I had hoped someone else would post this but prompted by comments in recent threads I am posting to see if there is interest in a group project on the topic of high fidelity on-wall speakers.
Outline speaker design:
- against wall required with mounting options to be fleshed out in discussion
- clean 105+ dB peaks at listening position
- 60 - 20kHz passband with subwoofer support below 80 Hz
- smooth directivity with details to be fleshed out in discussion
- standard range rather than premium range drivers
- budget in $1000 range but subject to movement in discussion
- active crossover for R&D with passive version later if interest
Outline project approach:
- focus on simulation to develop design and options in a quantitative manner
- no specific drivers to be considered until the speaker configuration is designed
A few points to kick off discussion:
No passive crossover constraints in the initial design but if there is still interest in a passive crossover version after the active crossover version has settled this would be performed later. So not rejecting passive crossovers just suggesting pushing their additional constraints to later in the project.
Handling the front wall well in the simulations may require a modified approach with BEM being one way to handle the 3D details. I have some experience with the open source acousto software though not all good. Others appear to have experience with AKABAK/ABEC. Is the latter suitable? Other options?
Commercial products like COMSOL are likely the fastest and easiest option for detailed simulations but few if any are going to have access. Should the project use only software that everyone can use to check the design and/or develop further for their own designs or is expensive commercial software OK?
Will need a repository to store simulations, design documents, possibly lengthy detailed posts and the like. Options? How well did the Open Source Speaker group project get on with their choice?
Options for woofer loading in the presence of the wall? Given sub support and wall mounting the smaller size of a sealed cabinet is perhaps the way to go?
If we opt to place the woofers on the side, the midrange on the front with the crossover frequency between the frequencies of the wall reflection nulls for the drivers then a pair of opposed woofers is looking attractive. It will also reduce cabinet vibration substantial and possibly help reduce the complexity of wall mounts. On the downside it would require space on both sides which may limit things like TV screens.
Cardioid is an alternative approach to handlng the front wall which we should consider. It will need more cone area and more complexity but would seem to be a candidate for high-fidelity speakers.
Another option to consider is a shallow speaker with a wide baffle. This is the simplest approach but also the one likely to handle the wall least well. The pros and cons to be quantified by simulation even if we don't adopt?
Outline speaker design:
- against wall required with mounting options to be fleshed out in discussion
- clean 105+ dB peaks at listening position
- 60 - 20kHz passband with subwoofer support below 80 Hz
- smooth directivity with details to be fleshed out in discussion
- standard range rather than premium range drivers
- budget in $1000 range but subject to movement in discussion
- active crossover for R&D with passive version later if interest
Outline project approach:
- focus on simulation to develop design and options in a quantitative manner
- no specific drivers to be considered until the speaker configuration is designed
A few points to kick off discussion:
No passive crossover constraints in the initial design but if there is still interest in a passive crossover version after the active crossover version has settled this would be performed later. So not rejecting passive crossovers just suggesting pushing their additional constraints to later in the project.
Handling the front wall well in the simulations may require a modified approach with BEM being one way to handle the 3D details. I have some experience with the open source acousto software though not all good. Others appear to have experience with AKABAK/ABEC. Is the latter suitable? Other options?
Commercial products like COMSOL are likely the fastest and easiest option for detailed simulations but few if any are going to have access. Should the project use only software that everyone can use to check the design and/or develop further for their own designs or is expensive commercial software OK?
Will need a repository to store simulations, design documents, possibly lengthy detailed posts and the like. Options? How well did the Open Source Speaker group project get on with their choice?
Options for woofer loading in the presence of the wall? Given sub support and wall mounting the smaller size of a sealed cabinet is perhaps the way to go?
If we opt to place the woofers on the side, the midrange on the front with the crossover frequency between the frequencies of the wall reflection nulls for the drivers then a pair of opposed woofers is looking attractive. It will also reduce cabinet vibration substantial and possibly help reduce the complexity of wall mounts. On the downside it would require space on both sides which may limit things like TV screens.
Cardioid is an alternative approach to handlng the front wall which we should consider. It will need more cone area and more complexity but would seem to be a candidate for high-fidelity speakers.
