Directiva r2 project: market requirements gathering

[sarumbear]

Requirements for Directiva r2

1. The r2 speaker should be active and simple to build. It will utilize a bass module and a upper monitor module. Use of commercially available cabinets (notably for the monitor) is preferred if possible.
2. The design should be reusable and extensible to allow for design variants. Minidsp 2x4 HD will be used as the active crossover for design and costing purposes. Seas DXT tweeter is to be re-used from r1 unless design team finds a less expensive substitute(s) with better performance.
3. Bass extension is targeted to be comparable to r1, but the bass module should provide additional bass output to allow for use in rooms up to 150 cubic meters. Listening distance is targeted to be 3-5 meters. The monitor module itself can have less bass extension than r1 to allow for use with existing subwoofers or other bass modules.
4. The bass module should be designed for accurate bass reproduction and low distortion. A sealed cabinet design is preferred.
5. The design may include passive crossover circuitry to help reduce the number of active channels and allow for a passive version of the monitor section.
6. Directivity should exceed r1. The primary requirement is for more consistent directivity over a more extended frequency range. Extensibility of bass module design may consider use of variable directivity, but this is NOT a primary requirement for r2. For the monitor, custom waveguides should be avoided.
7. The cost target for the drivers and crossover would be approximately $1600 US retail per pair. This cost includes the cabinets, and active crossover but NOT the amplification.
8. The combined cabinet volumes should be 60-75 liters or smaller. Combined height should not exceed 1.2m
9. The monitor module should be no wider than 50 mm more than the midwoofer used and the bass module should be wide enough to allow monitor to be securely fitted on top. The target width of the bass module is a maximum of 260 mm.
10. The electronics should be external, but the modules may consider added volume (above stated target) for those who prefer internal active electronics.
11. The design tools should be free but flexible enough to allow external data import/export. Plan to continue with VituixCAD and REW.
12. The design should be openly shared and FOR NON-COMMERCIAL USE ONLY.

As the main Directiva r2 stakeholder and project lead, questions or comments should be directed to me.

The deadline for comments is October 31 and plan to finalize the requirements by November 1.

Thanks!

Rick

I went through the list and suddenly realised that you are specifying our Silver 5L & BH bass augmentor system.

However, I assumed that since 90s, speaker design can be a bit more risky. (Joke.
Your list is very good.)

 

[Lbstyling]

Requirements for Directiva r2

1. The r2 speaker should be active and simple to build. It will utilize a bass module and a upper monitor module. Use of commercially available cabinets (notably for the monitor) is preferred if possible.
2. The design should be reusable and extensible to allow for design variants. Minidsp 2x4 HD will be used as the active crossover for design and costing purposes. Seas DXT tweeter is to be re-used from r1 unless design team finds a less expensive substitute(s) with better performance.
3. Bass extension is targeted to be comparable to r1, but the bass module should provide additional bass output to allow for use in rooms up to 150 cubic meters. Listening distance is targeted to be 3-5 meters. The monitor module itself can have less bass extension than r1 to allow for use with existing subwoofers or other bass modules.
4. The bass module should be designed for accurate bass reproduction and low distortion. A sealed cabinet design is preferred.
5. The design may include passive crossover circuitry to help reduce the number of active channels and allow for a passive version of the monitor section.
6. Directivity should exceed r1. The primary requirement is for more consistent directivity over a more extended frequency range. Extensibility of bass module design may consider use of variable directivity, but this is NOT a primary requirement for r2. For the monitor, custom waveguides should be avoided.
7. The cost target for the drivers and crossover would be approximately $1600 US retail per pair. This cost includes the cabinets, and active crossover but NOT the amplification.
8. The combined cabinet volumes should be 60-75 liters or smaller. Combined height should not exceed 1.2m
9. The monitor module should be no wider than 50 mm more than the midwoofer used and the bass module should be wide enough to allow monitor to be securely fitted on top. The target width of the bass module is a maximum of 260 mm.
10. The electronics should be external, but the modules may consider added volume (above stated target) for those who prefer internal active electronics.
11. The design tools should be free but flexible enough to allow external data import/export. Plan to continue with VituixCAD and REW.
12. The design should be openly shared and FOR NON-COMMERCIAL USE ONLY.

As the main Directiva r2 stakeholder and project lead, questions or comments should be directed to me.

The deadline for comments is October 31 and plan to finalize the requirements by November 1.

Thanks!

Rick

Interesting on the sealed cab design. I think your right to do it given the price of amplification and the amount of problems resonances give small boxes.

What's the max size driver you can fit on the side face of a 26cm wide denovo bass cab?

