ASR Directiva Open Source Speaker Review

[RickS]

My comment was for the passive version that was posted here.

Thanks, I missed that. Probably best we move any potentials to the another thread (like the Introducing Directiva one)...

 

[sarumbear]

Thanks, I missed that. Probably best we move any potentials to the another thread (like the Introducing Directiva one)...

ASR is a large community and it’s growing everyday. It’s not logical to accept an open development thread to work well. The noise to substance ratio will be too high. It is already high.

May I suggest a closed thread where only the developers of a certain design can post and everyone else only reads/follows. From time to time a developer can cross comment a post from an open thread of a similar subject. This offers valid suggestions made by other members to be discussed in an orderly matter and possibly implemented to the final design. A sticky post may be kept to summarise the current or final state of the design which will also list the specs and parts. Naturally the developers can and I hope will accept more members to their “room”.

This is a pretty standard concept in open software development for a couple of decades. I believe if managed well this will help the developers as well as the interested members.

 

[RickS]

ASR is a large community and it’s growing everyday. It’s not logical to accept an open development thread. The noise to substance ratio will be too high. It is already high.

May I suggest a closed thread where only the developers of a certain design can post and everyone else only reads/follows. From time to time a developer can cross comment a post from an open thread of a similar subject. This offers valid suggestions made by other members to be discussed in an orderly matter and possibly implemented to the final design. A sticky post may be kept to summarise the current or final state of the design which will also list the specs and parts. Naturally the developers can and I hope will accept more members to their “room”.

This is a pretty standard concept in open software development for a couple of decades. I believe if managed well this will help the developers as well as the interested members.

Yes this pretty much the model we have once the team concurs on the high-level design...

Still it is a community approach, and willing to take the bad with the good.

We are at a stage where we need more team members for r2. So, if you or anyone else is interested, PM me. Not everyone has the time, so we have part-time contributors too. Thanks!

 

[ROOSKIE]

Well, if I am at the peak of mount stupid, then you can pull me down to the valley of despair by the design or measurements.
And again, my verdict is irrelevant to my level.
I can not say a soft thing to the Directiva "as a product", because it's so unfair to the "competitor" (if I can call them) such as Neumann, Genelec, etc., they are manufacturing well balanced, properly designed products.

Only midwits cite DKE. Because they cannot articulate their position/argument in a convincing fashion.

Due to @Acoustician signing in from Japan, I was actually asking. The behavior may have a different identity there, or not really be commonly known. There may also be some thing lost in translation here as going from Japanese to English might be creating some odd misunderstandings.

From time to time a Mozart is real, a Da Vinci is real, a Kurosawa is real.
I think it still makes sense to ask if @Acoustician understands he/she may be overestimating their ability.
In no way was I accusing them of being at the top of mount stupid. That is not what the DKE is dealing with. I does not mean or imply you are at the top of mount stupid, so I don't think you are familiar with it.
It does suggest that even if someone is having some level of success, even some high level success, you are over estimating the nature of your true skill level in a significant way and amount. The main thing is that the person is are unaware of it yet, so in the USA we would say that you "are that guy". Most folks have been "that guy", a few times over. It is to some degree normal, especially as a person works to develop genuine confidence.

Acoustician has shown some skill in modifying another persons 3 way tower speaker design. Most measurements are currently predicted. That speaker has a great chance at measuring well on the Klipple. It might end up with a Harman score over 8.
Therefore the Directive, a 2-way speaker using an off the shelf tweeter is crap. This is despite it measuring extremely well on a real Klippel test. It achieved a fully tested and vetted, Harman score of 6.5/8.0(with sub) which is excellent if one is into the score thing.

Regarding that DXT waveguide, I have no doubt the waveguide works reasonably well. Look at all the measurements I posted earlier from 3 sources and this speakers test results. Can it be improved? Well what are your goals? My understanding of this tweeter is that the guide was optimized so that the tweeter face is still a standard size. I appears given that characteristic they are performing quite well. IMHumbleO as a hobbyist.

