It depends to a significant extent on the acoustics of the room and the position of the speakers. In a large symmetric room with good acoustics and the speakers pulled out into the room the setup that is likely to work best is a widish beamwidth with a gentle narrowing with the rate detemined by the amount of absorption in the room. Though this depends to some extent on preference concerning concert hall vs studio acoustics. On the other hand in a small asymmetrical room with poor acoustics and the speakers placed against a wall such a speaker will not work well whereas a cardioid radiation pattern will have no problem being near a wall and the room modes will be less driven. A few years ago at an audio show using hotel bedrooms there was a pair of D&Ds arranged to play across the room. They worked noticeably better than conventional speakers playing down the room in the other rooms with noticeably less room boom.
There is an interesting discussion about cardioids and the their coupling to room modes
https://audiosciencereview.com/foru...ectivity-patterns-couple-to-room-modes.45518/
For a summary, I am copy-pasting the relevant part here from the opening post:
"Monopoles and dipoles couple to the room in opposite ways, so a good location for a monopole is probably a bad location for a dipole, and vice versa.
To get the desired effects from dipoles and cardioids, they need to be oriented to align with the room modes.
Cardioids are significantly less affected by room modes.
Cardioids and dipoles are inefficient sound radiators compared to monopoles."
More interesting material is also found in subsequent posts, where it says:
"cardioid provides faster modes decay and more even room modes excitation"
https://audiosciencereview.com/foru...terns-couple-to-room-modes.45518/post-2215146
https://audiosciencereview.com/foru...ectivity-patterns-couple-to-room-modes.45518/
For a summary, I am copy-pasting the relevant part here from the opening post:
"Monopoles and dipoles couple to the room in opposite ways, so a good location for a monopole is probably a bad location for a dipole, and vice versa.
To get the desired effects from dipoles and cardioids, they need to be oriented to align with the room modes.
Cardioids are significantly less affected by room modes.
Cardioids and dipoles are inefficient sound radiators compared to monopoles."
More interesting material is also found in subsequent posts, where it says:
"cardioid provides faster modes decay and more even room modes excitation"
https://audiosciencereview.com/foru...terns-couple-to-room-modes.45518/post-2215146
Ok..
Out of curiosity, I modelled a dipole cardioid source (using the same ascilab-stylewoofer module + cardioid sica coax) and did some positioning experiments
Here is how the setup looks
Note that it is just the DSP settings that have changed compared to the earlier full range cardioid speaker.
This one seems to be much more tolerant of positioning if the front wall has moderate absorption and a relatively absorbing carpet on floor 🙂
Out of curiosity, I modelled a dipole cardioid source (using the same ascilab-stylewoofer module + cardioid sica coax) and did some positioning experiments
Here is how the setup looks
Note that it is just the DSP settings that have changed compared to the earlier full range cardioid speaker.
This one seems to be much more tolerant of positioning if the front wall has moderate absorption and a relatively absorbing carpet on floor 🙂
Attachments
@vineethkumar01 - Your work here is the first time I have seen a this sort of room mode analysis used to compare monopole to cardioid speakers. It is also very useful and relevant because of your personal experience with all of these speakers. Very informative, and thank you!
@andy19191 - I have never changed the title of one of my design/build threads, even when the final design became significantly different than the original concept. You can forever edit the opening post to add comments, updates, graphics, even an index. I would advise against creating a new thread because this one has a lot of useful discussion buried in it.
j.
@andy19191 - I have never changed the title of one of my design/build threads, even when the final design became significantly different than the original concept. You can forever edit the opening post to add comments, updates, graphics, even an index. I would advise against creating a new thread because this one has a lot of useful discussion buried in it.
j.
If a new thread was created this thread would be linked in the OP. Other than waiting till the project has a name I'm happy with either option. Currently I think the score is 1-1 which doesn't help.
Just to be clear and not intended to be semantics but this is not modal analysis but reflection analysis. How the rooms boundary positions will create interference. Modes exist on top of this.
