• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Directiva r2 project: market requirements gathering

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,223
Likes
17,799
Location
Netherlands
Is there consensus on whether it's better to have the sharp (off-axis) transition between omni and narrow-constant directivity of the 8c:
To quote yourself:
I don't know, are you asking for me personally or generally?
;)

But seriously: obviously, in most systems, the bass is omni. However the gradient system has specific directivity for mid and high, and I'm not sure that is the best idea. You'd either want a smooth transition or quite uniform throughout, exactly like the two examples you posted already.

I'm inclined to think the D&D version is the better deal overall.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
Really want to avoid a waveguide unless it is part of the tweeter design. Amir has reviewed a number of speakers with good directivity without needing waveguides. This was already discussed in r1.

Linkwitz's design is a dipole and my question was around the OB bass module. Ofc, the upper part is going to need space as the those drivers are open baffle and have significant energy output from the back.

I had already planned a separate the bass module and upper monitor, so not sure where the confusion may be. The high level goal is to improve directivity over r1. So r1 is the reference. Am still not clear on where the most audible improvement in directivity may come from. At the risk of repeating myself, am interested in the audible benefit. Lacking this, not interested at all in something superficial like making a prettier polar map or chasing other cool designs without knowing the problem they are trying to solve...

For example, if controlling bass directivity has the benefit of clearer bass and better room integration, then this seems like a good goal. However, if improving the directivity above 5 kHz will be more audible, then would not spend time or money for better bass directivity. As I have not heard a kii 3 or the d&d 8c, am relying on others to articulate the benefits.

Hoping the discussion turns towards what benefits a given feature has and less of how a feature may have been accomplished. :)

Since you brought up the point about audible benefits, I wonder how much the subject of high sensitivity is worthy of exploration, as per the topic below:

 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
To quote yourself:

;)

But seriously: obviously, in most systems, the bass is omni. However the gradient system has specific directivity for mid and high, and I'm not sure that is the best idea. You'd either want a smooth transition or quite uniform throughout, exactly like the two examples you posted already.

I'm inclined to think the D&D version is the better deal overall.

It doesn't look to me like the D&D produces a smooth transition between constant and omni... But, personally, I do like the constant directivity both from a little of experience with horns and thoretically. (not seriously though, that's only for you)

SPL%20Horizontal.png

 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,223
Likes
17,799
Location
Netherlands
It doesn't look to me like the D&D produces a smooth transition between constant and omni...
Well, it's probably the best one can do given the size restrictions. All in all, it's pretty amazing for such a small speaker.
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
Can this be modeled in vituixcad directly I wonder?
Theoretically, kind of. I haven't tried it, so I can't say how realistic the result is.

For a good comparison you need the three front plus rear enabled/disabled. Having 2 or three woofer on the front already changes directivity. I think things might improve when you filter the rear for < 500 Hz.
With the available 4 or 8 ohm drivers, the combination of three drivers is rather unusual. Therefore, two front woofers were chosen for the comparison.

But as already mentioned (in relation to four front woofers), the directivity does not fundamentally change. Horizontally, almost nothing changes and vertically, the radiation narrows because the "driver array" is extended.
I have nevertheless listed the case below to reassure the skeptics ;)

Normalized sonogram for bass module with two front driver and three front driver:
1634891213708.png

1634891228874.png



Also try mounting the rear driver all the way down like Genelec goes. That will make the path length different to the listener longer.
Yep, that's on my list too. I actually wanted to wait with further simulations until it is clear whether such a concept is being considered, since the effort for the simulations is not insignificant.
But since you're already making the suggestion.... :)

Normalized sonogram for bass module with rear driver up and rear driver down (don't get confused, in the first vertical sonogram the negative angles are up, in the second sonogram positive angles are up):
1634893027830.png

1634893040258.png


On the whole, this brings a slight improvement in the radiation. This is probably the better option.

Before anyone else asks, my list also includes
the rear driver with acoustic slot,
side drivers,
side drivers with additional acoustic damping,
side drivers with acoustic slot.
But there the simulation effort is again significantly higher. Therefore, do not expect to see this in the near future - unless someone makes me an "Indecent Proposal" (and I don't mean having sex with an old geezer) ;)



You could also experiment with an all-pass passive filter to make the delay longer. That won’t be cheap though…
This gets complicated quite quickly due to the wiring of the drivers and that you also have to keep an eye on the overall impedance of the bass module.