In any case if you read about the stated goals for the project, they have been met successfully (except for cost which is why the Purifi becomes the issue as it drives the cost way up. Though it is a great woofer if cost is ignored).
This project had specific goals and many of those goals are limiting to a degree.

via @ricksykora
"Towards the end of last year, @amirm and I started discussing a potential audio system project that would include an open source streamer and an active speaker design. I had already been working on a DIY modular speaker concept and planned to start a design based on it, but had some other projects that needed to be completed first. This allowed me to make some adjustments to my plans and reconsider some of my original ones. Here were my goals for the speaker:

1. The speaker would be active and simple to build (for me, this meant no tools beyond a circular saw but for more optimal results is more like a table saw and router).
2. The design tools would be free but flexible enough to allow external data import/export.
3. The design would be openly shared but FOR NON-COMMERCIAL USE ONLY.
4. The cost target for the drivers and crossover would be approximately $1000 US.
5. Keep the cabinet size to around 15 liters or about .5 a cubic foot or smaller (allowing it to be reasonably shipped and Amir to test). The electronics would be external. (EDIT)
6. The design would allow for modest flexibility and scalability so others could (eventually) modify and add value. So, here is where Amir intervened with his ideas/goals...
7. The speaker should have good directivity. Note, this did force me to reconsider some of my (mainly cabinet) plans as we'll see later.
8. The speaker should use the new Purifi 6.5 woofer. As it was not in my original plans, I was hesitant due mainly to cost. However, after working on the SPK5, had them in hand and agreed to use...

This put some pressure on the budget, but the tweeter candidates were not pricey, so on to the next task (getting some help)..."


Acoustician, after you complete your tower speakers, I really do hope they sound amazing and measure well.
Share what you have done here.
If you really do have a high level of design skill, and you certainly may have it or be on your way to having it, then please contribute it here.
If the chosen tweeter and it it incorporated waveguide is a poor choice for a 6.5" 2 way, then suggest another option.
If no off the shelf option currently exists. Help the team make the speaker even better. Help design a great open source, printable waveguide that fits a great tweeter.
I'd love if I could design such a tweeter and waveguide myself.

 

[ctrl]

I was always curious why @ctrl didn't go all in since it's an active design. However now one might argue that losing around 3dB of sensitivity is not worth the improvements!

Well, the prototype was intended to show the potential of the design, both for the active and potential passive version. With six or more PEQ in the tweeter crossover, the result would no longer be feasible as a passive crossover (or only with very many notch filters) - Therefore, only as many PEQ as absolutely necessary.

I always find @Maiky76's tweaks very exciting, as they rise to the level of how well or not so well the various aspects of a speaker's dispersion work together.

But as said, one should not overestimate the preference score. However, I find it quite helpful in the development or optimization of a design, since it provides an unbiased ranking in addition to the differences one observed visually.

he has some little points on the DXT problems

The problem is not so much the SeasDXT tweeter, but rather the not fully optimized baffle.

For small waveguides (anything under 180mm in diameter) that need to support the lowest possible crossover frequency, the waveguide should be designed specifically for the baffle and cabinet. Or less effectively the other way around.

This was discarded for Directiva, as has already been said, for the simplest possible build. Next, the baffle could have been optimized for the SeasDXT, but then no off-the-shelf cabinet could have been used, so this option was also discarded.

So the goal was to make the build as simple as possible, with a "shoebox-sized" off-the-shelf cabinet.

If you want, you can equip the Directiva with an optimized baffle or cabinet, like the ones A. Heissman used in his DXT concepts, or as presented here in the "Introducing Directiva" thread - then with newly designed crossover.

The Directiva project has further potential. Those who are not afraid of complicated manufacturing can still significantly improve the directivity, especially in the frequency range 4-10kHz.

Simulation of the measured Directiva Prototype:


Directiva with optimized cabinet:

Significant improvement in horizontal dispersion, slight improvement in vertical dispersion.
The preference score confirms the visual impression.

If you want to know more about it, please discuss this in the "Introducing Directiva" thread. Just wanted to address statements that claim the SeasDXT waveguide was miserably developed by Seas.

 

[Tangband]

Well, the prototype was intended to show the potential of the design, both for the active and potential passive version. With six or more PEQ in the tweeter crossover, the result would no longer be feasible as a passive crossover (or only with very many notch filters) - Therefore, only as many PEQ as absolutely necessary.

I always find @Maiky76's tweaks very exciting, as they rise to the level of how well or not so well the various aspects of a speaker's dispersion work together.

But as said, one should not overestimate the preference score. However, I find it quite helpful in the development or optimization of a design, since it provides an unbiased ranking in addition to the differences one observed visually.



The problem is not so much the SeasDXT tweeter, but rather the not fully optimized baffle.

For small waveguides (anything under 180mm in diameter) that need to support the lowest possible crossover frequency, the waveguide should be designed specifically for the baffle and cabinet. Or less effectively the other way around.