Cardioid patterns do very well to eliminate the rear wall as a problem but most designs are wide radiating to the sides and everywhere else so side wall and floor ceiling problems are still there.
Dipoles do well at reducing side wall problems from the side nulls.
Line arrays do well with floor and ceiling.
All of these things have an effect on perception that goes beyond what happens to the frequency response.
My post suggesting options for a spec a page or two back hasn't generated much discussion so I will pencil in a spec and ask for objections/support.
Floor standing to optimise the use of space. Separate woofer and mid/tweeter cabinets enabling different design approaches for each cabinet, passive isolation between them, supports a modular approach.
Designed to be positioned near or against the front wall. Given the use of an active crossover positioning out in the room should also work well though possibly not as well as it might had it been the only position.
Clean output at standard levels (80-85 dB average) at 3-4m with low frequency extension to 30-40 Hz. Will enable supporting a distributed sub system or working reasonably standalone without sub support.
The radiation pattern requires research but tentatively omnidirectional below 80-100 Hz with a narrowing beamwidth above. Lean towards clarity rather than spaciousness.
The build complexity to be pitched at inexperienced speaker DIYers that intend to continue with the hobby. A level above inexperienced one-off kit builders on the kitchen table but below experienced speaker DIYers with well equipped workshops.
To use good standard range drivers and components. Not budget ones and not premium/prestige ones. Budget for drivers £500-1000 increasing by £500 if an active cardioid woofer module.
External active crossover and amplifiers to support a modular approach.
Proposed initial modules subject to discussion:
An active cardioid woofer module driven by a pair of DSP controlled amplifier channels can also radiate as a monopole enabling a cardioid radiation pattern to be assessed by listening.
A cardioid midrange has been mentioned more than once and is a candidate for a top module. Separate waveguides on tweeter and midrange seem to be the conventional way to control the radiation pattern in high technical performance speakers and so is another candidate for a top module.
Start with an active crossover. If there is an interest in a passive crossover address it later.
Suggestions? What has been missed? What isn't quite right?
Floor standing to optimise the use of space. Separate woofer and mid/tweeter cabinets enabling different design approaches for each cabinet, passive isolation between them, supports a modular approach.
Designed to be positioned near or against the front wall. Given the use of an active crossover positioning out in the room should also work well though possibly not as well as it might had it been the only position.
Clean output at standard levels (80-85 dB average) at 3-4m with low frequency extension to 30-40 Hz. Will enable supporting a distributed sub system or working reasonably standalone without sub support.
The radiation pattern requires research but tentatively omnidirectional below 80-100 Hz with a narrowing beamwidth above. Lean towards clarity rather than spaciousness.
The build complexity to be pitched at inexperienced speaker DIYers that intend to continue with the hobby. A level above inexperienced one-off kit builders on the kitchen table but below experienced speaker DIYers with well equipped workshops.
To use good standard range drivers and components. Not budget ones and not premium/prestige ones. Budget for drivers £500-1000 increasing by £500 if an active cardioid woofer module.
External active crossover and amplifiers to support a modular approach.
Proposed initial modules subject to discussion:
- active cardioid woofer module
- coaxial in waveguide tweeter/mid module
- raspberry pi based active crossover module
- AVR and/or stack of stereo amplifiers and/or whatever
An active cardioid woofer module driven by a pair of DSP controlled amplifier channels can also radiate as a monopole enabling a cardioid radiation pattern to be assessed by listening.
A cardioid midrange has been mentioned more than once and is a candidate for a top module. Separate waveguides on tweeter and midrange seem to be the conventional way to control the radiation pattern in high technical performance speakers and so is another candidate for a top module.
Start with an active crossover. If there is an interest in a passive crossover address it later.
Suggestions? What has been missed? What isn't quite right?
Yes of course you are correct... Thanks!
Andy - overall your initial spec looks good to me.
I agree here. I have always thought that the large stand-mount speakers were rather impractical, though it is the current fashion. Building an attractive stand requires some work, and the space underneath is under-utilized. An 80 liter retro-box raised up on a 20" high stand consumes a lot of space in a room, as much as a 120 liter floor standing speaker.
