Optimizing Small Room Acoustics 

Also how to stop mechanical room noise. How to seal the doors and so on.

Yes that too would be very helpful Matthew. I am about to start work on building a dedicated theater room and this is one thing i am interested in knowing more about. Thinking about doing a staggered stud build. And also more about the acoustic treatment types and placement.

Yes. I finally have a man cave but it's made of bricks wall with cathedral wood style ceiling. I can hear my neighbors talking and dogs barking so clearly. Any suggestion would be appreciated ☺️

Yes, that would be great. Please include information for us people with smaller UK size rooms, eg 15'x10'.

@@guppy9713 good point. I dont think many people realize rooms in other countries vary from what a small us home is. Principle still apply Id think

Yes a soundoroofing video would be great. Like my theater is upstairs ground floor. It has two windows at one of its corners. Also my neighbors driveway is on that front wall. I like them and there blind so Id be interested in a soundproofing vid and options for walls that have things like windows and such. Ive bern looking at double drywalling the wall part with greenglue between as a option. There is also soundproof drywall from a few companies which I think is drywall with the greenglue already added but it costs like 6 timed a regular drywall sheet. I would like to stop ad much of the ht noise from reaching my neighbors and the livingroom right next to the theater. I look at some of these soundproofing web sites and they are selling solid doors for above 2500! There's got to br some reasonable priced solutions for those who wish to keep there home theater sound in it.

It's a lot more popular than I anticipated! I'm glad people liked it.

@@PoesAcoustics Matt I know treating 100% of the front wall is recommended for home theater use but how does it affect your 2 channel listening if you even do any more 2 channel with the Auromatic upmixer becoming more popular? I have to sit against my back wall in my setup so the advice for treatment was a help in that respect. Should I just build a rock wool or fiberglass absorber and leave the cloth off of it or is going with someone like GIK Acoustics along with their support a better option? I have 2 closet doors on either side so I'd probably only be able to add a single panel directly behind my head. TIA.

Peter does it feel like you’re in space when you clap?

That for a lack of a better word sounds amazing!

Clap Ecco is pretty non existent in most places of the room. The 2.4m wall of Quadratic Defusers on rear wall at head level when seated to make the room feel larger than it is was a great addition.

Your setup is not $1K if you have spent all that money treating the room.

Brent Povey I spent around $350 on mine and did it all myself, it is not expensive if you make panels yourself. Each panel cost me $30, 4” acoustic wool, mdf wood for frame and coloured hessian material to cover front and sides. Acoustic treatment does NOT have to be expensive. It’s a very cheap audio improvement technique that most people don’t do and would rather get way more expensive speakers. If you room isn’t treated you will never get the benefit those speakers could give you.

Size will not matter in a small room (quality of the speaker will). See how much space you have for the number of speakers you want (2.1 vs 5.1 vs 7.1). Figure our your budget and where you would place them (before buying). How many bare walls do you have (this is bad for reflections), how much furniture? (more stuff the better for absorption). Problems come up when you have bare walls and floors, if you clap and hear and echo, there is your first red flag.

Not according to this video! They're both basically saying don't bother with any acoustic absorption on a small room because the reverb time is already small. Thus conflicts with pretty much everybody else who say you need more bass absorption to tame a small room! Confusing or what? ☹️

Some people like that. I really think people should do what they like over all else. There are technical problems with the approach, but it doesn't mean you shouldn't do it. If you like it, do it. The technical problems may not be problems either, depending on the room and speakers.

Yep, and the way it calibrates bass is incredible.

Where's Hugo at these days?

@@GothamsFinest2010 Gene needed the protien, so he ate Hugo! Just a rumor, but it could be true :p

If that means having your mains more spread out on the longest wall you should be very happy with the results.

Thx i will have Matt look into this if he hasn't already.