This was discarded for Directiva, as has already been said, for the simplest possible build. Next, the baffle could have been optimized for the SeasDXT, but then no off-the-shelf cabinet could have been used, so this option was also discarded.

So the goal was to make the build as simple as possible, with a "shoebox-sized" off-the-shelf cabinet.

If you want, you can equip the Directiva with an optimized baffle or cabinet, like the ones A. Heissman used in his DXT concepts, or as presented here in the "Introducing Directiva" thread - then with newly designed crossover.

The Directiva project has further potential. Those who are not afraid of complicated manufacturing can still significantly improve the directivity, especially in the frequency range 4-10kHz.

Simulation of the measured Directiva Prototype:
View attachment 158141

Directiva with optimized cabinet:
View attachment 158142View attachment 158138
Significant improvement in horizontal dispersion, slight improvement in vertical dispersion.
The preference score confirms the visual impression.

If you want to know more about it, please discuss this in the "Introducing Directiva" thread. Just wanted to address statements that claim the SeasDXT waveguide was miserably developed by Seas.

Nice work indeed.
During work on the active dsp loudspeaker HYBRID , and the original passive loudspeaker, we/they didnt find major benefits with changing the cabinet shape. Most of the diffractions from the tweeter, if using a waveguide, dissapears almost completely.

HYBRID - a loudspeaker with dsp crossover

Hi. I will tell a little about a project that started in Sweden and in the beginning was a passive loudspeaker. The goals where to create a good loudspeaker with good directivity with driveunits that were good but not really expensive. This passive loudspeaker were not developed by me. I bought...

www.audiosciencereview.com

There is however some real gains in sound quality to have the right damping material inside the cabinet. Sheep wool is really good at damping soundwaves inside the cabinet. Or other material can make the cabinet acoustically bigger.

Read more here:

faktiskt.io • Visa tråd - Allram moffar i lådan, Del 1 + Del 2 + Slutsatser

About resonances:
I have had good results using bitumen glued on the sides inside the loudspeaker cabinet in HYBRID. One should avoid having the same dimensions on the thickness of the mdf baffle and rear wall. At my investigations with bitumen, I reckognized that damping all the walls inside the cabinet only lowered the cabinet resonances and made it sound worse when listening to music. But damping half of the cabinet walls made the sound better, spreading the resonances in different frequencies - It could easily be heard on piano recordings that it was an improvement.

Loudspeaker stands and feets:
Good results in sound quality has been made with blutack on the bottom of the loudspeakers fixating it on a loudspeaker stand.
Read more here :

The Sound of Surprise (the loudspeaker/stand interface) Page 4

The first thing to note is that, while the amplitude of the speaker's Helmholtz or bass tuning frequency around 63Hz doesn't change much with the change in interface material, the mode three octaves higher, at 156Hz, is vanquished by all the compliant supports. Fig.1, for example, shows the...

www.stereophile.com

 

[sarumbear]

About resonances:
I have had good results using bitumen glued on the sides inside the loudspeaker cabinet in HYBRID. One should avoid having the same dimensions on the thickness of the mdf baffle and rear wall. At my investigations with bitumen, I reckognized that damping all the walls inside the cabinet only lowered the cabinet resonances and made it sound worse when listening to music. But damping half of the cabinet walls made the sound better, spreading the resonances in different frequencies - It could easily be heard on piano recordings that it was an improvement.

I understand that one of the criteria of this project is to use off-the-shelf enclosures but adding braces shouldn't be big ask to the DIY'er as long as we tell them where to place the braces. Any DIY shop will help you to cut a piece of of wood to length. The rest is glue.

During the development of our Silver5L we researched cabinet resonances extensively. We knew of the Matrix system that B&W had started to use on the first generation 800 series. The Matrix 801 was launched a few years before we started our project but we understood the concept of enclosure resonance issues since 70s having discussed it with my friend Siegfried RIP intensively after his seminal Wireless World article of 1978. By using a modified torn-arm/cartridge setup we investigated every part of our enclosure against a frequency sweep while using different type of damping materials. We built various marble enclosures and with the help of a friend who owned a foundry we even had a cast iron one made! That was the best by far but naturally it was not practical. We found out that braces work much better than any material that is practical. B&W was certainly on the right track.

The simplest test you can have is to tap on a speaker enclosure with your knuckle. If you hear a sound, the enclosure walls are resonating.

 

[Trdat]

Great work Rick, your effort is appreciated. Now for a very similar DIY build but 3 bass drivers coupled with 2 mid drivers or thereabouts.