Everyone is invited to try their hand at this ;)


As an attachment again the corresponding VCAD project - bass module with rear driver down (instead of up).
Important: Set the frequency range in VCAD to 50 - 2000Hz, only in this range the simulation is valid.
 

Attachments

  • BM_v2.zip
    366.5 KB · Views: 56

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
If three processing channels per audio channel is the limit, one could do a hybrid crossover - passive MT unit (passive parts to roll off the drivers and correct for individual driver breakup, use the processing channel to set the response, a la JBL 7-series i-line) + bass. That gives you 2 channels for bass.
Yes, you are right and for a single project this would be a very good option, but the bass module is now intended to serve as a platform for many other speaker tops. A single channel for the MT unit is quite limiting, makes the top speaker more expensive (passive crossover parts) and less flexible.
But it would definitely be an option.
 
Last edited:

fluid

Addicted to Fun and Learning
Joined
Apr 19, 2021
Messages
691
Likes
1,196
Theoretically, kind of. I haven't tried it, so I can't say how realistic the result is.
There is an example here which is about as accurate as a Vituix sim can be of side slots, plenty of real prototype measurements later to compare with

https://www.diyaudio.com/forums/mul...-active-speaker-insipired-8c-post6773689.html
But there the simulation effort is again significantly higher. Therefore, do not expect to see this in the near future - unless someone makes me an "Indecent Proposal"
If you carry on in the thread linked above DonVK has made some ABEC simulations and there is some information on the simulation of side ports through lumped elements. If Don manages to make good plane wave tube measurements of different acoustic materials then a better model could be made. Given that most of the things necessary to make a good passive cardioid are known, measuring them is harder than making them.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,313
Location
UK
I don't agree in this case. What instead is you have a leaky box--the "vent" is not storing energy, and instead is modeled as a resistor, Rx, whose value is independent of frequency, and introduces another order to the system with the loss the RC product of Rx and Cab (cabinet compliance). In any e-vent, the system is now 3'rd order and rolls off at -18dB/octave. Been kicking around for 50 years now, but is seldom employed except unintentionally by crappy cabinet construction . http://diyaudioprojects.com/Technical/Aperiodic/Aperiodic-Closed-Box-Loudspeakers.htm
If you don't agree how about showing the math behind your argument? Or a reference where someone did?
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
3,514
Likes
7,026
Location
Stow, Ohio USA
Since you brought up the point about audible benefits, I wonder how much the subject of high sensitivity is worthy of exploration, as per the topic below:


Had read this thread earlier and reread again. Have comparable experiences with higher sensitivity speakers. Have always suspected it could be more important to have more dynamic bass than more extended bass.

My CBTs allow some clear comparison of the potential benefit of greater sensitivity for me. Since more sensitive is likely better, the question is how much more and other potential tradeoffs. r2 was already targeted to increase system sensitivity via added woofers. Pushing the design towards even more sensitive drivers limits the selection significantly though…

Since @ctrl has demonstrated that better bass directivity may be more readily achieved than I thought, certainly seems a more worthwhile pursuit in this new light. Even if the monitor needs to be more passive to keep channel count lower, we still could devote a channel to it for allow overall adjustments. :cool:
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
I don't agree in this case. What instead is you have a leaky box--the "vent" is not storing energy, and instead is modeled as a resistor, Rx, whose value is independent of frequency, and introduces another order to the system with the loss the RC product of Rx and Cab (cabinet compliance)
If you don't agree how about showing the math behind your argument? Or a reference where someone did?
We had this discussion back in March when Erin's D&D 8c review came out.

I think it is a question of definition. Without the damping fleece in front of the slot, the slot behaves like a classic Helmholtz resonator. We had discussed that based on simulations here and the following posts.
You can also see that implied on the simulations @fluid linked.

Unfortunately, so far no reviewer had the idea to measure the slot in the near field. The question now is whether the damping material changes the behavior of the slot so much that the characteristic of the Helmholtz resonator is completely lost.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,313
Location
UK
The question now is whether the damping material changes the behavior of the slot so much that the characteristic of the Helmholtz resonator is completely lost.
The question is what is that math behind the suggested behaviour, which is different to the standard resonator? Why would reducing the efficiency of the port will do anything other than reduce the Q at the resonance frequency?