Another option to consider is a shallow speaker with a wide baffle. This is the simplest approach but also the one likely to handle the wall least well. The pros and cons to be quantified by simulation even if we don't adopt?
Nice opening post, at least I'm glad you did it! 🙂
The BEM simulations would be very interesting I think. After all, on wall placement is just that, wall very near and what that brings into the soup at listening position. I think an on wall system design can be brought quite far without BEM, just like any other design, but eventually measuring the setup might be the hardest task unless it was just the in room listening position measurements. I mean, the BEM sims are propably easier to get all sorts of off-axis measurements than real world measurements the speaker on the wall.
I've been building mine with basic quasi-anechoic home measurement and simulation setup like any other speaker, just paying attention to the angles towards early reflections as well as minimizing distance to the wall. In general, trying to minimize the influence of front wall to the sound at listening position. This would be benefitical to any speaker design I believe, if it comes without bad trade-offs elsewhere.
Is there any other differences to normal speaker positioning than the front wall proximity, shift of interference pattern? Sidewall, floor and ceiling reflections should stay pretty much the same no matter front wall proximity, but it just occured to me that there might be quite strong reflections from floor and ceiling through the front wall, as secondary path whose path lenght is almost the same as direct from driver to ceiling /floor? I've reasoned waveguide and cardioid mid for these, shallow bass box to get lowest SBIR effects into the cardioid mid bandwidth.
The classic near wall boom bass seems to be merely EQ thing but I'm not too deep thinking the bass box yet for my system. I think sealed is reasonable way to go for the bass section especially additional subwoofer system in mind but lets see how tall on-wall speaker we are gonna get 🙂
Hopefully there is lot of input and interest! Speaker like this fits into a family living room very well.
The BEM simulations would be very interesting I think. After all, on wall placement is just that, wall very near and what that brings into the soup at listening position. I think an on wall system design can be brought quite far without BEM, just like any other design, but eventually measuring the setup might be the hardest task unless it was just the in room listening position measurements. I mean, the BEM sims are propably easier to get all sorts of off-axis measurements than real world measurements the speaker on the wall.
I've been building mine with basic quasi-anechoic home measurement and simulation setup like any other speaker, just paying attention to the angles towards early reflections as well as minimizing distance to the wall. In general, trying to minimize the influence of front wall to the sound at listening position. This would be benefitical to any speaker design I believe, if it comes without bad trade-offs elsewhere.
Is there any other differences to normal speaker positioning than the front wall proximity, shift of interference pattern? Sidewall, floor and ceiling reflections should stay pretty much the same no matter front wall proximity, but it just occured to me that there might be quite strong reflections from floor and ceiling through the front wall, as secondary path whose path lenght is almost the same as direct from driver to ceiling /floor? I've reasoned waveguide and cardioid mid for these, shallow bass box to get lowest SBIR effects into the cardioid mid bandwidth.
The classic near wall boom bass seems to be merely EQ thing but I'm not too deep thinking the bass box yet for my system. I think sealed is reasonable way to go for the bass section especially additional subwoofer system in mind but lets see how tall on-wall speaker we are gonna get 🙂
Hopefully there is lot of input and interest! Speaker like this fits into a family living room very well.
Last edited:
Some inspiration - Bookshelf-3WC
Within budget if you include cabinet material and crossover components and designed to be used in a bookcase or against the wall.
also the SpeakerDesignWorks Statement monitor has a close to wall crossover option - Statements_Monitor
just something to get the juices going.
Within budget if you include cabinet material and crossover components and designed to be used in a bookcase or against the wall.
also the SpeakerDesignWorks Statement monitor has a close to wall crossover option - Statements_Monitor
just something to get the juices going.
When placed on-wall reflections from the front wall largely cancel the direct sound from the speaker drivers creating nulls at various frequencies in front at the listening position. At 90 degrees there are no nulls with the sound remaining at full loudness. If the level of the sound at frequencies around a null is raised to correct the direct sound at the listening position the reflected sound will also be raised by the same amount and be much too loud. Equalisation cannot correct an incorrect radiation pattern plus the system is likely to be non-minimum phase at frequencies around the null.
