Home room measurements

[Keith_W]

RT60 - doesn’t it also depend on the speakers as well? I mean, at the very least, the directivity should have some effect, especially if the walls, ceiling, and floor have significantly different/variable absorption or reflection properties. I could imagine a measurement that, for specific locations of the future speakers and the listening spot, would capture the characteristics of the room that are complementary to the directivity of the speakers - the directivity of the room, if you will - and then try to predict which speaker directivity pattern would work better.

About the repeatability of the low-frequency response - doesn’t the same go for speakers? I mean, you can take an otherwise perfect speaker with a flat anechoic response, place it in a real room, and that flatness is just gone in the modal region. Still, we value those speaker measurements.

Of course you are right. The RT60 correlates with the sound power of speakers, and as you know, with nearly all speakers the sound power has a distinct downwards tilt from 20Hz to 20kHz. Some more pronounced than others. Speakers which have narrower directivity towards lower frequencies, e.g. cardioid designs or dipoles, have a less pronounced downwards tilt.

However, the actual RT60 measured at any location in a small room (i.e. nearly all domestic listening rooms) is specific to the observation point and the location of the speakers. This is because the "R" in RT60 stands for "reverberation", IOW you need a reverberant field. A reverberant field is thousands of overlapping room modes, such that the SPL of the reverberant field is the same no matter where in the room it is measured. Small rooms do not form reverberant fields in all but the shortest wavelengths. As Toole says, it is more accurate to call it "early reflection time" because you are measuring specular reflections and not reverberant fields. And, as you can imagine ... how much specular reflection you get depends on speaker setup, toe-in, observer position, as well as the speaker characteristics you mentioned.

This does not mean the RT60 measurement is useless, but that it should be interpreted with caution. Yes, it gives you an idea how your room is behaving. But the error window is reasonably large, especially with long wavelengths. How large? You can find out by measuring the RT60 at different points in your room. If these were true reverberant fields, all the measurements would be the same. But I guarantee they won't be.

The answer to your question is easy - narrow directivity does win hands down in this room, as I’m sure you’d expect. But it still doesn’t sound good. Yes, the view is great, and I really wanted to have a second listening location in the house in this room. But alas, it just doesn’t work for that purpose.

Yes, I was expecting that. I think you mentioned somewhere that stereo imaging is poor, so your problem is reflections from somewhere. I can totally understand why you are not willing to desecrate that gorgeous wall or that stunning window with room treatment. Maybe go to your other purpose-built listening room when you want good sound, and go to this room when you want the view

 

[Multicore]

Of course you need speakers and a microphone to measure a room.

But do you really? There's something quite appealing about measuring the room's impulse response directly. Balloon pops are fairly well studied. I guess a blank in a pistol could deliver a bit more energy. But you'll be in the lead on style points with the Hungarian whip.

 

[tengiz]

Maybe go to your other purpose-built listening room when you want good sound, and go to this room when you want the view

Yes, this is what I have to endure . Maybe - if I’m lucky - I’ll eventually find some miracle in-ceiling replacements that actually work magic. The current ones, as I mentioned, sound reasonably balanced overall, but they produce what I can only describe as a “coming-from-behind-a-corner” audio effect. It’s spatially vague, and while the sound is consistent, it’s far from ideal for anything that calls for immersive or focused listening. That said, it almost works when you're distracted by the visuals , especially with the right kind of music.

 

[Hear Here]

Odd question perhaps on this forum, but have you considered looking for speakers that you reckon would be good in your room without even thinking about measurements?

Your room features should point towards a particular TYPE of speaker - horn, electrostatic, omni, conventional box, etc. Getting this right is the biggest factor in achieving life-like music in your room. The more clutter and angled walls (within reason of course) the better, and carpets, curtains, etc will influence the result too.

Nice view and the acreage of glass may influence the best choice of speaker type. If you are particularly concerned about wall reflections messing up the music, consider the most directional of speakers - horns - but they have a small but very sweet sweet spot. If you want to take advantage of reflections (from the wall behind the speakers) then perhaps electrostatics may be ideal, but they don't like some room features, such as my own curved floor-to-ceiling glazed wall a long way behind the speakers. If you want good music at multiple listening positions, omnis are the obvious choice, though your sweet spot will never be as sweet as some other types and imaging will be more vague.

