is left as an exercise to the reader; I'd either sew the fabric into a bag shape, or simply wrap it around and glue it on the backside.

The steels tend to come curved from the roll, requiring some bending to straighten them. Even 1.0mm is flexible enough to be bendable by hand (I recommend wearing gloves to protect from edges/burrs). I found it useful to stand the panel on its narrow edge, and try to flex it from side to side. If the panel was bent, it would have a "dead spot" of sorts in the centre where it would take some force to dislodge it across the centreline and get it flexing (similar to how measuring tapes work). If the panel was straight enough, there would be little to no resistance to flexing from one side to the other. The sound when tapping the steel also helps; if it's like "pinnngg" or "bonnngg" it's not good, but if it's like "buwuwuwuwuwnnngg" it's better.

 

[refractioncat]

I dug up some other measurements where absorbers can be compared against each other, from an empty room with various types behind the listener.

(Red is always the control curve)

100mm CIB

0.6mm and 0.8mm EPAs; there is very little difference between them.

1.0mm EPA; this one has a noticeably different curve for some reason.

This is a VPR with 0.8mm steel, but unfortunately I'm not 100% sure whether it was 50mm or 100mm behind it.

It's a bit hard to tell what's going on here, and whether the metal-containing ones actually outperform just plain 100mm Iso-Bond.

 

[tehas]

hi @refractioncat - thanks for running these experiments ! apologies if I missed it, what does the EPA acronym stand for and is the only difference from a VPR the 'sandwiching' of the steel plate in between CIB vs the steel plate exposed in the VPR ?

Where are you located? if you are located in the USA, where did you source the CIB - don-audio or elsewhere?

 

[refractioncat]

The EPA refers to Enhanced Porous Absorber from Gearspace, and yes, the only difference is that the steel is sandwiched from both sides.

I'm in Britain, using mostly Thomann, ebay and B&Q (similar to Home Depot) for the materials.

Insofar as sourcing materials in the US is concerned, I think the basic design has a fair bit of flexibility and wouldn't discount options like polyester mattress toppers (some 4" versions of which were tested by NWAA to have some membrane activity down at 40 Hz even without any steel), polyurethane foams or even rockwool/fiberglass (which is used in some commercially sold options nowadays).

 

[Curvature]

Just looking at the first post, I'd say you don't have enough panels to make any real difference.

Conventional rule of thumb is 25-50% of room boundary surface area should be covered for decent bass effects to show up.

It's something that makes bass treatment impractical and expensive for DIY. Time is money too, IMO.

RPG BAD panels have already been mentioned, but I think the right product would be the Modex Plate: https://www.rpgeurope.com/products/product/modexplate.html

Artnovion Acoustics is also worth considering.

DIY is hard for a lot of reasons, including tuning panels.

 

[refractioncat]

I know I don't have enough to actually treat the room, which is why I was looking for ways to isolate the effects of single panels so that I would not need to make a large number of test samples just to determine whether they work in the first place. It is true that a large number is needed to get significant effects on bass frequencies, but if a single panel can have a measurable effect that is a fraction of a larger quantity's effects that's already useful information.

Also, this particular panel type is supposed to be highly broadband and not require tuning in the first place. The RPG modus plates are over 2x as large and weigh 3-5 times as much as the heaviest DIY version, making them significantly more difficult to handle and mount on walls without leaving marks. A single one also costs 2x as much as all 3 of my DIY versions combined. Artnovion doesn't seem to even have pricing in public anywhere. The time expenditure is less relevant when I'm doing it for the joy of DIY and research.

 

[kemmler3D]

Just want to say, I haven't had time to respond properly to all of the info in this post, but really appreciate you sharing your plans and test results. Of course it's not enough to fully treat the room, but as you say, it's informative anyway, and this is something I've wanted to see more of on the web. Not too many VPR threads out there in the first place, let alone with multiple sets of good measurements showing effects.

One thing I was curious about was your use of tape vs. glue. My understanding is that the glue is supposedly playing a role in damping which is why many people use special types of adhesive for this. On the other hand, I have also wondered why the glue matters if you've attached the thing to a big sheet of foam. I've seen people use green glue (which does have damping properties) and Super 77 spray (which doesn't) so maybe choice of adhesive can be one of convenience?