A high fidelity on wall speaker has to eliminate the front wall reflection to restore a decent radiation pattern using design not equalisation. Few on wall speakers do this including some of those posted as examples. Gathering a few examples that do is likely to be a useful but it will have to wait for tomorrow at least from me.
Would the project be open to a very wide, very thin (depth) speaker that is almost flush with the wall? Something like 40" high, 30" wide, 5" deep. Trying to get most of the advantages of a flush mount without having to remodel the room.
I am interested in participating in this project.
I am interested in participating in this project.
ABEC and AKABAK are suitable to see the effects and model any design whether it is theoretical with the use of pressure boxes or actual when the whole design is meshed with specific driving parameters.
The rest of the images are here
Dutch & Dutch 8c Review | Page 38 | Audio Science Review (ASR) Forum
Reflections, directivity and boundary interference are all within ABEC/AKABAK's BEM capability, COMSOL doesn't really offer more until it's multi physics or FEM are needed and without a commercial application or access to it in another way the price is pretty wild.
Some aspects of this have been discussed elsewhere recently and it is not just the front wall that needs to be considered, the side, floor and ceiling are often more trouble than the front wall alone. Vituix can demonstrate some of this too in a limited way. Like with any speaker absorption and directivity can help.
Dutch & Dutch 8c Review | Page 38 | Audio Science Review (ASR) Forum
Markus had some ideas of on wall in the same thread
Dutch & Dutch 8c Review | Page 39 | Audio Science Review (ASR) Forum
Attachments
For sure, EQ for the bass where there is only constructive interference. Tooles findings mentioned in ASR thread are very easy to reason / test in vituixcad as well. The interference depends on wavelenght and distance to the reflective surface (path length difference through reflection vs direct sound). and is seen as dips in various frequencies in the listening spot.
And there is nothing we can do about it is there, put the speaker in the wall, otherwise it will reflect the sound. In context of on wall design. Acoustic treatment would help to reduce reflections but I'm not sure it will ever be effective enough. Worst dip happens with 1/2wl path lenght difference (and multiples of). Just keep the speaker closer to wall than this, or reduce radiation to the wall as you'll quickly notice getting close enough to wall requires pretty much the in wall placement higher but on the low frequencies.
If you have 1/4wl distance from the transducer to the wall, listen directly ahead of the speaker so that the reflection happens directly behind the speaker, you get exactly 1/2wl from the roundtrip of the reflection from transducer to wall and back. Instead, make the reflection sum perfectly with the direct sound, coupling, consider the distance to wall <1/8wl and the path lenght difference is <1/4wl. If this was ideal for front wall reflections, then it makes sense to go cardioidish on this frequency and above.
Time delay of the front wall reflection is reduced to the listening position as well, so one might want to keep the reflections minimum. If speakers are 1.5meters out from the wall, reflection has roughly 10ms delay making it already to the late reflection category? Less than that and it is early or reflection.
The thing is, side and ceiling and floor reflections stay the same as with any speaker placement if you keep the listening distance constant, listening distance is relative to the speaker not to the wall. One is effectively moving the front wall not the speaker if looking at early reflections perspective (ignoring back of the room). In reality the sofa stays put and the speakers move, this just has to be taken account, shifts the nulls a bit. This means we can and should use same methods to battle reflections as with any loudspeaker system.
We have to remember to inspect and wonder various phenomena from the listening point.
Last edited:
This is the on-wall group project thread.
The thread is open, however it wasn't started with the intention of being an introduction to on-wall design. There are some members here with on-wall experience, including traditional methods and BEM. If your interest is to learn their methods, you'll read all about it here. If you have many questions, why not open a second thread dedicated to asking those questions, so that this one can remain on topic.
Yes sorry missing some words from my quick sentences 🙂 There is nothing we can do about it, the front wall proximity, since we are designing assuming the front wall is there. Hence we don't use in wall speakers, or assume speakers outside, I'm fully aware of this.
I admin I haven't read any lenghty threads on the subject, I've got only few years thinking and prototyping with on wall speakers behind the words. This doesn't mean I'm right, it only means I've got lot to say and ask what the others think. Could you point to on-wall speaker thread you particularly had in mind? It could be helpful for others as well. I'll paste them here as soon as I find them.