Try to generate a short list of a few likely contenders and arrange home demos for each, as speakers will sound far different in your room compared with the dealer showroom or at shows. This needs research by reading reviews, asking for recommendations on forums such as this (and more subjective ones!), spec sheets, prices, aesthetics, etc, Don’t be over-influenced by measurements as the room will chuck them out the window. Good solid well-engineered speakers from a respected brand are more likely to shine and of course they will keep their value if and when you move house or change to even better speakers.

Don’t be tempted to go for any speaker you have doubts about and think that chucking DSP at them will sort out their shortcomings! Get speakers that don’t need the support of unnecessary signal processing

Others may disagree!

 

[Sjkrabbe]

Bookshelves, plants, sofas, people etc., all tend to help the room acoustics.

I agree - from practical experience in my own living room. Look up "IK Media ARC roomcorrection"

 

[DanielT]

Edit: I mean, this is one of those cases where you really do want to have a pretty good idea before cutting into walls or the ceiling - unlike floor-standing speakers, which you can just move on from if they don’t work out.

You have a point. It's a practical aspect to take into account. Regarding in-celinig speakers, I don't really know anything about that type of speaker. Amir has tested some which resulted in rather mediocre results if I remember correctly. If anyone can recommend some good ones, please do. Now I'm curious myself.

There’s easy access to the attic above this room. It’s unfinished, and there’s no floor directly above the speakers, so they’re actually pretty straightforward to replace - they’re fully exposed from above. There’s also plenty of space to install larger speakers if needed. I’ve spent some time looking into possible replacements, but the real question is how to choose speakers that would actually work well in this room and in that particular ceiling location. I know KEF makes architectural models with their coaxial drivers, but would it even make sense to install something that advanced so close to the corners? Plus, the ones that I looked at are not sealed, so they assume the enclosure created by the in-wall installation. This is not my case - I would need to add some box or something.

Mount some speaker drivers on a suitable board and screw it in yourself? If you ever think about some type of DIY in-celinig speakers , start a new thread in the DIY section here at ASR and you will most likely get good tips and advice.

Have you tried placing the speakers really close to you?

Then we have headphones, but that may not be your thing?

 

[DanielT]

Odd question perhaps on this forum, but have you considered looking for speakers that you reckon would be good in your room without even thinking about measurements?

Your room features should point towards a particular TYPE of speaker - horn, electrostatic, omni, conventional box, etc. Getting this right is the biggest factor in achieving life-like music in your room. The more clutter and angled walls (within reason of course) the better, and carpets, curtains, etc will influence the result too.

Nice view and the acreage of glass may influence the best choice of speaker type. If you are particularly concerned about wall reflections messing up the music, consider the most directional of speakers - horns - but they have a small but very sweet sweet spot. If you want to take advantage of reflections (from the wall behind the speakers) then perhaps electrostatics may be ideal, but they don't like some room features, such as my own curved floor-to-ceiling glazed wall a long way behind the speakers. If you want good music at multiple listening positions, omnis are the obvious choice, though your sweet spot will never be as sweet as some other types and imaging will be more vague.

Try to generate a short list of a few likely contenders and arrange home demos for each, as speakers will sound far different in your room compared with the dealer showroom or at shows. This needs research by reading reviews, asking for recommendations on forums such as this (and more subjective ones!), spec sheets, prices, aesthetics, etc, Don’t be over-influenced by measurements as the room will chuck them out the window. Good solid well-engineered speakers from a respected brand are more likely to shine and of course they will keep their value if and when you move house or change to even better speakers.

Don’t be tempted to go for any speaker you have doubts about and think that chucking DSP at them will sort out their shortcomings! Get speakers that don’t need the support of unnecessary signal processing

Others may disagree!

Horn speakers may be an appealing solution for tengiz. With horns /WG you can choose the dispersion pattern that suits your needs.

Horns & Waveguides

Explore our stock of horns and waveguides. Find the partner for your horn driver to deliver midrange and high frequencies, with reliable range and coverage.

www.parts-express.com

Speaking of room acoustics and omnis.
In the 1970s, when many homes were heavily furnished, these OA 5 type II, not fully omni but close to it, were the most popular speakers in Swedish homes:

Then the years passed and the rooms became more sparsely furnished. The last model that Carlsson created was the OA 52.2. Not the least bit omni with the 1990s model:



Conditions and so also probably the sound ideal changed as the years passed.

 

[tengiz]

Odd question perhaps on this forum, but have you considered looking for speakers that you reckon would be good in your room without even thinking about measurements?
...

Yes, I’m open to considering anything - within reason

Ceiling-mounted or very unobtrusive designs preferred. I previously tried Martin Logans in this room, and while the sound quality was ok, not great, they visually dominated the space. It just didn’t look right. So I’m aiming for something much less imposing.