 

[Curvature]

I know I don't have enough to actually treat the room, which is why I was looking for ways to isolate the effects of single panels so that I would not need to make a large number of test samples just to determine whether they work in the first place. It is true that a large number is needed to get significant effects on bass frequencies, but if a single panel can have a measurable effect that is a fraction of a larger quantity's effects that's already useful information.

Also, this particular panel type is supposed to be highly broadband and not require tuning in the first place. The RPG modus plates are over 2x as large and weigh 3-5 times as much as the heaviest DIY version, making them significantly more difficult to handle and mount on walls without leaving marks. A single one also costs 2x as much as all 3 of my DIY versions combined. Artnovion doesn't seem to even have pricing in public anywhere. The time expenditure is less relevant when I'm doing it for the joy of DIY and research.

Standard method outside of a controlled room is to put the microphone and panel in opposite corners, maximizing the effects.

Good luck.

 

[refractioncat]

Just want to say, I haven't had time to respond properly to all of the info in this post, but really appreciate you sharing your plans and test results. Of course it's not enough to fully treat the room, but as you say, it's informative anyway, and this is something I've wanted to see more of on the web. Not too many VPR threads out there in the first place, let alone with multiple sets of good measurements showing effects.

One thing I was curious about was your use of tape vs. glue. My understanding is that the glue is supposedly playing a role in damping which is why many people use special types of adhesive for this. On the other hand, I have also wondered why the glue matters if you've attached the thing to a big sheet of foam. I've seen people use green glue (which does have damping properties) and Super 77 spray (which doesn't) so maybe choice of adhesive can be one of convenience?

People have had success with and without glue. I've seen a teardown of a commercial model where the front absorber sheet was held in place only by a couple of strips of adhesive tape, and another showing dabs of glue only in the corners of the panel. Some of the best-measured implementations were done by just sandwiching loose sheets of steel between loose panels of absorber.

I used 50mm wide heavy-duty tape in a double-square pattern, like the number 8 on a 7-segment display, with additional dabs of glue in the corners of the steel on the side that is facing the wall. In hindsight I might use a more dense pattern of tape just to ensure that the absorber and steel can't detach from each other in the midpoints forming thin hollow cavities inside the panels. I chose tape because I didn't want to bother with the fumes and mess of spray glue, and the local B&Q didn't have the Sikaflex glue people usually use. So it was either using the less flexible glue from Thomann that I ordered with the panels, experiment with different glues to find out what works, or simply get the tape. I think the tape probably ends up behaving somewhat similarly to spray glue; my main motivation was ensuring that I don't end up with harder lines of glue covering significant areas of the steel and possibly interfering with its bending modes.

 

[cooldude]

I am looking to build something like this. What can i use with steel plate? would fibreglass panel work?
I am in USA .Trying to see what i can source

 

[kemmler3D]

I am looking to build something like this. What can i use with steel plate? would fibreglass panel work?
I am in USA .Trying to see what i can source

It's meant to use foam, the plate needs to be able to flex somewhat freely because that's how it absorbs the sound, the foam damps the vibration in the steel. I think a cheap mattress topper might be ok in a pinch but that's a total guess on my part. Fiberglass is probably too stiff.

 

[refractioncat]

Afaik some commercial products use fiberglass. The dense iso-bond is noticeably stiffer than melamine foam, which is already quite rigid for its density, and I could see the right density of semi-rigid fiberglass being comparable. I haven't handled semi-rigid fiberglass myself, so I can't give exact numerical recommendations and this is a bit of a guess, but if the fiberglass is strong enough to hold the steel, yet flexible enough that it doesn't feel hard, and requires moderate force to compress (if I press on an absorber with one finger, I need around 1 lb of force to get around 1/4" of indentation), it's probably good. Many forms of rockwool seem to have the right characteristics too. A suitable mattress topper would probably work as well. The design seems generally quite forgiving of exact material variations; efficiency can be somewhat affected, but a lot of materials seem to work reasonably well.

I would recommend going for the largest steel size that's practical to source, handle and mount. A large steel with a sub-optimal absorber is probably better than a small steel with the perfect absorber.