I like the wording for point 6. Reminded me that the DXT tweeter may be surpassed also, at some point by somthing like Brandon's waveguided SB26... We just need Denovo to start producing the waveguide...

 

[Lbstyling]

If you're willing to accept an unusual baffle design, you can modify the Directiva to control dispersion down to about 400Hz.
The Purifi woofer can be replaced by a cheap but also very good driver.

The horizontal -6dB limit could then look like this for the tweeter and woofer.
View attachment 161152
A crossover frequency around 1.8kHz would provide a very even dispersion down to 400Hz.

The normalized sonogram for the woofer alone, would look like this:
View attachment 161153
For even deeper control of radiation, one must resort to alternative concepts - such as slots, OB, ...

Do you have any rough drawings of what the 'unusual baffle' design might look like? I wonder if is has any measurement correlation to effects (path lengths) of positioning cardioid side vents.

 

[Rick Sykora]

Interesting on the sealed cab design. I think your right to do it given the price of amplification and the amount of problems resonances give small boxes.

What's the max size driver you can fit on the side face of a 26cm wide denovo bass cab?

I like the wording for point 6. Reminded me that the DXT tweeter may be surpassed also, at some point by somthing like Brandon's waveguided SB26... We just need Denovo to start producing the waveguide...

The 26 cm bass module width does not allow for much more than an 8 inch woofer on the front.

 

[sarumbear]

The 26 cm bass module width does not allow for much more than an 8 inch woofer on the front.

Use two, you get the equivalent of a 12”

 

[Lbstyling]

The 26 cm bass module width does not allow for much more than an 8 inch woofer on the front.

If it's this one- https://www.soundimports.eu/en/denovo-audio-300-7066.html

11 inches deep.

10 inch side firing bass driver crossed at 100hz or so.

Or dual opposed 10 inch drivers (one either side) and staggered in height to give space in the cabinet for the magnets to miss each other.

Unless a good 10inch bass driver with a mounting depth of less than 4 inches exists.....

I have a distant memory of one being used on HTGuide forum by Doc Brown.....?

 

[voodooless]

The 26 cm bass module width does not allow for much more than an 8 inch woofer on the front.

You might be able to fit a 10” woofer. For instance Faital 10PR320 is just 1mm to wide. Assuming a paint job or veneer, you should be good. Visually it might even be interesting, but for sure it’s not for everybody. Also you cannot have roundovers.

 

[Lbstyling]

You might be able to fit a 10” woofer. For instance Faital 10PR320 is just 1mm to wide. Assuming a paint job or veneer, you should be good. Visually it might even be interesting, but for sure it’s not for everybody. Also you cannot have roundovers.

Don't have to have roundovers all the way down the cabinet though do you?.....

 

[voodooless]

Don't have to have roundovers all the way down the cabinet though do you?.....

Nope, you don’t, but some people might like it.

 

[ctrl]

Do you have any rough drawings of what the 'unusual baffle' design might look like? I wonder if is has any measurement correlation to effects (path lengths) of positioning cardioid side vents.

Something like this - using two 4'' midwoofer.



The trick is the ratio of baffle width and cabinet depth. The depth of the cabinet should usually be between 80-200mm. I'm currently working on three other designs using this technique to control directivity.
The finished simulations have been sitting on my hard drive for years, but I usually don't have the time or the motivation to implement the projects. But I'm working on them at the moment.

Since virtually everything in audio has been used before, this should not be new. Only the use of cabinet depth to explicitly control directivity has not been discussed as much in forums.

If this is done correctly, controlled horizontal directivity down to 400Hz is possible.
But of course only for relatively wide radiating speakers.

 

[abdo123]

Something like this - using two 4'' midwoofer.

View attachment 161214 View attachment 161217

The trick is the ratio of baffle width and cabinet depth. The depth of the cabinet should usually be between 80-200mm. I'm currently working on three other designs using this technique to control directivity.
The finished simulations have been sitting on my hard drive for years, but I usually don't have the time or the motivation to implement the projects. But I'm working on them at the moment.

Since virtually everything in audio has been used before, this should not be new. Only the use of cabinet depth to explicitly control directivity has not been discussed as much in forums.

If this is done correctly, controlled horizontal directivity down to 400Hz is possible.
But of course only for relatively wide radiating speakers.

This is unbelievably simple and smart and I would love to keep this design moving forward.

As long as we get the W371A clone to complement it at some point though.

 

[tuga]

The 26 cm bass module width does not allow for much more than an 8 inch woofer on the front.

A pyramidal shape perhaps?

Estelon

Or a bit of clever chamfering?