Wall-mounted is also on the table, possibly placed on either side of the large window. But that would need to be a one-shot success . I know Focal offers collinear driver designs that allow you to aim the midrange/tweeter unit, but I still worry about reflections and interference from the side walls.

I’ve also seen some in-ceiling speakers with angled, recessed baffles that aim the drivers toward the listening area. But without firsthand experience with that kind of design, I’d be flying blind.

Wide listening area is not a priority. I’m the only one in the family who really is that picky about sound quality, so I’m totally fine with an “ultimate egoist” setup. That said, if there’s a way to find a reasonable compromise that broadens the sweet spot - even if it makes it a bit less “sweet” - it would be fine too.

I don't know if wall limits precludes the horn design - what I've seen is pretty deep and won't fit in the wall, but it might in the ceiling.

 

[tengiz]

Then we have headphones, but that may not be your thing?

I’ve never been able to shake the feeling that music through headphones is happening inside my head - right between the ears. It just doesn’t convince me spatially. I can still enjoy it, sure, but it never comes close to what I get from speakers.

 

[Holmz]

Horn speakers may be an appealing solution for tengiz. With horns /WG you can choose the dispersion pattern that suits your needs.
View attachment 466425

Horns & Waveguides

Explore our stock of horns and waveguides. Find the partner for your horn driver to deliver midrange and high frequencies, with reliable range and coverage.

www.parts-express.com

Speaking of room acoustics and omnis.
In the 1970s, when many homes were heavily furnished, these OA 5 type II, not fully omni but close to it, were the most popular speakers in Swedish homes:
View attachment 466426View attachment 466428
Then the years passed and the rooms became more sparsely furnished. The last model that Carlsson created was the OA 52.2. Not the least bit omni with the 1990s model:

View attachment 466429

Conditions and so also probably the sound ideal changed as the years passed.

A narrow dispersion speaker can work well in a reflective room.

 

[Hear Here]

Yes, I’m open to considering anything - within reason

Ceiling-mounted or very unobtrusive designs preferred. I previously tried Martin Logans in this room, and while the sound quality was ok, not great, they visually dominated the space. It just didn’t look right. So I’m aiming for something much less imposing.

Wall-mounted is also on the table, possibly placed on either side of the large window. But that would need to be a one-shot success . I know Focal offers collinear driver designs that allow you to aim the midrange/tweeter unit, but I still worry about reflections and interference from the side walls.

I’ve also seen some in-ceiling speakers with angled, recessed baffles that aim the drivers toward the listening area. But without firsthand experience with that kind of design, I’d be flying blind.

Wide listening area is not a priority. I’m the only one in the family who really is that picky about sound quality, so I’m totally fine with an “ultimate egoist” setup. That said, if there’s a way to find a reasonable compromise that broadens the sweet spot - even if it makes it a bit less “sweet” - it would be fine too.

I don't know if wall limits precludes the horn design - what I've seen is pretty deep and won't fit in the wall, but it might in the ceiling.

Frankly I think you'll be hard pressed to find ceiling or wall mounted speakers that sound anything better than mediocre. Imaging (a feature I highly prize) will inevitable be poor, but perhaps you'll find a system that sounds acceptable. Unfortunately a home demo will be impossible to add to your problems! Good luck with your search.

PS - One speaker that I was most impressed with that was installed in contact with the wall behind them was a hORNS speaker, I think it ws this model

FP10 MK 3 – hORNS

horns.pl

Having a respectable 10" bass driver with horn-loaded tweeter, it offered a very convincing sound. I wouldn't go for the stands shown but they may look acceptable mounted in some other way

 

[Mike123]

Odd question perhaps on this forum, but have you considered looking for speakers that you reckon would be good in your room without even thinking about measurements?

It is very long, expensive and with an uncertain result. DSP very quickly, almost free (Laptop + microphone + ADC + REW + equalizer APO) will allow you to evaluate the problems, hear the result very quickly and practically understand what problems there are in a specific room with specific speakers. And choose speakers by price and appearance. For me, a revelation, which I even added to bookmarks, is testing the AmirM acoustics Revel F35 and manipulating the equalizer. I would always quote one paragraph from there in such disputes:

Past experience with equalizing the room in low frequencies had taught me that removing bass resonances not only improves bass, but has a remarkable (good) effect on notes above that range. Resonances boost certain bass frequencies. In time domain that lengthens the amount of time it takes for those high energy notes to go away. And that steps on much lower level notes in higher frequencies.