In science you either experiment or calculate. We can discuss as much as you want but without mathematical representation of the suggested behaviour there’s no possible way to agree on anything. As the proof of the said behaviour is not demonstrated either there is nothing to discuss as we only have a suggestion.
 

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
The entire history of loudspeaker design is people imagining things that are audible and important and then spending enormous amounts of time implementing their ideas only to find that their design priorities are not supported by psychoacoustic research. Most of the 'errors' that designers seek to correct are things they see in measurements first, not from reported listener experiences. The history of technical 'issues' which designers seek to correct is exactly the same as the history of acoustical measurements. First they were obsessed with flat axial FR and bass Q, because that's all they could measure or model. Then in the 80s, you have software acoustic measurement and designers become obsessed with 'transient perfection' because it's one of the new things their tools showed them. More measuring tools create more imaginary issues and invented solutions.

Harmonic distortion and IMD is an example of this. Harmonic distortion is a crude metric that mathematically characterizes one simple type of deviation from an ideal signal. There is absolutely no reason to assume that it is audible except in extreme cases, and if it is audible, it is almost certainly the wrong metric to use to characterize the audibility. People have this intuition that orchestral music is 'complex' presumably because it has a lot of instruments, and this is more susceptable to IMD distortion, but let's be real. What is 'complexity'? A square wave has an infinite series of odd harmonics, is it 'complex'? The sound of a single familiar voice almost never sounds correct played back - is it 'complex'? Is IMD the correct metric to describe why your orchestral recordings sound like a cloudy soup? Unlikely.

A lot of audio designers clown audiophiles like Romy the Cat who has this very famous multiway horn system, all DIY, with a mishmash of exotic and unusual drivers such as a compression driver from the 50s, an english PA speaker for the bass, and a huge RAAL tweeter for the treble, all connected to passive line level crossovers and tube amps. And I would agree that it is not a cost effective way to create good sound - but consider the methodology. Romy the Cat doesn't 'suppose' something sounds bad due to measurements - he tries to make determinations about what he wants his system to sound like. He identifies the problem with the reproduction first, in his own experience and in his own terms. Then, he makes a change and listens to it.

There are obvious problems with this approach from an engineering perspective, but making assumptions about the importance of technical issues is not a shorter path to audio quality. Rick seems to want to bring the focus to audibility, but what do we really know about what performance metrics matter?

Lots of bass, user adjustable for quantity and to correct room issues. Smooth off axis without major resonances down to Schroeder. Narrow mid/treble dispersion less room-dependent than wide, but not necessarily better. What else does psychoacoustic research say about speakers?
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
We can discuss as much as you want but without mathematical representation of the suggested behaviour there’s no possible way to agree on anything.
my expertise on this topic is abysmal in comparision to the two of you, but the simulations he produced should be 'mathematically representative' of what is going on. Isn't that true?
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,313
Location
UK
my expertise on this topic is abysmal in comparision to the two of you, but the simulations he produced should be 'mathematically representative' of what is going on. Isn't that true?
I apologise if I missed but what I remember was a discussion about the possible cause of distortion on the 8C. Those simulations were about standard port designs that acted like ports. May I ask you to please point me to calculations where port stuffing causes the port to stop acting like a resonator and shifts the frequency as you suggested.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
The entire history of loudspeaker design is people imagining things that are audible and important and then spending enormous amounts of time implementing their ideas only to find that their design priorities are not supported by psychoacoustic research. Most of the 'errors' that designers seek to correct are things they see in measurements first, not from reported listener experiences. The history of technical 'issues' which designers seek to correct is exactly the same as the history of acoustical measurements. First they were obsessed with flat axial FR and bass Q, because that's all they could measure or model. Then in the 80s, you have software acoustic measurement and designers become obsessed with 'transient perfection' because it's one of the new things their tools showed them. More measuring tools create more imaginary issues and invented solutions.

Harmonic distortion and IMD is an example of this. Harmonic distortion is a crude metric that mathematically characterizes one simple type of deviation from an ideal signal. There is absolutely no reason to assume that it is audible except in extreme cases, and if it is audible, it is almost certainly the wrong metric to use to characterize the audibility. People have this intuition that orchestral music is 'complex' presumably because it has a lot of instruments, and this is more susceptable to IMD distortion, but let's be real. What is 'complexity'? A square wave has an infinite series of odd harmonics, is it 'complex'? The sound of a single familiar voice almost never sounds correct played back - is it 'complex'? Is IMD the correct metric to describe why your orchestral recordings sound like a cloudy soup? Unlikely.