Pass SR-1

 

[Lbstyling]

Something like this - using two 4'' midwoofer.

View attachment 161214 View attachment 161217

The trick is the ratio of baffle width and cabinet depth. The depth of the cabinet should usually be between 80-200mm. I'm currently working on three other designs using this technique to control directivity.
The finished simulations have been sitting on my hard drive for years, but I usually don't have the time or the motivation to implement the projects. But I'm working on them at the moment.

Since virtually everything in audio has been used before, this should not be new. Only the use of cabinet depth to explicitly control directivity has not been discussed as much in forums.

If this is done correctly, controlled horizontal directivity down to 400Hz is possible.
But of course only for relatively wide radiating speakers.

Fantastic post. Much appreciated!

It's been clear for some time now that the cabinet shape is responsible for half or more of the quality of sound from a speaker. And it should really be seen as an extension of a waveguide rather than the box you put the parts in.

But alas, the consumer has difficulty accepting change in shape like this. I have alway thought the solution would be to bake a simple wire frame that holds the shape of a aesthetically pleasing 'box' with an acoustically transparent sock pulled over it. Inside the frame would be the true form of necessity hidden from view.

I have seen it done with dipole speakers effectively in the 90s

 

[ctrl]

But alas, the consumer has difficulty accepting change in shape like this.

I know, such a design is probably very difficult to accept for most people.

Since I don't have to make any commercial compromises and the speakers are in my study, there are no limits - except symmetry, asymmetrical speakers are not my thing

 

[abdo123]

@ctrl if we use the Purifi long stroke mid-bass drivers it would exceed in output several regular bookshelves measured here anyway.

 

[gy-k]

One can just drop ideas here at this point in the project?

For great controlled directivity performance, down to 1-1.5kHz in the following idea, I believe the best "off-the-shelf waveguide design" available is Marcel Batík's (mabat's) ST260 waveguide: https://at-horns.eu/ST260.html

I might be streching things a bit here by saying "off-the-shelf design", sort of what I mean by that it's not a ready to go component the one can just buy, one must put further work and time into it to get the waveguide manufactured (3D printed), but the source files of the design are available under a Creative Commons Non-Commercial license. (Which is a good fit for the project, and based on this I hope it's okay enough with mabat that I'm just taking it and bringing it up here like this.)

So the waveguide design is ready (including an ABEC project which is a bonus for while system simulation), so things wouldn't turn into a waveguide design project. But it to not to be way too deeply DIY, requiring DIY 3D printing, it would need to be worked out how others could get the waveguide easily manufactured in a nice quality, like finding a handful of 3D printing companies in different geographical locations from which others could order the waveguide with a few clicks, something like that.

I'd like to suggest a 2-way design with the freestanding ST260 waveguide on top of a mid-bass cabinet built out of Denovo flat packs with a 10" mid-bass driver.


HF: as far as spinorama performance, lot of it would come from the waveguide and the HF unit. The freestanding ST-260 waveguide is to be mounted on top of the mid-bass cab. This requires constructing something out of a few pieces of wood that holds the waveguide. (It's more DIY, but this one should be easy.) The exact construction of this thing depends on the driver spacing on the Y and Z axes, which depends on the crossover point, or vice-versa. But I think the spacing/crossover point could be worked out using ABEC and Vituix along with eyeballing the datasheet FR of the candidate drivers. (I think the crossover would be 1-1.3kHz, and Kimmosto's rule of thumb of 1-1.4x wavelength spacing is something to keep in mind.)

Once a 3D-printed waveguide, or actually the whole speaker would exist, there could be test of a number of HF drivers to see wich one is the nicest performer. Some candidates could be: 18Sound HD1050/ND1090/95, B&C DE502, Beyma CP380M, BMS 4550, Faital HF109, RCF ND350/CD400. Most of these have their impedance peak at lower frequencies and their exit angle is less than the waveguide's entry, but if that doesn't matter the following could also be interesting: BMS 4538, Faital HF108, Peerless DFM-2535R00-08. (Sorry if I've left out other good ones.)
This many measurements would be best outsourced to a volunteer with a Klippel NFS, I imagine it would still take a lot of time, but in theory changing the driver on the waveguide between measurement runs would be easy.

LF: a floorstander cabinet could be could be constructed using 2x "Denovo Audio Knock-Down MDF 1.0 cu. ft." subwoofer cabinets. Glueing or bolting them together. That would make a 73.39cm tall cab or >98.39cm tall speaker with the waveguide on top, with 56.63l volume. If a 38cm long 10cm/4" diamater pipe or something equivalent would fit inside it that would result in a 30Hz tuning. I think the best would be putting the port on the bottom. This would require speccing feet or an additional base for the speaker. (Which is more DIY again, but this one should be easy as well. The bottom port behavior could be simulated using ABEC.)