I admin I haven't read any lenghty threads on the subject, I've got only few years thinking and prototyping with on wall speakers behind the words. This doesn't mean I'm right, it only means I've got lot to say and ask what the others think. Could you point to on-wall speaker thread you particularly had in mind? It could be helpful for others as well. I'll paste them here as soon as I find them.
Yes, because people think like that, making on (or in) wall speakers. Like seen with the threads fluid posted, would anyone make speakers like that further out from the wall? No, not everyone because they have issues. The edge diffraction would be still there, sidewall would be still there. Addressing floor and ceiling is nice with MTM or array configurations but that is not specific for on wall speakers. Requirement for sweet spot would be still there. SPL and bandwidth requirements. In general, all aspects one would think of a good sound and what is needed from the loudspeaker would be there, only thing that changes is the distance between front wall and the speaker.
You have to think reversed, take best possible speaker (given other requirements and limitations than the front wall) and just tweak it enough to work close to front wall. This means constant directivity and all. Why proximity of front wall suddenly make one ignore the sidewall for example, or baffle edge diffraction? There is nothing we can do to the walls as long as we listen in rooms, so better just get used to it and figure out workaround that best addresses the walls, the loudspeaker and listening spot in the room. In fact I think there are some things we can perhaps take advantage of, like hang the speaker on the wall, height can be optimized between ceiling and floor if there is meaningful difference doing so. Flat speakers yeah, arrays why not, just don't overlook the sharp edges, internal dimension related resonances and what not. Toe in is almost impossible with wide speakers for example, but not all think this toe in is necessary. Toe in is something almost none of the (commercial) on wall speaker designs try to solve. Fluid pasted DIY speakers have toe in baked in and that is nice, it can be optimized and built for the particular room and listening position.
I'm not saying this thinking would make the best on wall speaker or that I would know what makes better on wall speaker, just makes most sense to me to think about stuff this way. Just keep in mind the issues and try to take them into account as well as possible. There is just one specialty / requirement here, distance to the front wall, and front wall is there for all speakers in the normal positioning situation as well. From the listening spot this is just path length and incident angle issue, what is the sound of the reflection and how much delayed. I think it is still possible to design the on wall speakers for same criteria like any speakers found good sounding, smooth on axis DI and power response, reduced early reflections and what have you for balanced nice sound.
You have to think reversed, take best possible speaker (given other requirements and limitations than the front wall) and just tweak it enough to work close to front wall. This means constant directivity and all. Why proximity of front wall suddenly make one ignore the sidewall for example, or baffle edge diffraction? There is nothing we can do to the walls as long as we listen in rooms, so better just get used to it and figure out workaround that best addresses the walls, the loudspeaker and listening spot in the room. In fact I think there are some things we can perhaps take advantage of, like hang the speaker on the wall, height can be optimized between ceiling and floor if there is meaningful difference doing so. Flat speakers yeah, arrays why not, just don't overlook the sharp edges, internal dimension related resonances and what not. Toe in is almost impossible with wide speakers for example, but not all think this toe in is necessary. Toe in is something almost none of the (commercial) on wall speaker designs try to solve. Fluid pasted DIY speakers have toe in baked in and that is nice, it can be optimized and built for the particular room and listening position.
I'm not saying this thinking would make the best on wall speaker or that I would know what makes better on wall speaker, just makes most sense to me to think about stuff this way. Just keep in mind the issues and try to take them into account as well as possible. There is just one specialty / requirement here, distance to the front wall, and front wall is there for all speakers in the normal positioning situation as well. From the listening spot this is just path length and incident angle issue, what is the sound of the reflection and how much delayed. I think it is still possible to design the on wall speakers for same criteria like any speakers found good sounding, smooth on axis DI and power response, reduced early reflections and what have you for balanced nice sound.
Last edited:
Yes. Wide and shallow is the most popular of the 3 approaches I am aware of to design out the front wall reflection for a high fidelity on wall speaker. It has pros and cons like the other 2 approaches. I hope this project will first analyse and quantify the performance of the alternative approaches in order to make an informed choice about which one/s to adopt. Hence the interest in simulating at the BEM level in order to handle 3D details rather than just simple sound sources.