A lot of audio designers clown audiophiles like Romy the Cat who has this very famous multiway horn system, all DIY, with a mishmash of exotic and unusual drivers such as a compression driver from the 50s, an english PA speaker for the bass, and a huge RAAL tweeter for the treble, all connected to passive line level crossovers and tube amps. And I would agree that it is not a cost effective way to create good sound - but consider the methodology. Romy the Cat doesn't 'suppose' something sounds bad due to measurements - he tries to make determinations about what he wants his system to sound like. He identifies the problem with the reproduction first, in his own experience and in his own terms. Then, he makes a change and listens to it.

There are obvious problems with this approach from an engineering perspective, but making assumptions about the importance of technical issues is not a shorter path to audio quality. Rick seems to want to bring the focus to audibility, but what do we really know about what performance metrics matter?

Lots of bass, user adjustable for quantity and to correct room issues. Smooth off axis without major resonances down to Schroeder. Narrow mid/treble dispersion less room-dependent than wide, but not necessarily better. What else does psychoacoustic research say about speakers?

I think this is a very important discussion that we should have. Is this design rooted in fidelity (minimize destructive room interactions with dipoles, coaxials, cardioid radiation .etc) or is it rooted in enjoyment / psychoaccoustic preference?

I feel something like the Neumann KH310 but made with the Philharmonic BMR tweeter would be a phenomenal design from a psychoaccoustic point of view. which is something completely different than the current discussions.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
I apologise if I missed but what I remember was a discussion about the possible cause of distortion on the 8C. Those simulations were about standard port designs that acted like ports. May I ask you to please point me to calculations where port stuffing causes the port to stop acting like a resonator and shifts the frequency as you suggested.

you're right the simulations were about the increased excursion / harmonic distortion. sorry about that.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
The entire history of loudspeaker design is people imagining things that are audible and important and then spending enormous amounts of time implementing their ideas only to find that their design priorities are not supported by psychoacoustic research. Most of the 'errors' that designers seek to correct are things they see in measurements first, not from reported listener experiences. The history of technical 'issues' which designers seek to correct is exactly the same as the history of acoustical measurements. First they were obsessed with flat axial FR and bass Q, because that's all they could measure or model. Then in the 80s, you have software acoustic measurement and designers become obsessed with 'transient perfection' because it's one of the new things their tools showed them. More measuring tools create more imaginary issues and invented solutions.

Harmonic distortion and IMD is an example of this. Harmonic distortion is a crude metric that mathematically characterizes one simple type of deviation from an ideal signal. There is absolutely no reason to assume that it is audible except in extreme cases, and if it is audible, it is almost certainly the wrong metric to use to characterize the audibility. People have this intuition that orchestral music is 'complex' presumably because it has a lot of instruments, and this is more susceptable to IMD distortion, but let's be real. What is 'complexity'? A square wave has an infinite series of odd harmonics, is it 'complex'? The sound of a single familiar voice almost never sounds correct played back - is it 'complex'? Is IMD the correct metric to describe why your orchestral recordings sound like a cloudy soup? Unlikely.

A lot of audio designers clown audiophiles like Romy the Cat who has this very famous multiway horn system, all DIY, with a mishmash of exotic and unusual drivers such as a compression driver from the 50s, an english PA speaker for the bass, and a huge RAAL tweeter for the treble, all connected to passive line level crossovers and tube amps. And I would agree that it is not a cost effective way to create good sound - but consider the methodology. Romy the Cat doesn't 'suppose' something sounds bad due to measurements - he tries to make determinations about what he wants his system to sound like. He identifies the problem with the reproduction first, in his own experience and in his own terms. Then, he makes a change and listens to it.

There are obvious problems with this approach from an engineering perspective, but making assumptions about the importance of technical issues is not a shorter path to audio quality. Rick seems to want to bring the focus to audibility, but what do we really know about what performance metrics matter?