Right now I'm looking at the Beyma 10G40 mid-bass driver. Simming it in a 56l cabinet tuned to 30Hz and applying Linkwitz transform (fo: 61.6, Qo: 0.4, fp 35, 0.5) I get a basically flat response with f3: 26.3, f6: 23.4, f10: 20.9. Might be a bad example actually, pushed too low and without a HPF, the point is, flat response is possible combining low port tuning and Linkwitz transform. The Beyma 10iX is an interesting driver, with a large Xmax and a lot heavier cone, without active filters it models kind of flat with a -1dB kind of tilt below 200hz with the following response: f3: 31.7, f6: 27.9, f10: 24.5. Though this driver is lot more expensive and seems to be unavailable. The FR of both of these Beymas seems fairly good? The PHL 3411 is another interesting driver, chances are the Genelec S360 uses this driver and the "symmetric BL(x) and Le(x)" and the excursion specs in the datasheet sound pretty cool, but it would need more filters to get a flat response with this tuning. These are the mid-bass drivers I've looked at so far.


In summary, I think it would be a fairly solid cab for the effort, a good 10" mid-bass could do bass well, as well as playing nicely up to 1-1.3kHz, and I'm pretty sure mabat's ST260 waveguide would sound great and show really nice performance, either with active or passive crossover with the right drivers.

 

[voodooless]

@ctrl if we use the Purifi long stroke mid-bass drivers it would exceed in output several regular bookshelves measured here anyway.

With two 4" Purifies per side that would bring you to 1200 euro already, so that does not leave a lot of budget for the rest.

A single 4" Purify would limit SPL to about 105 dB. I would aim for > 110 dB max output @1m.

There is also a bit of a contradiction between the added bass output capability and a proposed closed box design. With 2x8" closed, you will probably not have much more (maximum) output than the single Purify in the R1 < 50 Hz.

 

[abdo123]

There is also a bit of a contradiction between the added bass output capability and a proposed closed box design.

How? the absolute amount of bass is almost always (roughly) the same because of excursion limits anyway, a port might raise output at 40Hz but it won't do anything to output at 100Hz. With a bass module we don't need to worry about getting the response flat.

With two 4" Purifies per side that would bring you to 1200 euro already, so that does not leave a lot of budget for the rest.

Yeah but because it's a 4-inch woofer we can achieve Neuman KH80 levels of linearity easily, While with a constant directivity waveguide you're looking at milled aluminum or 3D printing, neither of which are particularly cheap to do, and the majority of people will not be paritcularily happy with the later option in their domestic spaces.

The mid-bass drivers can be repurposed for other projects so they're an 'investment' in a way, it's important to keep in mind that not only the cost of the speaker should be manageable but also the cost of developing it, it might take several iteration of waveguides to get the constant directivity right for the tweeter of choice.

 

[Rick Sykora]

One can just drop ideas here at this point in the project?

For great controlled directivity performance, down to 1-1.5kHz in the following idea, I believe the best "off-the-shelf waveguide design" available is Marcel Batík's (mabat's) ST260 waveguide: https://at-horns.eu/ST260.html

I might be streching things a bit here by saying "off-the-shelf design", sort of what I mean by that it's not a ready to go component the one can just buy, one must put further work and time into it to get the waveguide manufactured (3D printed), but the source files of the design are available under a Creative Commons Non-Commercial license. (Which is a good fit for the project, and based on this I hope it's okay enough with mabat that I'm just taking it and bringing it up here like this.)

So the waveguide design is ready (including an ABEC project which is a bonus for while system simulation), so things wouldn't turn into a waveguide design project. But it to not to be way too deeply DIY, requiring DIY 3D printing, it would need to be worked out how others could get the waveguide easily manufactured in a nice quality, like finding a handful of 3D printing companies in different geographical locations from which others could order the waveguide with a few clicks, something like that.

I'd like to suggest a 2-way design with the freestanding ST260 waveguide on top of a mid-bass cabinet built out of Denovo flat packs with a 10" mid-bass driver.


HF: as far as spinorama performance, lot of it would come from the waveguide and the HF unit. The freestanding ST-260 waveguide is to be mounted on top of the mid-bass cab. This requires constructing something out of a few pieces of wood that holds the waveguide. (It's more DIY, but this one should be easy.) The exact construction of this thing depends on the driver spacing on the Y and Z axes, which depends on the crossover point, or vice-versa. But I think the spacing/crossover point could be worked out using ABEC and Vituix along with eyeballing the datasheet FR of the candidate drivers. (I think the crossover would be 1-1.3kHz, and Kimmosto's rule of thumb of 1-1.4x wavelength spacing is something to keep in mind.)