The quickest way to get started analysing this approach would likely be with a 2D rather than a 3D approach. That is simulating a circular cylinder on an infinite wall to study the influence on the radiation pattern of varying the height to width ratio and chamfering/rounding the edge/side. Once acceptable ratios have been established squarer more practical 3D shapes for the speaker could be examined in 3D requiring more effort to setup and run.
Before that we need to sort out appropriate software and settle on how to store the documentation and simulation data for a group project.
Thanks. Good to see these simulations and links to prior work.
COMSOL does have advantages over ABEC/AKABAK like better accuracy, faster for large simulations such as speakers in rooms, ease of use and support, hooking in to other software, runs on platforms I could use for crunching numbers (dual Xeon) not just Windows (laptop with Athlon),... but it also has disadvantages like availability which isn't one of ABEC/AKABAK strengths either.
There is clearly a task to examine our options for solving the 3D linear wave equation for a group project like this. Given my background in this sort of thing I guess it is one for me.
We have a related decision to make concerning the type of software to use within the group project. Open source software that everyone can freely use and fix/tweak is obviously preferable but it often isn't available in easily usable forms and can be poor quality.
Should we embrace closed source but free software? I would suggest yes.
Should we embrace closed source commercial software? I would suggest only if we cannot get free software to work adequately. It is not going to be as well supported, slick and easy to use as commercial software so there is a judgement call involved.
Thoughts?
I am unfamiliar with this software due to the Windows requirement but it is currently widely used and seems to be liked. Can someone summarise it's capabilities with respect to on-wall speakers?
^ Advantage of VituixCAD to me is that it is real time! I mean you can change the simple room model by mouse wheel and see the interference move around at the listening spot. You can inspect floor, ceiling, side and front wall reflections in relation to the speaker and listening spot but it does include the one corner of the room only so there is no means to see if late reflections from the back affect. This is a good thing I believe, we need data about the front wall proximity.
The simulated speaker works with real or generated measurements. It is just fast and easy to whip up and tune around to see what affects what. But, it is not very accurate to reality other than being simple math based on heights, widths, angles toe-ins, path lengths and what not to display sum of direct and first reflections at single observation point as single estimated in room response line. But for example it is pretty straight forward to test out if there is optimal height for a bandwidth in relation to floor and ceiling and listening spot (height) or not! Or what is effect of floor (and/or ceiling) bounce and does woofer height affect on that or what crossover point would be nice between woofer and mid try and smoothen the interference.
Lots of generalized information can be gained, like if something seems to be possible or impossible or what seems to be the phenomenon and what affects what in a very quick way. But, one has to be able to ask questions and figure out how to test them in simple form and if the test gives any valid information or is it something else. It is very interesting to see things move even if there wasn't enough understanding to interpret what that might be. First thing that will pop into mind is how it is all about wavelengths 😀
The simulated speaker works with real or generated measurements. It is just fast and easy to whip up and tune around to see what affects what. But, it is not very accurate to reality other than being simple math based on heights, widths, angles toe-ins, path lengths and what not to display sum of direct and first reflections at single observation point as single estimated in room response line. But for example it is pretty straight forward to test out if there is optimal height for a bandwidth in relation to floor and ceiling and listening spot (height) or not! Or what is effect of floor (and/or ceiling) bounce and does woofer height affect on that or what crossover point would be nice between woofer and mid try and smoothen the interference.
Lots of generalized information can be gained, like if something seems to be possible or impossible or what seems to be the phenomenon and what affects what in a very quick way. But, one has to be able to ask questions and figure out how to test them in simple form and if the test gives any valid information or is it something else. It is very interesting to see things move even if there wasn't enough understanding to interpret what that might be. First thing that will pop into mind is how it is all about wavelengths 😀
Attachments
Last edited:
There are 3 things that I am aware of that have been done in the past and there may be more! 1) wide and shallow is by far the most common, 2) cardioid-type radiation patterns that radiate little-to-none at the wall are currently fashionable, 3) woofers on the side, midrange on front with appropriate crossover frequency is perhaps the most elegant.