I am just curious if we implement our own DSP in particular one that is automated such as Audiolense or say something like Acourate would the results be similar? I get the benefit of minimal distortion throughout the chain with the choice of DAC's just not sure how the filters through DSP software will effect directivity...?

 

[ctrl]

During work on the active dsp loudspeaker HYBRID , and the original passive loudspeaker, we/they didnt find major benefits with changing the cabinet shape. Most of the diffractions from the tweeter, if using a waveguide, dissapears almost completely.

As luck would have it, I have a simulation of a 170mm waveguide on my hard drive - comparable to the one you used.

First, let's look at the waveguide in infinite baffle (normalized on-axis sonogram):

The directivity at smaller angles looks a bit strange, this is because the waveguide was designed for a specific baffle width.
So if you are right, then the frequency response should not change when installed in a loudspeaker cabinet as long as the waveguide controls the radiation.
In our example the -6dB limit is drawn in black and up to 2.7kHz the Waveguide seems to control the radiation.

If the baffle and speaker enclosure have little effect, then there should be virtually no change above 2.7kHz when installed in a speaker enclosure (horizontal, normalized on-axis sonogram):

As expected, the control of the waveguide increases to lower frequencies due to the interaction with the loudspeaker cabinet. The radiation is now controlled down to about 1.3kHz.
But what is also noticeable is that in the frequency range 2.7-7kHz the frequency response curves have also changed significantly for small angles.

You can see this a little better if you look at the frequency responses in direct comparison (IB=infinite baffle, LS=loudspeaker cabinet, horizontal, normalized on-axis sonogram):

Hence the difference in horizontal directivity between the normal and the optimized version of the Directiva cabinet shown in post446.

 

[RickS]

Great work Rick, your effort is appreciated. Now for a very similar DIY build but 3 bass drivers coupled with 2 mid drivers or thereabouts.

I am just curious if we implement our own DSP in particular one that is automated such as Audiolense or say something like Acourate would the results be similar? I get the benefit of minimal distortion throughout the chain with the choice of DAC's just not sure how the filters through DSP software will effect directivity...?

Thanks, but as I mentioned, much of the credit belongs to @ctrl. He has much more design experience than I.

You are suggesting an entirely different design, so how much Directiva r1 applies is rather limited IMO.

 

[kyle_neuron]

I understand that one of the criteria of this project is to use off-the-shelf enclosures but adding braces shouldn't be big ask to the DIY'er as long as we tell them where to place the braces. Any DIY shop will help you to cut a piece of of wood to length. The rest is glue.

During the development of our Silver5L we researched cabinet resonances extensively. We knew of the Matrix system that B&W had started to use on the first generation 800 series. The Matrix 801 was launched a few years before we started our project we understood the concept of enclosure resonance issues since early 80s having discussed it with my friend Siegfried RIP intensively after his seminal Wireless World article of 1978. By using a modified torn-arm/cartridge setup we investigated every part of our enclosure against a frequency sweep while using different type of damping materials. We built various marble enclosures and with the help of a friend who owned a foundry we even had a cast iron one made! That was the best by far but naturally it was not practical. We found out that braces work much better than any material that is practical. B&W was certainly on the right track.

The simplest test you can have is to tap on a speaker enclosure with your knuckle. If you hear a sound, the enclosure walls are resonating.

Some great research on bracing here:

E-library page - AES

AES E-Library ← Back to search

www.aes.org

And resonances here:

E-library page - AES

AES E-Library ← Back to search

www.aes.org

It seems that lengthwise bracing across the longest dimension is the most effective, of the types tested.

 

[Trdat]

Thanks, but as I mentioned, much of the credit belongs to @ctrl. He has much more design experience than I.

Yes, that is correct I saw his input you both deserve credit. DIY is right down my alley and this was a great project to follow. Ctrl appreciate your input and looking forward to another larger build.

You are suggesting an entirely different design, so how much Directiva r1 applies is rather limited IMO.

I have nothing against minidsp(maybe the hiss mentioned), no doubt it is the most practical and most affordable DSP, and I might even go with the whole outfit its just I already got Audiolense. I suppose I can use the proposed filters with other hands on DSP software.

Regarding how to factor in directivity with the crossover I might start a new thread.

 

[sarumbear]

Some great research on bracing here:

E-library page - AES

AES E-Library ← Back to search

www.aes.org

And resonances here:

E-library page - AES

AES E-Library ← Back to search

www.aes.org

It seems that lengthwise bracing across the longest dimension is the most effective, of the types tested.

