Room is acoustically solved with uniform rt60 about 0,3 - 0,34 seconds 200-20000Hz (speakers as a source) and other speakers sound neutral. Treatment is mainly on back wall and celling, less on the other walls.
FR is attached, measured about 1,3m from speaker at tweeter level (tweeter is like 10° below ear level)
FR is attached, measured about 1,3m from speaker at tweeter level (tweeter is like 10° below ear level)
Attachments
Your on-axis response is flat(ish) and the room is leaning towards the dead side for stereo. If the high frequency is to be perceived as too strong then the off-axis high frequency response will be strong. Some typical examples of this are constant directivity studio monitors, dipoles and omni-directional speakers which radiate more off-axis high frequencies than conventional speakers with a narrowing beamwidth with frequency. Your simulations show a constant(ish) directivity from the waveguide which would fit assuming your other speakers that sound better balanced had a more conventional radiation pattern. A slight downward slope of the on-axis response may well be the way to go.
here are some very rough, quick and dirty tests with measurements for a wall speaker project I am about to start.
general setup:
I took measurements in 40 cm distance, to get most of the "wall" (big plywood board) influence and have the microphone next to the wall at 90° angle.
as this is a single, small FR driver I suppose 40 cm is enough to get a rough idea of directivity and baffle/wall influence.
the driver is a visaton FRS-8 (cheap 3" fullrange), about 15 or 20 years old.
the TSP differ quite significantly from the actual data sheet values, just as a side note (see comparison in the image above).
The driver is mounted centrally on the small "baffle" and this degrades the on-axis response quite badly. The on-axis hump around 2 k and the dip at 4k is because of the baffle layout.
To reduce the baffle effect I chose a reference angle of 10°.
My main interest was the midrange behaviour and the possibilities with different setups.
here are the six most interesting variants:
bad on axis and polar response due to near square baffle with centrally mounted driver. comb filtering due to step to the back wall.
on-axis and polar responses are far better (see directivity index). response above 5 kHz (driver is beaming) and below 500 Hz (omni/half space) is of course similar. remaining drop of directivity index at 500 Hz is due to distance to back wall.
response obviously drops constantly below the ~1 kHz peak.
below 500 Hz there is a drop at 60° off axis (the "rotated" OB side null?), and presumably the rear radiation exiting at 90°.
above 1 kHz this variant seems to behave quite omnidirectional. I presume this is due to the rear driver output being restricted by driver cutout (not beveled) and absorption by polyester batting.
This variant is very much inspired by mabat and vineethkumar - thanks!
There is still a peak at 1 kHz but the response does not drop so evenly (and steeply) below that.
The speaker seems to behave "cardioidish" below 1 kHz, with a constriction at 700 Hz.
The "spread out" dipole eight has disapperared (no red borders at 90°).
further tweaking would definitely be necessary regarding the thickness of melamine foam, baffle size and driver position.
I also tried covering the front part of the side melamine-foam-openings with cardboard to increase lower output and still keeping the cardioid behaviour, but without much success.
this version is "less cardioid", mostly at the lower end. I guess the main influence is the thicker melamine foam, being more resistive.
the worse 3,5 kHz notch is probably a "wall" reflection resonance.
This was an attempt to increase the directivity above 2 kHz, which did work out, but just very specificly at 3 kHz.
some more work is be necessary to combine a full range driver with a wave guide and a cardioid enclosure.
And here are the respionse and polar-map results once more to be able to toggle through them.
for my (low-fi-ish) project I will probably just go on with a wall speaker with slanted side walls.
I am still very curious about mabat's recent resistive midrange project, to be combined with an ATH waveguide.
general setup:
I took measurements in 40 cm distance, to get most of the "wall" (big plywood board) influence and have the microphone next to the wall at 90° angle.
as this is a single, small FR driver I suppose 40 cm is enough to get a rough idea of directivity and baffle/wall influence.
the driver is a visaton FRS-8 (cheap 3" fullrange), about 15 or 20 years old.
the TSP differ quite significantly from the actual data sheet values, just as a side note (see comparison in the image above).
The driver is mounted centrally on the small "baffle" and this degrades the on-axis response quite badly. The on-axis hump around 2 k and the dip at 4k is because of the baffle layout.
To reduce the baffle effect I chose a reference angle of 10°.
My main interest was the midrange behaviour and the possibilities with different setups.
here are the six most interesting variants:
1. "closed" box sitting in front of the wall, driver ~ 8cm distant
bad on axis and polar response due to near square baffle with centrally mounted driver. comb filtering due to step to the back wall.
2. same box but with added slanted sides elements, ~ 30° angle
(did I mention it's a very quick and dirty setup? no slanted wall at bottom) on-axis and polar responses are far better (see directivity index). response above 5 kHz (driver is beaming) and below 500 Hz (omni/half space) is of course similar. remaining drop of directivity index at 500 Hz is due to distance to back wall.