Again, this is a good choice, and it is wise to simply specify it in advance. It also keeps the weight down on each component.
At first I thought that this one might be difficult to live up to, but as long as we are willing to sacrifice some aesthetics it is doable... High performance AND beauty are a combination that usually requires advanced fabrication skills/processes. But if we accept that form follows function, and that might mean a form which is a bit funky, then I believe we can meet this requirement.
Again, good to specify this right away.
Eventually we will need to define a no-kidding minimum requirement. F3 <= xy, F10 <= yz for example... "30 - 40 Hz" is a big range, and there is a big difference in capability between getting an F3=40 vs F3=30. The limiting factor in SPL is almost always in the deep bass, so we need to eventually specify some minimum SPL at some xy Hz... No need to nail all this down yet. It would probably be a good idea to explore what is possible and realistic within our other constraints (particularly cabinet size and driver costs).
j.
Depending on woofer crossover frequency and distance to front wall the cardioid woofer might not even be necessary, imo.
But it is interesting nontheless.
Midrange radiation pattern control via resistive cardioid enclosure may help reducing necessary waveguide size.
To just point out some possibilities regarding aesthetic, driver-related and other considerations.
In all the pattern control study-related sims I have done and given in the above posts, I have just considered a coax (passive cardioid) + two woofers (one placed on the front of the cabinet and one on the back, like the ascilab-woofer). All the pattern control below 400ish Hz was done by changing the DSP settings applied on the two woofers and coax in the passive cardioid box. So the concept can be applied from desktop/standmount-sized speakers to floorstanding or monkey box type speakers.
1) The single speaker with passive cardioid coax (with an easy-to-do passive crossover, at least with SICA drivers) + two woofers in the cabinet can give the following radiation pattern configurations just by changing DSP settings: omni (bass)+cardioid, cardioid(bass)+cardioid, dipole(bass)+ cardioid
2) The speaker can look like the below (with an additional back woofer in these cabinets) with bigger coax+woofer combinations. As a separate coax top + woofer module as shown in the first pic or as one integrated box with internal partitions to separate the woofer cab and passive cardioid coax cab (all pictures are from some excellent projects by @airvoid 🙂 )
After all, it can get that 'monkey box' looks if needed 😉
For those interested in slimmer cabinets and taller floor standers, the ascilab-cardioid speaker form factor can be adopted..
Whichever way, the dual woofer concept remains the same (The box width-to-depth ratio, baffle treatment etc need to be agreed upon based on woofer module volume requirements/directivity transitions in the lower-midrange to midrange and any other considerations ). Here is a closer look into the baffle for the above speaker (just for fun and to appreciate the attention to detail that went into this build 🙂 )
2) The above concept speaker needs 3 DSP output and amp channels (per side) with a passive crossover on the passive cardioid top module. This can relax requirements on amp and DSP channels. And with an 8-channel output DSP setup, two channels can be reserved for dedicated subs later, if needed.
3) A sweet spot would be to have a nice 8inch coax + 8-10 inch woofers. And for those who don't like coaxes, it can always be replaced by a waveguided tweeter+passive cardioid midrange or non-waveguided tweeter + passive cardioid midrange, etc. The choices of woofers and the pattern control can be done as per max SPL/distortion/ low frequency reach requirements. In fact, if anyone is allergic to any potential low frequency passive cardioid related harmonic distortion rises, one can always replace passive cardioid mid with normal mid and side drivers (like ascilab speakers/kii speakers) for pattern control in the midrange 😉
Those additional 2 DSP channels that we have reserved by making a passive crossover-based mid-tweeter/coax crossover (out of the 8output channels from DSP) can then be put to use with the side woofers for required pattern control.
In short I feel this concept can work to some extent in addressing some of the boundary interference-related concerns along with speaker positioning and some absorption.
Also, with this kind of speaker, one can have a modular, configurable/application-driven directivity speaker.. How cool is that 😉 and how many manufacturers make such a speaker (as of now 🙂 ).