As I understand it, resonances clog the ears with loud sound and the ears stop hearing everything except loud resonances. Solving these problems by selecting speakers is a lost cause.

 

[DanielT]

A narrow dispersion speaker can work well in a reflective room.

And here are some narrow dispersion speakers that Audioholics recommends:

 

[alankila]

By room measurements, I mean the characteristics of the room that are independent - or mostly independent - of the speakers.

Picture this: the room is already set up, furniture and all. I know where the speakers and listening spot will go. The goal is to preserve the layout and design first - then pick gear that fits.

So how do I measure the room in a way that helps me choose speakers that’ll actually work well in this specific setup?

We already have solid, measurement-based speaker rankings - directivity, off-axis response, and so on. But in-room response predictions still seem based on simplified models - maybe not a spherical room in a vacuum, but close enough.

They’re grounded in real data, sure, but how well do they handle real-world spaces? Most rooms are far from ideal - full of furniture, odd shapes, slanted ceilings, asymmetrical openings, random nooks. That’s what people are actually dealing with.

So is there anything that can take a real room, run some measurements, and suggest speakers likely to work well in that space?

It’s physically doable. The real question is whether a reasonably skilled DIY-er can pull it off without too much cost or hassle. I know one way to approach it - I’ve done something similar before, just not for room acoustics. But honestly, it’s probably overkill for this kind of use. You’d need specialized hardware and a pretty niche skill set to build it from scratch at home.

Just wondering if anything new has emerged on this front.

The Spinorama method for the predicting the in-room response involves the following constants and measured datasets:

- 12 % of the sound pressure perceived is from on-axis
- 44 % of the sound pressure is early reflection bounces at various angles towards the walls, ceiling and floor.
- 44 % of the sound pressure is the sound power generally, which is the entire spherical radiation field of the speaker.

Note that the sound power contains the reflection bounces and the on-axis audio as part of itself, as that is the full radiation field of the speaker. Therefore, the contributions of on-axis and early reflections become greater than these % values actually indicate. I haven't calculated how much, but essentially this method evaluates the entire spherical field with a weighing function that prioritizes some directions of radiation more than others, and this produces the ultimate estimate. It's also worth remembering that the only thing behind the speaker of typical boxed speaker is the bass which inside a room will hit the front wall, and mostly can't go through it but bounces off and gets summed with the direct sound, so sound power likely mostly just adds that into the total prediction.

Any specific placement involves computable distances that give rise to time of flight of sound that induces phase shift that causes cancellations as the reflected sound sums with direct sound. This type of reasoning is likely wiggling the response up and down between +6 dB and some negative dB value, generally averaging lower. I think the prediction from Spinorama can be reasonable if side walls, floor, and similar are not very close to the speaker, and if you consider the prediction to concern a larger volume near the listening seat, rather than just the main listening point.

 

[Holmz]

Yes, one of them. But it clearly is not suitable for listening with decent fidelity when paired with otherwise fantastic floorstanding speakers with dynamic transducers.

Even if you left curtains open, having some curtains bunched up in the corners would not hurt.
Nor would a nice plants on each side of the large window.
Some hanging rug on the walls above the sofas would not hurt.
One does not need to use “treatments” per se, to get some treatment and still look pleasing to the eye.
Or an animal pelt.

 

[tengiz]

The Spinorama method for the predicting the in-room response involves the following constants and measured datasets:
...

Thank you. So, as expected, the model - while not exactly trivial - is still idealized. And that was the point of my original question: it seems like the model could be made more complete if one were motivated enough to take some (probably non-trivial) room measurements and provide additional input parameters into the Spinorama-based prediction. For example, the ratios between on-axis, early reflections, and total radiated sound. There’s not much more to it, really - at least when it comes to predicting room response and estimating tonality.

What prompted this was the Spinorama disclaimer: “The tonality (Preference) score does not make sense for panel speakers. Don't give too much credit to the value.”

I’m sure that’s because the directivity pattern of panel speakers doesn’t align with the assumptions of the Spinorama model. But speaker directivity always interacts with what you could call the “room directivity” - that is, the absorption and reflection properties of the room’s surfaces, depending on the speaker and listener positions.

Of course the simplified model assumes that some "canonical room" characteristics are baked into it.

 

[tengiz]

Even if you left curtains open, having some curtains bunched up in the corners would not hurt.
Nor would a nice plants on each side of the large window.
Some hanging rug on the walls above the sofas would not hurt.
One does not need to use “treatments” per se, to get some treatment and still look pleasing to the eye.
Or an animal pelt.