Thank you for these. As I said in my post research on managing cabinet resonance had been available for a long time. Any manufacturer disregarding this and still selling loudspeakers that have enclosure resonance is simply [fill your word here].

 

[thewas]

In below link there are audio comparative recordings of cabinet resonances with different thicknesses, material and damping combinations (use google to translate to english) which impressively show the possible audible differences (by the way also a DXT based loudspeaker)

Hifi-Selbstbau - DUO-DXT

HiFi-Selbstbau - Das Online-Magazin

www.hifi-selbstbau.de

 

[Arc Acoustics]

It's not your preference which was criticised but the way you commented a very good DIY attempt here. By the way even the 2 way Genelec and Neumann have their small directivity problems.

Well what are your goals? My understanding of this tweeter is that the guide was optimized so that the tweeter face is still a standard size. I appears given that characteristic they are performing quite well.

Well, I am posting my opinion to the Directiva "as a product" consistently, which hypothetically cost 400-500 grand a pair (according to the 1st post), that is KH310 or even close to 8341 money in some store.
Are you guys sure?

Some might think I'm in the DKE, others just hate me.
Never mind, you can interpret me as you like, If that makes you feel better, that's great.
But is that really relevant?

My point was, whether it is competitive among other products or not, it's flaws and their degrees (acceptable or not, considering the price, balancing of design conflicts, etc.).

I can see flaws, they might be not insignificant for me, and once I asked the reason, I have a responsibility to pointing out. (Or should I evade?)
I honestly do not think that crossing a pricy 6.5" woofer and cheap 1" tweeter at 2.7kHz in a cheaply made box (even if it has fancy finish) is a competitive design for that amount of money.
You can do it in a DIY project, you can do it in a low-ish priced model (such as 8040), but as a 400-500 grand pair product, I really do not think it is a good design concept (of course I know there are much much worse products in the market though).

Acoustician, after you complete your tower speakers, I really do hope they sound amazing and measure well.
Share what you have done here.
If you really do have a high level of design skill, and you certainly may have it or be on your way to having it, then please contribute it here.
If the chosen tweeter and it it incorporated waveguide is a poor choice for a 6.5" 2 way, then suggest another option.
If no off the shelf option currently exists. Help the team make the speaker even better. Help design a great open source, printable waveguide that fits a great tweeter.
I'd love if I could design such a tweeter and waveguide myself.

Thank you for your kind words but sorry, I am not active in ASR because of the "thread starter" (you can see the reason here, I almost banned for mocking his "scientific attitude", you can add my so-called warning point and ban me by reporting my post! LOL) so I would not post ASR (maybe FB or DIYAudio) and for the very same reason, I can not participate to the ASR project.

 

[thewas]

Well, I am posting my opinion to the Directiva "as a product" consistently, which hypothetically cost 400-500 grand a pair (according to the 1st post), that is KH310 or even close to 8341 money in some store.
Are you guys sure?

In the original post #1:

Rick tells me that the cost for the components to build a pair is about US $1,300 to $1,400.

That is less than half than what a KH310 or 8341 cost which makes the rest of your post and comments invalid.

 

[abdo123]

That is less than half than what a KH310 or 8341 cost which makes the rest of your post and comments invalid.

you're missing the cost of DSP and amplification that the Neumann KH310 has already included.

 

[thewas]

you're missing the cost of DSP and amplification that the Neumann KH310 has already included.

Which still don't make the price difference if chosen reasonably (the amplification modules of most monitors cost are far from expensive "highend").

 

[abdo123]

Which still don't make the price difference if chosen reasonably (the amplification modules of most monitors cost are far from expensive "highend").

consumers usually don't have access to those affordable options, the cheapest equivalent thing would be the Hypex FA253. As Class A/B designs would require you to design dedicated cooling solutions and other Class D plate amplifiers are not really up to standard considering the quality of the actual speaker.

the Hypex F253 would be close to ~500-600$ per speaker.

 

[thewas]

By the way for people having no access to affordable solutions there is also the passive option which costs only 50 bucks per speaker

Post Directiva r1 Passive Crossovers here

This looks very promising. As long as the loudspeaker can still be called a 4 Ohm LS according to the official designation (impedance always >= 3.2 Ohm), I don't see any problem if the impedance falls below 4 Ohm. But your min 4 ohms are of course even better. Make sure that the component...

www.audiosciencereview.com

which would be often even preferred by many people who have already good integrated or power amps.