One speaker-many avatars at the flick of a switch (or DSP preset) 😛
In all the pattern control study-related sims I have done and given in the above posts, I have just considered a coax (passive cardioid) + two woofers (one placed on the front of the cabinet and one on the back, like the ascilab-woofer). All the pattern control below 400ish Hz was done by changing the DSP settings applied on the two woofers and coax in the passive cardioid box. So the concept can be applied from desktop/standmount-sized speakers to floorstanding or monkey box type speakers.
1) The single speaker with passive cardioid coax (with an easy-to-do passive crossover, at least with SICA drivers) + two woofers in the cabinet can give the following radiation pattern configurations just by changing DSP settings: omni (bass)+cardioid, cardioid(bass)+cardioid, dipole(bass)+ cardioid
2) The speaker can look like the below (with an additional back woofer in these cabinets) with bigger coax+woofer combinations. As a separate coax top + woofer module as shown in the first pic or as one integrated box with internal partitions to separate the woofer cab and passive cardioid coax cab (all pictures are from some excellent projects by @airvoid 🙂 )
After all, it can get that 'monkey box' looks if needed 😉
For those interested in slimmer cabinets and taller floor standers, the ascilab-cardioid speaker form factor can be adopted..
Whichever way, the dual woofer concept remains the same (The box width-to-depth ratio, baffle treatment etc need to be agreed upon based on woofer module volume requirements/directivity transitions in the lower-midrange to midrange and any other considerations ). Here is a closer look into the baffle for the above speaker (just for fun and to appreciate the attention to detail that went into this build 🙂 )
2) The above concept speaker needs 3 DSP output and amp channels (per side) with a passive crossover on the passive cardioid top module. This can relax requirements on amp and DSP channels. And with an 8-channel output DSP setup, two channels can be reserved for dedicated subs later, if needed.
3) A sweet spot would be to have a nice 8inch coax + 8-10 inch woofers. And for those who don't like coaxes, it can always be replaced by a waveguided tweeter+passive cardioid midrange or non-waveguided tweeter + passive cardioid midrange, etc. The choices of woofers and the pattern control can be done as per max SPL/distortion/ low frequency reach requirements. In fact, if anyone is allergic to any potential low frequency passive cardioid related harmonic distortion rises, one can always replace passive cardioid mid with normal mid and side drivers (like ascilab speakers/kii speakers) for pattern control in the midrange 😉
Those additional 2 DSP channels that we have reserved by making a passive crossover-based mid-tweeter/coax crossover (out of the 8output channels from DSP) can then be put to use with the side woofers for required pattern control.
In short I feel this concept can work to some extent in addressing some of the boundary interference-related concerns along with speaker positioning and some absorption.
Also, with this kind of speaker, one can have a modular, configurable/application-driven directivity speaker.. How cool is that 😉 and how many manufacturers make such a speaker (as of now 🙂 ).
One speaker-many avatars at the flick of a switch (or DSP preset) 😛
A cardioid woofer module could be active or passive (or hybrid though perhaps not for a first go). Active is tunable in blending monopole and cardioid, moving nulls around whereas passive is more fixed. Active is more expensive. Passive likely more fun to design. Are we leaning towards active?
What size drivers do we need to meet the clean SPL requirement with a cardioid configuration?
For an active cardioid how best to arrange the drivers? Possibly one on the front but how many on the sides and rear? Is a woofer bottom and tweeter/midrange top a good configuration for a cardioid?
What size drivers do we need to meet the clean SPL requirement with a cardioid configuration?
For an active cardioid how best to arrange the drivers? Possibly one on the front but how many on the sides and rear? Is a woofer bottom and tweeter/midrange top a good configuration for a cardioid?
Thanks @vineethkumar01, the two piece speakers are actually 12” coaxial tops with 12” isobaric woofers firing at the floor.