That’s all true - but imagine if, for whatever reason (perhaps not a convincing one to you), there’s no option to make any changes that would visibly alter the look of the room.

 

[alankila]

Thank you. So, as expected, the model - while not exactly trivial - is still idealized. And that was the point of my original question: it seems like the model could be made more complete if one were motivated enough to take some (probably non-trivial) room measurements and provide additional input parameters into the Spinorama-based prediction. For example, the ratios between on-axis, early reflections, and total radiated sound. There’s not much more to it, really - at least when it comes to predicting room response and estimating tonality.

What prompted this was the Spinorama disclaimer: “The tonality (Preference) score does not make sense for panel speakers. Don't give too much credit to the value.”

I’m sure that’s because the directivity pattern of panel speakers doesn’t align with the assumptions of the Spinorama model. But speaker directivity always interacts with what you could call the “room directivity” - that is, the absorption and reflection properties of the room’s surfaces, depending on the speaker and listener positions.

Of course the simplified model assumes that some "canonical room" characteristics are baked into it.

I suppose one could provide reflectivity for the side walls, floor and ceiling separately. SBIR cancellations from floor bounce and front wall could be estimated based on positioning and dimensions using simple early reflection line paths. Panel speakers are thorny because they are dipoles, and they send bass both front and back, but the backside is out of phase and destructively interferes. So, the summation from sound power is unlikely to predict < 200 Hz bass very well.

The in-room curve doesn't really predict other than what speaker might sound like in some kind of typical room within a wide volume around the listening position, and that hypothetical room also has neither modes nor cancellations. Developing more accurate models might require incorporating phase of the radiation field, more accurate description of the spherical sound field because the spins are somewhat coarsely distilled in the data, and there should be some kind of high resolution wave propagation algorithm to estimate the impulse response resulting in the wave propagation simulation to the listening position. All of these might be worthwhile goals, perhaps, but also gets multiple orders of magnitude more difficult to do.

 

[tengiz]

Panel speakers are thorny because they are dipoles, and they send bass both front and back, but the backside is out of phase and destructively interferes. So, the summation from sound power is unlikely to predict < 200 Hz bass very well.

That, and also the deviation from the 1/R^2 falloff, I presume.

The in-room curve doesn't really predict other than what speaker might sound like in some kind of typical room within a wide volume around the listening position, and that hypothetical room also has neither modes nor cancellations. Developing more accurate models might require incorporating phase of the radiation field, more accurate description of the spherical sound field because the spins are somewhat coarsely distilled in the data, and there should be some kind of high resolution wave propagation algorithm to estimate the impulse response resulting in the wave propagation simulation to the listening position. All of these might be worthwhile goals, perhaps, but also gets multiple orders of magnitude more difficult to do.

Right, but much of that data is already collected by systems like Klippel for speakers. What’s missing is the other half - a kind of “Klippel for rooms” . Depending on how it’s done - and I know it’s possible - propagation, reflection, scattering, and absorption in the room could already be folded into the measurement results. Imagine something like a full 3D transfer function from the speaker locations to the listening positions.

 

[Holmz]

…interferes. So, the summation from sound power is unlikely to predict < 200 Hz bass very well.

…

That <200 Hz is the only place that one would have a hope.
These graphs usually like like a lice comb, and the only thing one really knows is the envelope of the peaks and nulls out at the high frequency.

That’s all true - but imagine if, for whatever reason (perhaps not a convincing one to you), there’s no option to make any changes that would visibly alter the look of the room.

Ok - so we are giving up on taming the room.
And I guess that is what the speakers are hidden in the ceiling?

….

So is there anything that can take a real room, run some measurements, and suggest speakers likely to work well in that space?

It’s physically doable. The real question is whether a reasonably skilled DIY-er can pull it off without too much cost or hassle. I know one way to approach it - I’ve done something similar before, just not for room acoustics. But honestly, it’s probably overkill for this kind of use. You’d need specialized hardware and a pretty niche skill set to build it from scratch at home.

Just wondering if anything new has emerged on this front.

There are ray tracing algorithms. But what is the point?
They will tell you how the room sounds in an SPL graph, but they may be less likely to tell you how the speakers sound.

A line array should be narrower in the vertical axis.
And the planar speakers might have a narrow radiation.

But if nothing can be visible, then just make sure that the sound has tone controls and call it a day.