Yes I'm familiar, this is an old article. We will cover this in more detail and have Sean comment as well. Maybe we can get him to join us! A few points of clarity, this study doesn't show that fully bandwidth automatic room correction like Audyssey or Dirac is good or better than only correcting at low and mid frequencies. All they reveal is that Harman's own method of correction is best, which makes use of knowledge of the speakers anechoic response in its correction. It was designed around JBL Synthesis speakers, which have a flat smooth DI. All of this is critical to room correction working correctly at mid and high frequencies. Dirac was not tested in that study, Audyssey was and is widely believed to be one of the worst performing. Sean hasn't publicly confirmed that, as far as I know, but we can ask. I intentionally simplified the concept of the transition frequency, but since it is being questioned, I need to clarify. First, nobody should or is questioning the idea that such a transition exists in a room. It not only can be readily measured in any room of any size, but anyone, without much training, can spot the transition. What is questionable is the notion that it is a hard transition (but this isn't a small vs large room, just a problem with the original theory). Schroeder chose a constant in the formula that artificially picked the transition point, other constants could have been chosen based on different assumptions and this would shift the transition frequencies slightly. What does this tell us, its not a hard transition. What happens is that the response gradually shifts from one of clear distinct modal interactions to a point where that isn't true, where we can refer to it as Stochastic. What that means is that between the start of the transition point and some high frequency point of clear stochastic behavior, it seems to be both. What that means is that the application of EQ above the transition frequency can be ok as long as it is done carefully. Generally lower Q being the safest choices. You can make more robust EQ choices when you have the knowledge of the speakers actual free space polar response. This can't be easily obtained in room, though some correction systems, notably Dirac, have a sophisticated method of measuring that can do a good job obtaining this, as long as the measurements are taken correctly. As for the "dynamic" nature of the target curve that Dirac uses, your description is not exactly correct. I wish what you said was true, but Dirac has a user adjustable correction curve, but the ends are automatically set based on the maximum badnwidth. In other words, it looks where the response naturally rolls off and then tries to limit the amount of correction. The overall shape is still static, picked by whatever curve you set. As for what Dirac would do with regard to the bass tilt, a couple important things. First, As far as I know, Toole and Olive haven't discussed this publicly much if at all. However, I have had a conversation with Olive about just this, asking Olive and Welti about the idea of different curves. I have a different idea than yours, which is that the "curve" is actually speaker directivity and room reflectiveness dependent. What I thought, but wasn't sure about, was that our brain is so used to elevated bass, that even in anechoic or free space conditions, we would prefer this. The standard response curve for outdoor music venues involves such a tilt. Welti and Olive confirmed this, that even in an anechoic chamber, we prefer the rise in bass, which normally happens naturally. As for what Dirac would do, it still follows the target curve. Bass is elevated in the standard target curve. How do I know, because I have long been a beta tester for DIRAC and I have used it on speakers sitting on a stand well up in the air, basically free space. I was curious what it would do. It followed the target curve very closely basically. The target curve was a mix of what I set and the bandwidth of the speaker, but it didn't set the target to flat.

​@@PoesAcoustics I'm going to use quotes because there's a lot of points to cover: "A few points of clarity, this study doesn't show that fully bandwidth automatic room correction like Audyssey or Dirac is good or better than only correcting at low and mid frequencies." Let me clarify my statement: it shows full bandwidth automatic correction can be better than transition frequency limited. Per their keynote it showed power response issues at around 3K which were well above transition frequency - when these were corrected they received subjective evaluation improvement. This is an example of a clearly off-axis response issue being "corrected" for the listening position. "All they reveal is that Harman's own method of correction is best, which makes use of knowledge of the speakers anechoic response in its correction. It was designed around JBL Synthesis speakers, which have a flat smooth DI" The speaker being tested (B&W 802N) had directivity problems and was being measured in the international reference listening room - I'm not familiar if they preloaded their solution with anechoic information though that would likely defeat the purpose of the test IMO. Even if we assumed the anechoic response was used to tweak their solution, another solution was used that clearly did not have that information and performed better than stock while filling in the higher frequency power response issue. As such there's presumably merit to going above transition if done properly including handling off-axis issues. The problem with handling them which Sean had mentioned in some comments at one point was risk of ruining listening for those on-axis - as such care would have to be taken. "What that means is that the application of EQ above the transition frequency can be ok as long as it is done carefully. Generally lower Q being the safest choices. You can make more robust EQ choices when you have the knowledge of the speakers actual free space polar response. This can't be easily obtained in room, though some correction systems, notably Dirac, have a sophisticated method of measuring that can do a good job obtaining this, as long as the measurements are taken correctly." I agree on the low Q, I naturally arrived at that solution doing my EQ by hand which gave me results extremely similar to Dirac (ironically ended up refunding Dirac because I thought it wasn't working on ver. 1.2). Generally I have higher Q filters around Davis frequency and below (mentioned by a prior Audioholics article) and low Q as frequency rises. I got these results using the primary listening measurement method described in Sean's keynote, a UMIK, and REQ shaping my target to the speaker's behavior in my room. Worked beautifully ¯\_(ツ)_/¯ "As for the "dynamic" nature of the target curve that Dirac uses, your description is not exactly correct. I wish what you said was true, but Dirac has a user adjustable correction curve, but the ends are automatically set based on the maximum badnwidth. In other words, it looks where the response naturally rolls off and then tries to limit the amount of correction. The overall shape is still static, picked by whatever curve you set." Dirac actually offered a recommend curve that varied from what I remember, but I could be wrong. Under that assumption looking at the roll-off to limit correction would effectively do something very similar to what I'm describing wouldn't it? A nearfield response wouldn't have as much roll-off so correction would be applied to a higher frequency compared to a mid-field in such a case? The goal should be to not force the speaker into a curve that doesn't fit which Audyssey seems to try and brute force with poor results as expected. "I have a different idea than yours, which is that the "curve" is actually speaker directivity and room reflectiveness dependent." I actually do believe the same thing, I'm not sure where you think we disagree? The speaker in the room with its reflections dictate the curve, there isn't a simple curve that can be uniformly applied. The failure of the X Curve showed that a fixed curve cannot be applied everywhere and sound good - I believe I indicated that in my first post but maybe I wasn't that clear. "Welti and Olive confirmed this, that even in an anechoic chamber, we prefer the rise in bass, which normally happens naturally." This I didn't know, interesting information that I will take into consideration moving forward. I do know there was also a section in treble and bass that prior studies did agree was still subjective to an extent while the central part of the passband was overwhelmingly shown to be preferred as linear in Harman's testing. I'd be interested if this was found to be a to taste thing or if there was truly a quantifiable preference going forward. >As for what Dirac would do, it still follows the target curve. Bass is elevated in the standard target curve. How do I know, because I have long been a beta tester for DIRAC and I have used it on speakers sitting on a stand well up in the air, basically free space. I was curious what it would do. It followed the target curve very closely basically. The target curve was a mix of what I set and the bandwidth of the speaker, but it didn't set the target to flat. So based on what you said it has a slight bass elevation with a fixed curve specified by Live, but based on deviation it will stop correcting at some point. So if the speaker is relatively flat nearfield it will give it a slight bass bump and correct it until it rolls off, and since midfield ones roll-off sooner in the FR due to reflected content it just stops correction at that point, am I understanding correctly?