For my next adventure I’m working on this:
For my next adventure I’m working on this:
Regarding the general direction of this project, I feel that we should finalize who this project is for as @andy19191 had asked in an earlier post. This would help us decide the speaker form factor, cost/cost cutting involved, and maybe other details involved. From the looks of it so far, it looks like we have a target budget and an inclination to extract very good technical performance. Aesthetical considerations should fall in place as the form factor is fixed.
Given the potential use of DSP etc, it seems like it is intended at least for someone who already has some experience doing DIY speaker projects. Although with good enough documentation, it could be attempted even by a newbie.
So I am just asking again whether we want:
1) A stand-mount-sized speaker with pattern control capabilities from 100Hz up
2) A floor-stander-sized speaker with pattern control capabilities from 100Hz up
3) A monkey coffin-sized speaker with pattern control capabilities from 100Hz up
At least one of these choices should drive the bass cabinet volumes available, directivity control strategy to be put in place, low-frequency reach, and max SPL capabilities with/without pattern control to low frequencies.
Getting pattern control down to 300Hz or so is relatively easy with an 8-inch woofer + tweeter in a passive cardioid box or an 8-inch coax driver in a passive cardioid box. Even I have been able to do this with just a low-cost 5.5-inch sica coax in a passive cardioid enclosure. The THD at 86dBSPL/1m was good enough even with it. The Sigberg audio Manta project extracts very respectable distortion figures out of the same coax crossed over at 600Hz.
Given the availability of DSP/DSP amps these days, I think we should include one in the project. But there can also be a potential (even reduced capability) passive version. For example, I could do this with the cardioid coax speaker (this was never a final crossover for this project or optimized to full extent and was just a proof of concept)
Regarding the best way to arrange the drivers, I think people with better experience simulating and building these kinds of concepts can suggest. Requesting @fluid, @nc535, @tmuikku to comment about these aspects.
But from some preliminary simulations, etc, I think for better pattern control, side drivers might help better than back-mounted drivers. It makes more of a difference in higher frequencies than lower frequencies I think.
Regarding sizes of drivers and example woofers that can be used for a bass module, I guess the options are driven by cost and cabinet requirements.
1) For example, one can have up to 4 drivers per speaker and still have overall reasonable costs for the woofer module with this driver (Thanks to @nc535 who alerted me about the existence of such a driver sometime back)
https://www.parts-express.com/GRS-8...VRdRp-Eyw4mNohOrnUuYyzlj2bjsc8M6U5pvX9-so3qsw
PS: Dont ask why cardioid speakers these days.. It seems to be in vogue even if all/most kinds of considerations that apply to regular speaker placement applyies to cardioid speakers too 😛
Given the potential use of DSP etc, it seems like it is intended at least for someone who already has some experience doing DIY speaker projects. Although with good enough documentation, it could be attempted even by a newbie.
So I am just asking again whether we want:
1) A stand-mount-sized speaker with pattern control capabilities from 100Hz up
2) A floor-stander-sized speaker with pattern control capabilities from 100Hz up
3) A monkey coffin-sized speaker with pattern control capabilities from 100Hz up
At least one of these choices should drive the bass cabinet volumes available, directivity control strategy to be put in place, low-frequency reach, and max SPL capabilities with/without pattern control to low frequencies.
Getting pattern control down to 300Hz or so is relatively easy with an 8-inch woofer + tweeter in a passive cardioid box or an 8-inch coax driver in a passive cardioid box. Even I have been able to do this with just a low-cost 5.5-inch sica coax in a passive cardioid enclosure. The THD at 86dBSPL/1m was good enough even with it. The Sigberg audio Manta project extracts very respectable distortion figures out of the same coax crossed over at 600Hz.
Given the availability of DSP/DSP amps these days, I think we should include one in the project. But there can also be a potential (even reduced capability) passive version. For example, I could do this with the cardioid coax speaker (this was never a final crossover for this project or optimized to full extent and was just a proof of concept)
Regarding the best way to arrange the drivers, I think people with better experience simulating and building these kinds of concepts can suggest. Requesting @fluid, @nc535, @tmuikku to comment about these aspects.
But from some preliminary simulations, etc, I think for better pattern control, side drivers might help better than back-mounted drivers. It makes more of a difference in higher frequencies than lower frequencies I think.
Regarding sizes of drivers and example woofers that can be used for a bass module, I guess the options are driven by cost and cabinet requirements.
1) For example, one can have up to 4 drivers per speaker and still have overall reasonable costs for the woofer module with this driver (Thanks to @nc535 who alerted me about the existence of such a driver sometime back)
https://www.parts-express.com/GRS-8...VRdRp-Eyw4mNohOrnUuYyzlj2bjsc8M6U5pvX9-so3qsw
PS: Dont ask why cardioid speakers these days.. It seems to be in vogue even if all/most kinds of considerations that apply to regular speaker placement applyies to cardioid speakers too 😛
DSP is straightforward for beginners if they want to learn. Hardware like a raspberry pi has strong support for beginners and much of the addon hardware is push fit like lego. Whether everything will remain beginner friendly as issues arise remains to be seen. The processor seemed to get hot installing (but I wasn't monitoring temperature) and may well need some form of cooling. The DAC hat gets in the way of this to some extent. Some degree of fiddling seems likely.
The suggestion earlier was floor-standing with a midrange/tweeter cabinet on top of a woofer cabinet. However, since starting to sim the radiation patterns some of the natural driver arrangements (e.g. Kii) are not a good fit for this configuration. So tempted to let form follow function if in conflict with modularity.
Distortion at 86 dB average at 1m is not of much relevance. What is are clean peaks at 3-4m.
By DSP amp do you mean a board with power supply, power amplifier and DSP on it? Placed in the speaker? Something like hypex modules or something else?
Linear acoustic BEM (akabak, acousto, bempp,...) can be useful in sorting driver configuration and perhaps some in-room issues. A better job can likely be done with an acoustic FEM code due to modal analysis and sound absorption through foam, furniture and the like. Structural FEM for cabinet vibration. Nonlinear rather than linear 0D/1D analysis for the drivers would enable us to include better distortion levels in the simulations and as more driver manufacturers publish Klippel data this is becoming viable. My hope is the project will include much of this.
A large woofer like a 15" in a shallow cabinet can produce a cardioid like directivity down to about 300Hz all on it's own with the rear box diffraction creating the interference signal.
The time between the front sound and the interfering sound determines the wavelength where the cancellation and resulting directivity will occur. So side slots or side drivers can have effects to higher frequencies than rear drivers, the size and placement of the slot or driver will dictate it's upper limit.
In an active setup the group delay of an IIR lowpass on the side driver can be enough to get good even directivity. In a resistive slot the path from the rear of the cone through any absorptive material decides the same. This is difficult to simulate accurately without a detailed FEM model, but you can get pretty close with AKABAK if you accept some margin for error.
Active side drivers are the most flexible but with greater cost and complexity.
Leaky boxes and active cardioids can actually increase output in the baffle step region over a regular single forward facing driver alone. Below this often around 200 to 300Hz the extra 6dB per octave rolloff kicks in and you start to lose output. Linkwitz has a spreadsheet where you can compare the output needs of a dipole woofer vs a boxed one. A true cardioid has the same DI as a dipole so that spreadsheet would be a close facsimile. As you go down in frequency the needs get more and more to replace the output lost from interference.
Cardioid returning to monopole at 100Hz as suggested gives something like a 3dB loss in max output compared to a pure monopole. This is simple enough to do in an active setup, I posted a Vituix simulation of it somewhere with a front and rear facing woofer.
I can see the possible value of going cardioid over the whole range if you can live with the size requirement, I am struggling to see the value in the other types of cardioid unless you really must have the speaker right up against the front wall. In that scenario side drivers or slots makes more sense.
The D&D 8C is a clever solution as it uses slots for the midrange directivity to stop rear wall interference and cuts to rear facing subs at 100Hz to leverage coupling with the boundary. It is between 100Hz and around 200 or 300Hz that excursion will increase in the mid woofer over what it would be without the slot if the same SPL is required.