Once a 3D-printed waveguide, or actually the whole speaker would exist, there could be test of a number of HF drivers to see wich one is the nicest performer. Some candidates could be: 18Sound HD1050/ND1090/95, B&C DE502, Beyma CP380M, BMS 4550, Faital HF109, RCF ND350/CD400. Most of these have their impedance peak at lower frequencies and their exit angle is less than the waveguide's entry, but if that doesn't matter the following could also be interesting: BMS 4538, Faital HF108, Peerless DFM-2535R00-08. (Sorry if I've left out other good ones.)
This many measurements would be best outsourced to a volunteer with a Klippel NFS, I imagine it would still take a lot of time, but in theory changing the driver on the waveguide between measurement runs would be easy.

LF: a floorstander cabinet could be could be constructed using 2x "Denovo Audio Knock-Down MDF 1.0 cu. ft." subwoofer cabinets. Glueing or bolting them together. That would make a 73.39cm tall cab or >98.39cm tall speaker with the waveguide on top, with 56.63l volume. If a 38cm long 10cm/4" diamater pipe or something equivalent would fit inside it that would result in a 30Hz tuning. I think the best would be putting the port on the bottom. This would require speccing feet or an additional base for the speaker. (Which is more DIY again, but this one should be easy as well. The bottom port behavior could be simulated using ABEC.)

Right now I'm looking at the Beyma 10G40 mid-bass driver. Simming it in a 56l cabinet tuned to 30Hz and applying Linkwitz transform (fo: 61.6, Qo: 0.4, fp 35, 0.5) I get a basically flat response with f3: 26.3, f6: 23.4, f10: 20.9. Might be a bad example actually, pushed too low and without a HPF, the point is, flat response is possible combining low port tuning and Linkwitz transform. The Beyma 10iX is an interesting driver, with a large Xmax and a lot heavier cone, without active filters it models kind of flat with a -1dB kind of tilt below 200hz with the following response: f3: 31.7, f6: 27.9, f10: 24.5. Though this driver is lot more expensive and seems to be unavailable. The FR of both of these Beymas seems fairly good? The PHL 3411 is another interesting driver, chances are the Genelec S360 uses this driver and the "symmetric BL(x) and Le(x)" and the excursion specs in the datasheet sound pretty cool, but it would need more filters to get a flat response with this tuning. These are the mid-bass drivers I've looked at so far.


In summary, I think it would be a fairly solid cab for the effort, a good 10" mid-bass could do bass well, as well as playing nicely up to 1-1.3kHz, and I'm pretty sure mabat's ST260 waveguide would sound great and show really nice performance, either with active or passive crossover with the right drivers.

Hi and welcome to ASR!

Thanks for sharing your thoughts. If you have not done so already, please read the start of the Introducing Directiva thread and all of this one to get an understanding of basis for this project. This project is past the initial requirements, the design team has been formed and we are reviewing the requirements before a plan design start for next week.

The overall Directiva project is done in the spirit of DIY. We had a smaller team do r1 and a larger team (currently 5-6 volunteers) doing r2. One key aspect here is the smaller team as it enable focus towards completing the project. If what we specifiy for a given release fits your needs, great. If not, it is open source and you are welcome to reuse it more to your liking. r2 is intended to be an extensible design, but there needs to be an target (or two) implementation for the first release. One of the primary requirements is to keep the build simple. So, as an example, while custom waveguides can help directivity, do not plan to use them.

Hope this helps and hope you find it an interesting project!

Rick

 

[voodooless]

How? the absolute amount of bass is almost always (roughly) the same because of excursion limits anyway, a port might raise output at 40Hz but it won't do anything to output at 100Hz. With a bass module we don't need to worry about getting the response flat.

Above 100 Hz is not a problem, it's below 50. And I'm also not worried about a flat response, we have a DSP for that. The issue is max SPL capability. If you want to have it working for a larger room, you will need at least a bit more maximum output down there.

Yeah but because it's a 4-inch woofer we can achieve Neuman KH80 levels of linearity easily, While with a constant directivity waveguide you're looking at milled aluminum or 3D printing, neither of which are particularly cheap to do, and the majority of people will not be paritcularily happy with the later option in their domestic spaces.

I have no idea why you drag the waveguide into this discussion? I'm talking output vs budget of the midranges here, and the question is: would 105 dB @ 1m be enough? If so, you can do with a single one per side.