ZinjaShike DIRAC won’t try to extend the response of a speaker past its limits. The curve it targets doesn’t match the natural response of the speaker however. That is my point. If you measure near or far, it doesn’t change how much tilt in the bass it will or won’t apply. It still attempts to match a target curve. When you say that there should be no single correct target curve, you are right. My point is that no correction does that. They all apply a static curve. DIRAC lets you vary it. It’s up to the user to create an accurate target. As for the study, I need to relook and we will check with Sean. It was my understanding they didn’t explicitly test a bandwidth limited version of their correction. I know the hidden anchor was no correction, but I thought it was just a comparison of a set of different correction systems on the market. I may have missed that detail. None the less, I am certain that Olive agrees with the assertion that most full bandwidth correction is problematic and that in many cases it is just a bandaid for bad speaker design. That a better speaker is a better option. That was always Toole’s hang up. They corrected a B&W speaker with a seriously flawed response and found that giving it a more neutral response sounded better. But is that still true if the speaker is already neutral? A speaker with a perfectly flat anechoic response and a perfectly flat DI will still not measure flat in a room due to reflections. Should that be corrected? My point in the video is no it shouldn’t be. A lot of correction systems will flatten that though. Or more importantly, the risk here. If the average person doesn’t understand the measurements they are talking, can they do more harm than good? We don’t recommend full bandwidth correction mostly because of this risk. Sure it can be done right. I think DIRAC mostly does it right. But I think a lot of people apply eq based on in room measurements inaccurately. While it can technically be done right, it’s too complex for most. I also worry that people will focus on features like DIRAC and spend extra for it when that should be a bonus. That money should first go to better speakers. So what I meant in the video is that I do not recommend end users attempt to correct the system full bandwidth. In addition, that most room corrections do more harm than good here. Not that there is no acceptable full bandwidth correction. However my big caveat is that I believe a perfect speaker doesn’t need this correction. I bet nobody could hear the difference between correction up to 500hz and correction up to 20khz on a speaker with a near perfect response. Full bandwidth correction can make a speaker sound better, but there is no guarantee it will. The software

@@PoesAcoustics "DIRAC won’t try to extend the response of a speaker past its limits." I wasn't asserting it would. In my experience it seemed to mimic the speaker's natural response while cleaning up linearity issues along it, but it appears I was mistaken based on what you're saying. So effectively it will boost/lower to meet a static target assuming it's correctable via EQ and also assuming it's not requiring excessive gain. "When you say that there should be no single correct target curve, you are right. My point is that no correction does that. They all apply a static curve. DIRAC lets you vary it. It’s up to the user to create an accurate target. " That's a shame, I thought Dirac was onto that but if not then we're still at the crossroads. I think a system that considers general response and adjusts the curve to estimate the expected result would be the best attempt at automated correction. It's not fun spending 2-3 hours doing it by hand, though it's quite affordable! "None the less, I am certain that Olive agrees with the assertion that most full bandwidth correction is problematic and that in many cases it is just a bandaid for bad speaker design. That a better speaker is a better option. That was always Toole’s hang up." On one hand I would say yes, absolutely. On the other that becomes a practical concern when cost comes into play. If dynamic room correction that behaved in a way I was discussing existed for say ~$500, would it be worth it? The question is how much do you have to spend to get a well behaving speaker? $1K for a pair of bookshelves that behave really well in directivity and have reasonable distortion performance in small rooms? When talking larger stuff the price goes up drastically and even more as output turns into a real concern. My setup uses speakers with good directivity (compression drivers coupled to waveguides and solid xover design) with mild FR issues. Since directivity isn't the primary concern they're prime candidates for full-range correction (though I've used it on others with good results). The speakers I have were remarkably cheap used. To get ballpark output/directivity performance I'd have to spend 6x new, 3x used in the home enthusiast market - they would come without the minor FR hiccups at a premium. A solution for $500 would save significant money for my setup compared to the ideal speakers. The band-aid full bandwidth correction (by hand in my case) provides me a solid compromise getting me performance I desire at a price much lower than expected. The goal for those where cost is a real factor comes down to compromise mitigation, and correction can be a valuable tool in that context. > But is that still true if the speaker is already neutral? A speaker with a perfectly flat anechoic response and a perfectly flat DI will still not measure flat in a room due to reflections. Should that be corrected? My point in the video is no it shouldn’t be. A lot of correction systems will flatten that though. Agreed on all accounts. >I also worry that people will focus on features like DIRAC and spend extra for it when that should be a bonus. That money should first go to better speakers. Agreed to an extent, it's finding the right balance to where the tool serves a valid purpose without creating issues due to utilizing it whether it be misallocation of budget, misapplication, etc. I think we mostly agree. It definitely is easy to make a good speaker sound bad and takes understanding to make the compromise sound good. Hopefully we get a solution that can prevent the former and enforce the latter in the near future providing better tools and value to the consumer. It's been a pleasure discussing :)

wow lol that is deep

Slap echo can be measured but I don't have a way for you to quantify it. Most people look at RT60 time instead. As I mentioned, the theory behind RT60 is invalid in a small room, but the measurement of decay in a room is accurate. You can use Room EQ Wizard to do this. Just ignore all the fluff on using waterfalls, those are a waste of time in my opinion. Stick to RT60 and wavelets as a better alternative to the waterfall. As for acceptable range, as noted before, its about what is comfortable for you. We will need to go into this more later, but the problem is, eliminating slap echo totally may not sound any different than allowing for some of it, but eliminating it totally may make the room too dry. As for what I target, I prefer to target a flat and even decay rate from 100hz to 5khz and I prefer to see a gradual and minimal rise in the decay down to 20hz. RT60 won't be as accurate down past 100hz with REW, so I switch to wavelets. The wavelet will make a hickey stick shape, which is ok. The broadening at the base is fine. You want to see that it still remains linear and that the broadening is minimal. I also look at bandwidth limited impulse responses. So you do a 1/3 octave at 250 hz and what you look for is that it should have roughly an egg shape. After the egg, an anechoic room will have nothing. A reverberant room that is treated well will have a gap or space with nothing, but there will be obviously smaller eggs which are the reflections. What you don't want is an irregular egg with a long irregular tail. That implies modal interference and strong reflections.

Doug Hilton I mean, we can do that. Money is an object and we readily accept it.

Rudy Rutan the best option is probably to seek consultation from experts. Complex rooms are hard to treat without really digging into the details of the room. I gave away my free knowledge in the video and will do so in more videos. For targeted advice I would have to charge.

Alain my main gripe with the Minidsp is that it adds delay. If your main speakers don’t also go through it, that delay is group delay. If you don’t have a way to compensate for that group delay, it can become audibly significant. If you do run your speakers through it, it can audibly compromise fidelity. The few I’ve used added noticeable hiss. I wasn’t a fan of that. However, without that kind of DSP control, multisub can be tricky. The individual subs may need individual delay, phase, and crossovers. Most HT processors can’t handle that. For many people it’s a good option, they just need to be sure to compensate for the delay.

@@PoesAcoustics Man, they're MiniDSP 2x4 HD nuts over on the AVS subwoofer threads for BEQ. It's like 300 dollars with the wifi dongle they recommend to quickly upload the profiles. That's not cheap. I keep going back and forth whether to get one or not. I have Audyssey XT32 with he MultEQ mobile app already so I'm in no rush to add more DSP.

Thomas Hoctor do you have your own measurement capabilities? If so, i would take measurements of your system at a set of key locations around your listening position. If you have a good flat response in the bass and it sounds good, you don’t need the Minidsp. It’s just a tool, if you don’t need that tool, done buy it.

@@PoesAcoustics Yes I do Matt. I have REW on my Acer laptop and a CSL Umik-1 I bought used from an AVSer a while back. I've also watched your 2 REW tutorial videos several times as well even though I had a good understanding of how REW worked before that from following Austin Jerry's FAQ those videos also helped immensely. The wavelet advice and RT60 just to mention 2 tips from those videos. I know a lot of MiniDSP users who EQ their setups to death using REW. I have the advantage of a smaller sealed rectangular space with equidistant dual subwoofer placement from the MLP. Unfortunately that MLP is smack up against a back wall and there's not really much I can do about that seating wise sadly. I only recently bought the MultEQ mobile app though so I haven't run it through my laptop using the Android BlueStacks emulator just yet. I still just correct full range through the receiver with Audyssey XT32.

Hi Matt, yes I'll still probably try out the minidsp later just because of the multisub configuration. I plan add more bass trapping first to try to flatten the bass response to an acceptable level then add the minidsp after.

Kind of, but its a lot more complicated than that. It isn't something that can be done by most consumers. I could do a video on this, but it would have to be only informational. I have developed my own method to analyze binaural impulse responses to assess rooms and I would like to keep it proprietary for now. Simply measuring with two microphones doesn't achieve the right results. The microphones don't mimic our hearing. On top of that, for many measurements, a single omnidirectional microphone is actually most appropriate. For example, to assess room decay, modes, etc. a single mic is fine in a room. However, to measure the binaural reverberation of the room, we need a binuaral microphone and it needs to accurately mimic the response of the head and ears. I actually use a modified version of the MiniDSP EARS device. My experience with it is that it is fine for what I am doing as compared to the 100 times more expensive professional product. However, the trick is how you use it. You cannot take a software like REW, connect the two channel EARS, sweep the room, and use that. What I've done is developed a method to replay the tones, capture them in stereo, and then convert those recordings into the binaural impulse response. From there I process the two impulse responses in order to detect the source and direction of reflections in a room. Since doing this was a lot of work for me and required developing some proprietary programs in Octave, I don't want to just give it away. Keep in mind that a binaural measurement still isn't appropriate for EQ. It's just used to assess the acoustics of the room, and even then, only certain aspects of the acoustics. I have a client right now that I am working with who has requested full measurements of a new dedicated space, and so I will be doing binaural impulse response measurements at three stages. No walls, Bare Walls, fully treated. I can show the differences in these measurements over time to provide everyone with a sense of what I find. I hope to obtain a 3D impulse response setup in the future and if I can obtain this, I will be doing a lot of projects and videos for you all. A binaural measurement like I take is still limited in it's ability to detect direction. Since current software (or at least my current math knowledge) doesn't allow me to model an HRTF and then use that to detect direction across a full sphere of the head, I am only able to look at the lateral reflections. Basically, I can tell there are reflections and if they come from the left, middle, or right side. I can't tell if they come from above, below, in front, or behind. To figure that out, I use meausrements of the room to calculate the fly path and basically guess at the location. To make that feasible, the dummy head must be a specific non-symmetric location within the room. The 3D system uses 4 channels of mics and as a result can detect direction no matter where the mic is placed and requires a lot less guessing.

@@PoesAcoustics I really appreciate your response, with that said whenever you finish developing your system, I need you to come measure and then treat my room(s) using the method and equipment you develop ✊🏿. I look forward to it hitting the market, and wish you the best going forward. Once the again thank you, and Gene. 🙏🏾🙏🏾👌🏿👌🏿👌🏿.

isaacsykes3 it probably won’t come to market as a consumer retail product. I didn’t invent a new technology, just modified what was in the literature to meet my need. The problem, for now, is that nobody knows what a good binaural impulse response of a small room looks like. All of the research and tests is largely about either auralization or large acoustic spaces. For now I’m using this method to collect data that i can correlate with subjective impressions. I do use it to detect the direction of modes and their strength. I would like to see that technology built into consumer products but the approach I’m using is crude. A better method would be a spherical 3D mic. I am hopeful that I will be able to work with a certain free measurement software to implement some of what I’m doing. For now it won’t be possible until that free software moves to a 2-channel FFT. That is a long term project.

isaacsykes3 where are you located? I am able to measure people’s room using this technique now, I just can’t give the script I developed out to anyone.