Do you fancy a background review pulling previous examples together? There are plenty of 1) to find, 2) should be straightforward given it is currently back in fashion, 3) might require a bit more digging. The Allison speakers from the 70s with accompanying publications might be a good place to start. I recall seeing a Scandivian example, Finnish I think, a few years ago but it didn't seem to catch on. There are not many examples but there are some. Of course there are a great many conventional speakers claiming to be suitable for on-wall but that's the modern world.
I have looked up what the previous group project did to handle the group documentation and simulation files. They appear to have used github but there seems to be only final documentation, no simulation files and only 2 members. Was there activity elsewhere (apart from diyaudio obviously).
How should we handle things?
One possibility might be:
- setup github project
- post relevant text here first
- copy text to github as markdown file
- edit and revise markdown file on github
This would keep the relevant posts/articles containing the details visible rather than lost somewhere in a lengthy thread. It would also notify people of new material when it appears.
Thoughts and experiences of other group projects?
You asked if it could be done with ABEC/AKABAK and I showed that it could. Anybody can get a fully functioning student licence for either program for free on the condition that it is used for non commercial purposes only. What you choose is up to you as my interest in this project is purely academic.
The real time aspect of Vituix is great, the room simulator is my least favourite part of it, basic questions on room placement can be answered but not much more. A BEM room sim is better but the more generic it is the less valid it will be for any given setup. I agree with what you said before, and I said the same thing above, that the front wall is actually the least problematic to deal with in a good on wall design, bad ones make everything a problem 🙂
Yes this is what I was thinking, there is nothing we can do to the front wall because it is the basis for the design and assumed to be there from the get go 🙂 Writing in a hurry. We of course figure out how to address the wall in the design. Point was, the wall is there and the question is how we can address it in the speaker design and there are various ways to do it as you have listed. In addition lets not forget room acoustic treatment which could be utilized as well. There might be possibility for insulation material as part of the speaker, between it and the wall.
But, as soon as one figures out wavelengths and effectiveness of the insulation in relative thickness I never looked further into it and figured out cardioidish + waveguides and anything that prevents sound to the front wall seems the most effective way to reduce effect of the wall. There can be more ways to address the wall proximity, like additional acoustic stuff etc. Not ruling anything out.
It is the wavelengths that bite back and dictate. For the lowest frequencies, the long wavelengths, we are effectively coupled to the wall. Above that there is transition to shorter wavelengths we have the dips and nulls happening where the 1/2wl cancellation and multiples of happen, then the wavelenths are much shorter than the distance of the transducer to the wall and stuff turns into comb filter territory, which I believe hearing system handles just fine. For the high frequencies, comb filtering, some acoustic treatment could work nicely. For the transition region there is nothing that fits between the speaker and the wall that would help, there would always be transition region and here the cardioidish or some cancellation is almost only thing we can do. Shorter than 1/8wl distance we just couple to the wall and anything goes.
These very rough features stay with the system no matter what: coupling > transition to interference > interference which turns into comb filter. The distance to the wall (path length difference of direct and reflected sound) says where the transition happens from coupling to comb filter. The same thing happens with all the boundaries with all speaker positions, just the frequency where we transition changes with distance to the boundary. When we get closer to 0 distance, the transition happens above hearing range.
This kind of stuff is easy to see and think upon in the VCAD 😉 And as soon as you enable both walls, floor and ceiling you'll notice the interference doesn't line up very well no matter the speaker / listening position... Well, at the low mids (where the room modes are not dominating anymore) it looks like careful positioning can average out the interference at listening spot (ignoring rest of the room, furniture and stuff). Roughly about 3kHz it looks like the comb filtering is smoothed in the brain. Between about 500Hz and 3kHz is a region that seems to be a mountain range no matter what. It is here you want the interference to minimum to make the mountains and canyons turn into to hills and ditches. This is true for any speaker though, not just on wall speaker. I made up the 3kHz, took some rough number. Might be 6kHz, or something.
Well, if you tune the VCAD room model so that your speaker is at the ceiling or floor corner there is not much interference. Just cram a fullranger there and call it a day 😉 Or figure out what the hearing system is capable of hearing. Maybe there is frequency range we could exploit, where all problems can be piled on and just rely brain cleaning it all up. ? I haven't looked too much into hearing system other than the basics that popup in the forums every now and then.
Last edited: