Introducing Directiva - An ASR open source platform speaker project

[tktran303]

Subscribed.

That truncated pyramid cabinet is a form follows function design.

I think we OCD types get a bit obsessed with diffraction but can it really be audibly better?

I always liked big 1” round overs and they’re measurably different,
And felt around the tweeter.

but can I hear the difference?

Considering all the other effects of the room etc...

I went back to nice sharp edges.

Just food for thought..

 

[gingerbeer]

I know that the original premise was to make it simple to fabricate....

but I’m curious as to whether any of these tools can be automated to generate the dimensions to give the optimal response? (“Generative design”, eg: https://www.google.com/search?q=generative+design+parts. )

 

[bluefuzz]

That truncated pyramid cabinet is a form follows function design.

Yes, and I for one, would never let such a thing in the house. Regardless of how well they may measure we still have to look at our loudspeakers. Coffin-shaped or pyramid-shaped boxes are simply too fugly even for me, who doesn't have to contend with WAF as such, but I do have some æsthetic sensibilty. They are also challenging to build if you do not have a well stocked workshop and a fair bit of routine using it.

Large round-overs can look nice but, again, challenging to build. They require a wall thickness at least as thick as tthe round-over or a complicated construction with quarter-round dowels in the corners. Rounded boxes are also more difficult to finish. Veneer is difficult to make go round tight curves ...

The only realistic design for the average diy'er – also æsthetically – is a standard square box with optionally triangular chamfered corners. All that requires is a thick front baffle and a sharp plane which is doable for most people.

Perhaps I missed it but has an assymetric tweeter placement been modelled?

 

[LearningToSmile]

Yes, and I for one, would never let such a thing in the house.

Yeah, same. Still it's very cool to see all the different designs modeled and simulated, and like mentioned earlier in the thread, the finalized design can always serve as starting point for different variations.

Would love to see at the end to see the final design built as well as a completely standard rectangular box to see how the measurements compare given the same volume and components, and how much is gained from all the optimizations.

 

[bluefuzz]

Still it's very cool to see all the different designs modeled and simulated

Absolutely. It's a great initiative which I will follow with much interest. But as an inveterate DIY'er myself I know that 'perfection is the enemy of the good'. A project that is either too ugly or too difficult (or both) to build is doomed from the start. Expense (within reason) is less of a problem. The $1000 goal is fine but will vary wildly depending on where in the world you live and availability of parts, materials, tools etc.

 

[Rick Sykora]

Yeah, same. Still it's very cool to see all the different designs modeled and simulated, and like mentioned earlier in the thread, the finalized design can always serve as starting point for different variations.

Would love to see at the end to see the final design built as well as a completely standard rectangular box to see how the measurements compare given the same volume and components, and how much is gained from all the optimizations.

Yes! The Directiva project is meant to support freedom of choice. If you or your significant other does not like a truncated pyramid, there are other options. The beveled variant is arguably just as good from a directivity perspective. I have a wife and she disdains bevels. At one time I wanted to buy some closeout Emotiva speakers (that are beveled) and was promptly shot down.

Your other build idea is the plan. Am already prepping to build the simple Denovo cabinet version and later will build the more optimal version to allow for a comparison beyond the simulations.

 

[Rick Sykora]

Perhaps I missed it but has an assymetric tweeter placement been modelled?

I think this is a fair question, but @ctrl would have to simulate and like to be respectful of his time.

Please state what goal you have in mind that an asymmetric tweeter placement would accomplish and what your proposed offset would be. Thanks!

 

[ctrl]

Does the addition of a bass cabinet (or subwoofer) to turn this into a three-way negate the lower IMD benefit of using a Purifi woofer?

Every 2-way loudspeaker can benefit from a good subwoofer.

In the frequency range from 200Hz down, the Purifi chassis has its advantages over comparable drivers - extremely low harmonic distortion and low IMD (in the midrange) with large cone excursion.

However, if a 3-way concept is planned from the outset, where the crossover frequency is usually 200-300Hz to the midrange, there are drivers that can compete and are even slightly better, for example, in terms of harmonic distortion in the frequency range 200-3000Hz.
This then raises the cost-benefit question.

 

[bluefuzz]

Please state what goal you have in mind that an asymmetric tweeter placement would accomplish and what your proposed offset would be.

Well, it's a common, if not the most common, method of spreading edge diffraction effects on the tweeter. I simply wondered if it has been considered. From a DIY standpoint it is an easy 'fix' if indeed it does fix anything. But without modelling it we won't know. I don't have any specific offset in mind although due to the limited width of the baffle the options are few.

 

[Rick Sykora]

Well, it's a common, if not the most common, method of spreading edge diffraction effects on the tweeter. I simply wondered if it has been considered. From a DIY standpoint it is an easy 'fix' if indeed it does fix anything. But without modelling it we won't know. I don't have any specific offset in mind although due to the limited width of the baffle the options are few.

Fair enough, just wanted to check as there are other potentials to consider as well.

Those who know me, know I am a big advocate of identifying and reproducing known issues BEFORE trying to apply solutions. @ctrl has more experience with the DXT tweeter design AND diffraction, so will review with him.

 

[ctrl]

Perhaps I missed it but has an assymetric tweeter placement been modelled?

I think this is a fair question, but @ctrl would have to simulate and like to be respectful of his time.

No, it was not.
The technical reason for this is that with asymmetrical speaker cabinets the computing time increases exponentially. 1.5-2 hours of computing time (up-to-date CPU, 10 threads) can easily turn into 10 hours or more when the boundary elements are doubled, which is the case with asymmetric cabinets (this quickly adds up to days).

Unscientific reason: Horizontally different directivity is not my thing and actually I have no aesthetic demands on loudspeakers at all, but asymmetrical driver arrangement is my kryptonite, just the sight of it makes my b... shrink.

However, if the time arises, I will adapt my scripts for an asymmetrical loudspeaker design - without being able to give a concrete date.

 

[ctrl]

But without modelling it we won't know. I don't have any specific offset in mind although due to the limited width of the baffle the options are few.

Roughly you can simulate something like this in VituixCAD's "diffraction tool".
A bit more realistic results can be achieved with Boxsim, as long as you use the default tweeters (more info here).
Simulation of waveguides are not supported by either program.

 

[Rick Sykora]

Roughly you can simulate something like this in VituixCAD's "diffraction tool".
A bit more realistic results can be achieved with Boxsim, as long as you use the default tweeters (more info here).
Simulation of waveguides are not supported by either program.

Thanks! Will see what VC modeling indicates. As I stated earlier, need to identify a clear problem if there is one.

May have to wait until after I build one and measurements are done on it.

 

[ctrl]

As I stated earlier, need to identify a clear problem if there is one.

There is no clear problem. The only thing that cannot be achieved in this way is horizontally symmetrical directivity. Of course, the deviations in horizontal radiation vary from case to case.

So that it is not lost in the discussion: This alone says nothing about the possible sound of the loudspeaker - outstanding directivity is not everything.

As an extreme example of an asymmetrical driver arrangement, one could take the Neumann KH310. Coincidentally, Amir has already measured it.
That's where the asymmetrical arrangement of tweeter and midrange driver works pretty well.
If you look at the frequency responses normalized to 0° at +-30° and +-60°, they still run pretty well parallel to each other:

 

[HooStat]

The triangular shape made me wonder, could one build two boxes? One presumably small one for the tweeter, and a separate, larger one for the woofer, and stack them? From a construction perspective, I think that would be easier. Since I never see this, I assume it is a "bad idea". No need to answer this if it takes us too off topic or it isn't easy to do so.

 

[Rick Sykora]

There is no clear problem. The only thing that cannot be achieved in this way is horizontally symmetrical directivity. Of course, the deviations in horizontal radiation vary from case to case.

So that it is not lost in the discussion: This alone says nothing about the possible sound of the loudspeaker - outstanding directivity is not everything.

As an extreme example of an asymmetrical driver arrangement, one could take the Neumann KH310. Coincidentally, Amir has already measured it.
That's where the asymmetrical arrangement of tweeter and midrange driver works pretty well.
If you look at the frequency responses normalized to 0° at +-30° and +-60°, they still run pretty well parallel to each other:
View attachment 113328

Understood. I also took note as SPK5 tweeter is offset too...

Am speculating it may have been more for better nearfield demo purposes, but Amir's SPK5 directivity seemed just fine as well.

 

[Rick Sykora]

The triangular shape made me wonder, could one build two boxes? One presumably small one for the tweeter, and a separate, larger one for the woofer, and stack them? From a construction perspective, I think that would be easier. Since I never see this, I assume it is a "bad idea". No need to answer this if it takes us too off topic or it isn't easy to do so.

Maybe if you were building the version where the box is rectangular for the woofer.

Otherwise, for the truncated pyramid, it is more cuts, more pieces and more glue/assembly. This outweighs any other advantage that comes to mind. Notably, in my time around a table saw, I prefer not to do any angled small parts.

 

[HooStat]

Maybe if you were building the version where the box is rectangular for the woofer.

Otherwise, for the truncated pyramid, it is more cuts, more pieces and more glue/assembly. This outweighs any other advantage that comes to mind. Notably, in my time around a table saw, I prefer not to do any angled small parts.

I meant the rectangular box version. I was thinking of it as a "quasi-triangular" box for bad woodworkers.

 

[diyaudnut]

Was wavecor tweeter with waveguide considered?
http://www.wavecor.com/html/tw030wa11_12.html

 

[Zvu]

It's priced higher than DXT but inconclusively better performance.

Good suggestion though, as is Morel CAT378.