2-way 10'' speaker with swappable waveguides for constant/linear directivity

[CinamonRolls]

Kimmosto said Genelec dealers like the sound of his Aalto speakers. Kimmosto is using feedback from multiple people with industry experience to construct a subjective measure. You could formalize that but it would get expensive. Personally, I don't think loudspeakers are important enough to go to that much trouble and expense for more objective subjective measurements.

But, robotics and computation reduced the cost of anechoic chambers via the Klippel. Maybe new research & technology will reduce the cost of formal listening studies one day. Until then we will have to live with what we've got in measurements and shared subjective feedback/opinions.

 

[kimmosto]

...at what level of certainty can one predict a speakers subjective performance with its CTA2034 measurement alone?

It depends... I've probably mentioned also on ASR that it's possible to have almost "perfect" CTA-2034-A Directivity response, but the speaker can be very bland, artificial and boring. Unusable weak with bad microdynamics IMO. Soundstage could be scattered too, but that's different problem. Dr. Olive lists e.g. non-linear distortion and max SPL, but those two alone have not explained the problem, because both can be good with very bland speaker.
I try to minimize the following to avoid bland reproduction:
- Excess group delay down to 20 Hz. This usually means that subwoofer is not allowed.
- Short and long term compression.
- Spherical wavefront at low...mid frequencies i.e. radiators in front baffle only, and adequate radiating area.
- "Too" constant directivity which reduces too much relative power and reflection energy at midrange.

 

[CinamonRolls]

I didn't realize Kimmosto added the feature to show the data labels for the various slopes in VituixCAD2. I though he was doing it with a manual calculation. I love this. Thank you for adding it and also writing the document about preferred slopes of specific responses (Slope and linearity of sound power (SP), directivity index (DI) and sound power directivity index (SPDI); K Saunisto, 2024-11-24).

Now I can see how far off from preferred target slopes I am.

 

[Matias]

Aalto 9 was originally designed as symmetrical for flush-mount. The first three have Be-dome in quite shallow wave guide. That baffle model is available on request.
View attachment 409530

@kimmosto Are you the manufacturer of Aalto Speakers? If so would be good practice to announce to the moderators and get a Manufacturer badge for clarity.

 

[kimmosto]

@kimmosto Are you the manufacturer of Aalto Speakers? If so would be good practice to announce to the moderators and get a Manufacturer badge for clarity.

No. I just design for Aalto Speaker as a hobby, for free so far.

 

[Halbknapp]

@kimmosto, what i never understood is how to get such a smooth directivity (like in Aalto Speaker) in the crossover region between mid and high Frequency driver. Would not the vertical offset cause a small bump in directivity in anything other then a coaxial design?

 

[kimmosto]

@kimmosto, what i never understood is how to get such a smooth directivity (like in Aalto Speaker) in the crossover region between mid and high Frequency driver. Would not the vertical offset cause a small bump in directivity in anything other then a coaxial design?

Methods to smooth directivity around XO frequency:
1) Distance from mid/woofer to tweeter 1-1.4 x wave length at XO frequency. This reduces hump in directivity index down to ca. 1 dB assuming perfect "L-R" phase match.
2) Lower directivity of tweeter slightly above XO frequency. Wave guide/horn should be small and shallow enough to compensate directivity of midwoofer and directivity due to c-c distance.
3) Low order acoustical slopes distribute power dip to wider frequency range. This also reduces group delay with IIR filter.
4) Avoid excessive directivity of mid/wooofer cone. Select driver with less ideal piston directivity to reduce directivity right below XO frequency.
5) Intentional phase mismatch (extra delay in tweeter) at XO frequency to aim the highest SPL above (or below) the reference axis. This trick can be okay especially if listening elevation is temporarily high.

The rest can be compensated with small hump in on-axis response at XO frequency and slightly below. Shouldn't be much left.

 

[kimmosto]

For example c-c=1.1 x wave length at XO, in theory with omni radiators

In practice c-c is not a problem if listening point is far enough where power balance rules, or at designed listening elevation. The best speakers don't have coaxial so...