Lots of bass, user adjustable for quantity and to correct room issues. Smooth off axis without major resonances down to Schroeder. Narrow mid/treble dispersion less room-dependent than wide, but not necessarily better. What else does psychoacoustic research say about speakers?
I’m buying a pair of single driver speakers...
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,223
Likes
17,799
Location
Netherlands
With the available 4 or 8 ohm drivers, the combination of three drivers is rather unusual. Therefore, two front woofers were chosen for the comparison.
When doing these kind of experiments, I tend to not concern with those kind of things at first. It only leads too many variables that change at ones. Better to first prove or compare single concept, and if it has merit, one can see how things like impedance can work out.
Normalized sonogram for bass module with rear driver up and rear driver down (don't get confused, in the first vertical sonogram the negative angles are up, in the second sonogram positive angles are up):
View attachment 160667
View attachment 160668

On the whole, this brings a slight improvement in the radiation. This is probably the better option.
Looks like it indeed makes quite a difference. But I suspect removing the driver in favor of a lot will save one woofer, and a bigger effect due to the larger delay.
 
Last edited:
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
3,514
Likes
7,026
Location
Stow, Ohio USA
The entire history of loudspeaker design is people imagining things that are audible and important and then spending enormous amounts of time implementing their ideas only to find that their design priorities are not supported by psychoacoustic research. Most of the 'errors' that designers seek to correct are things they see in measurements first, not from reported listener experiences. The history of technical 'issues' which designers seek to correct is exactly the same as the history of acoustical measurements. First they were obsessed with flat axial FR and bass Q, because that's all they could measure or model. Then in the 80s, you have software acoustic measurement and designers become obsessed with 'transient perfection' because it's one of the new things their tools showed them. More measuring tools create more imaginary issues and invented solutions.

Harmonic distortion and IMD is an example of this. Harmonic distortion is a crude metric that mathematically characterizes one simple type of deviation from an ideal signal. There is absolutely no reason to assume that it is audible except in extreme cases, and if it is audible, it is almost certainly the wrong metric to use to characterize the audibility. People have this intuition that orchestral music is 'complex' presumably because it has a lot of instruments, and this is more susceptable to IMD distortion, but let's be real. What is 'complexity'? A square wave has an infinite series of odd harmonics, is it 'complex'? The sound of a single familiar voice almost never sounds correct played back - is it 'complex'? Is IMD the correct metric to describe why your orchestral recordings sound like a cloudy soup? Unlikely.

A lot of audio designers clown audiophiles like Romy the Cat who has this very famous multiway horn system, all DIY, with a mishmash of exotic and unusual drivers such as a compression driver from the 50s, an english PA speaker for the bass, and a huge RAAL tweeter for the treble, all connected to passive line level crossovers and tube amps. And I would agree that it is not a cost effective way to create good sound - but consider the methodology. Romy the Cat doesn't 'suppose' something sounds bad due to measurements - he tries to make determinations about what he wants his system to sound like. He identifies the problem with the reproduction first, in his own experience and in his own terms. Then, he makes a change and listens to it.

There are obvious problems with this approach from an engineering perspective, but making assumptions about the importance of technical issues is not a shorter path to audio quality. Rick seems to want to bring the focus to audibility, but what do we really know about what performance metrics matter?

Lots of bass, user adjustable for quantity and to correct room issues. Smooth off axis without major resonances down to Schroeder. Narrow mid/treble dispersion less room-dependent than wide, but not necessarily better. What else does psychoacoustic research say about speakers?

Seems to say there more than just good design principles to consider :)...

Otherwise we are just making educated guesses at some point. Despite trying to make a more science-based speaker, there will still be tradeoffs and can only try to make a speaker that most people will like. Since there are tradeoffs, if we cannot meaningfully rank them, then it comes down to a judgment call. As I know the design team will be torn at times, the decision may come to me. So, if this does happen, would like to know what the benefits may be gained or lost. The alternative is iterating our way there and really do not have time or money to do so.

I like the discussion over the last few days on bass directivity as an example. Am still a bit skeptical about taking on a bit too much, but I can see that a multi-driver design may be obtainable and more likely based on having more drivers. While the D&D approach would seemingly be lower cost, if we have to iterate multiple prototypes to get the tuning right, am less inclined to take on the likely added cost of time/materials to do so.

For r1, we initially tried a fancier design with more parts and it did not work out. This time would like to be more prepared to have a more reusable design that gives us more options for both r2 and post-r2. I will draft requirements tomorrow and then we can review. Maybe that should include some pass/fail criteria, but the team really is not that big, so not sure it is needed. I suggest if we identify key stakeholders first, this may be all we will need. :cool:
 
Top Bottom