 

[Halbknapp]

For example c-c=1.1 x wave length at XO, in theory with omni radiators
View attachment 415028
In practice c-c is not a problem if listening point is far enough where power balance rules, or at designed listening elevation. The best speakers don't have coaxial so...

Thanks a lot for your explanation!

 

[XMechanik]

- Spherical wavefront at low...mid frequencies i.e. radiators in front baffle only, and adequate radiating area.

How about a bipolar configuration with a tweeter? I wonder if this would be a valid approach to reduce the PR slope of a small "wave-guided" speaker? An additional (rear) speaker should have almost no effect on the axial response, spraying around additional energy in the HF, which will increase the power response characteristic and thus reduce directivity in the frequency region where it peaks.

I ran some simulations and filter designs based on measurements of a 2-way system. I added a copy of the tweeter and set its parameters Z=264mm and r=180deg to simulate mounting on the rear panel.

Simulation results:

Normalized to on-axis

Impact of the rear driver: reduced DI for f>2kHz allowing to reduce slope of the PR and close it to the preferred values

 

[kimmosto]

How about a bipolar configuration with a tweeter? I wonder if this would be a valid approach to reduce the PR slope of a small "wave-guided" speaker?

It's one option, but not without few possible problems/limits. It's quite common that small stand speakers are located quite close to front wall in a small room or to compensate dip due to floor reflection (delay 1/2 wave length) with hump due to front wall reflection (1/1 wave length). In that case rear tweeter locates very close to front wall creating strong early reflection. Maximum level for rear tweeter is usually ca. -2...-5 dB compared to maximum of front tweeter. Vicinity of front wall almost forces to reduce rear level extra 3 dB or more.
Another interesting phenomenon is that delay in rear tweeter's wavefront could increase directivity to front sector. Consequences depend on shapes, dimensions, polarity and XO slopes of rear tweeter. Inverted i.e. "dipole" could be slightly smoother in some cases. "Dipole" in quotation marks because rear usually has (/should have) own individual high-pass so difference in phase angle between rear and front is variable with IIR.

Larger constructions with horn play with different rules. Rear tweeter does not produce much...at all effect at sweet spot if directivity of front horn is strong. Sound to radical off-axis will be more balanced so trick improves power response. But generally, speaker with large and deep horn is not usually designed for casual listening in all possible hearing locations. Target is usually more extreme hifi in a spot only.

 

[kimmosto]

Vicinity of front wall almost forces to reduce rear level extra 3 dB or more.

Diffuser on front wall is recommended - especially with rear tweeter.

 

[XMechanik]

Maximum level for rear tweeter is usually ca. -2...-5 dB compared to maximum of front tweeter

rear usually has (/should have) own individual high-pass

In case of these simulations the rear driver had its own filter and its magnitude varied from -5dB (4k) to 0dB (>10k) with respect to the main tweeter.

Despite its small size it makes no sense to call this type of approach, much less use it, as a bookshelf speaker design. More like a monitor requiring some setup work and perhaps additional accessories like the mentioned diffuser. Thanks for your remarks and sorry everyone for another OT.

 

[Duke]

How about a bipolar configuration with a tweeter? I wonder if this would be a valid approach to reduce the PR slope of a small "wave-guided" speaker? An additional (rear) speaker should have almost no effect on the axial response, spraying around additional energy in the HF...

A well-implemented rear-firing tweeter can improve the power response with no effect on the direct sound. Or to put it another way, it can reduce the spectral discrepancy between the first-arrival sound and the reflection field.

It's one option, but not without few possible problems/limits. It's quite common that small stand speakers are located quite close to front wall in a small room or to compensate dip due to floor reflection (delay 1/2 wave length) with hump due to front wall reflection (1/1 wave length). In that case rear tweeter locates very close to front wall creating strong early reflection. Maximum level for rear tweeter is usually ca. -2...-5 dB compared to maximum of front tweeter. Vicinity of front wall almost forces to reduce rear level extra 3 dB or more.

Imo a separate level control for the rear-firing tweeter can be useful. Ime there is a "sweet spot" level above which the rear-firing tweeter's contribution degrades clarity, and that level can change with the reflection path length.

Here is one way to get a longer reflection path length for a rear-firing tweeter:

 

[kimmosto]