3. "open baffle" with some poly-batting behind the driver (not stuffed)
response obviously drops constantly below the ~1 kHz peak.
below 500 Hz there is a drop at 60° off axis (the "rotated" OB side null?), and presumably the rear radiation exiting at 90°.
above 1 kHz this variant seems to behave quite omnidirectional. I presume this is due to the rear driver output being restricted by driver cutout (not beveled) and absorption by polyester batting.
4. side openings 4,5 cm wide and filled with 3 cm thick melamine foam
I used foam called "dirt eraser" (german "Schmutzradierer") as can be bought cheaply in houshold shops.This variant is very much inspired by mabat and vineethkumar - thanks!
There is still a peak at 1 kHz but the response does not drop so evenly (and steeply) below that.
The speaker seems to behave "cardioidish" below 1 kHz, with a constriction at 700 Hz.
The "spread out" dipole eight has disapperared (no red borders at 90°).
further tweaking would definitely be necessary regarding the thickness of melamine foam, baffle size and driver position.
I also tried covering the front part of the side melamine-foam-openings with cardboard to increase lower output and still keeping the cardioid behaviour, but without much success.
5. side openings 3 cm wide (smaller space behind driver), filled with 4,5 cm thick melamine foam
I just removed one of the MDF spacers and rotated the melamine foam pieces. this version is "less cardioid", mostly at the lower end. I guess the main influence is the thicker melamine foam, being more resistive.
the worse 3,5 kHz notch is probably a "wall" reflection resonance.
6. same as 5. but with some added clay to make some kind of "wave guide" bulge around the driver
This was an attempt to increase the directivity above 2 kHz, which did work out, but just very specificly at 3 kHz.
some more work is be necessary to combine a full range driver with a wave guide and a cardioid enclosure.
And here are the respionse and polar-map results once more to be able to toggle through them.
for my (low-fi-ish) project I will probably just go on with a wall speaker with slanted side walls.
I am still very curious about mabat's recent resistive midrange project, to be combined with an ATH waveguide.
Last edited:
I should have mentioned that the web-links take you to the respective cardioid midrange/coax posts by @mabat and @vineethkumar01!
For reference, here are the responses and polar plots for the simple closed box (refer to number 1 above) and the resistive cardio speaker (number 4 "melamine") in 4pi space, free standing, without "wall" behind the speaker. note the 180° polar map scale, instead of 90°!
wiggles above 2 kHz and specially the widening at 4 kHz is effect of the bad baffle creating an on-axis dip at 4 kHz.
also have a look at vineethkumar's version with sica coax!
EDIT: sorry - images don't work, even if attached repeatedly. I'll try again later!
1. closed box free standing
4. side openings 4,5 cm wide and filled with 3 cm thick melamine foam, free standing
wiggles above 2 kHz and specially the widening at 4 kHz is effect of the bad baffle creating an on-axis dip at 4 kHz.
also have a look at vineethkumar's version with sica coax!
Last edited:
Mic distance is 40 cm (the first photo might be misleading!).
8 cm is the distance from baffle surface to the "wall" (board) behind the speaker!
Stv, nice experiments! Have tried doubling the baffle diameter? Or, halve it to minimal baffle? Baffle could be made from thick absorbing material? I think relationship between edge diffraction and wall reflection is unfavorable now 🤔
Exactly ... even slanted side walls from melamine foam - resistive + absorptive elements!
Not tried yet - I'm thinking maybe simulation is the best way to move on.
Open driver close to wall is not a good solution IMO. Of course stv's tests will show the real world...
Bass will suffer in spl, but above mid the pathlength difference of directed vs. reflected will make lots of interferences in wide spectrum.
A big question is how to control dispersion through spectrum. And what is the desired pattern?
The Atmos installation and demos I have seen and heard all used Genelecs... here in Finland. Big in-wall mains and G-series auxiliaries. https://www.genelec.com/home-theatres
JBL offers wide and narrow radiation speakers for HT https://jblpro.com/en/product_families/immersive-rooms
Bass will suffer in spl, but above mid the pathlength difference of directed vs. reflected will make lots of interferences in wide spectrum.
A big question is how to control dispersion through spectrum. And what is the desired pattern?
The Atmos installation and demos I have seen and heard all used Genelecs... here in Finland. Big in-wall mains and G-series auxiliaries. https://www.genelec.com/home-theatres
JBL offers wide and narrow radiation speakers for HT https://jblpro.com/en/product_families/immersive-rooms
Last edited:
I must add an important note for multichannel speaker systems.
All speakers should have similar crossover topology to keep phase through spectrum identic. Localization just doesn't happen if this rule is broken. Delay/distance adjustment cannot correct this.
Early stereo test was to change polarity of the other speaker's wires and listen how image was lost. But nowdays the most common fault I see is to have 3-way mains and a 2-way in others. As well LR2 vs. LR4 turns phase/polarity differently. My HT system has identical coaxial L/C/R but different 2-way rears, and it sounds unbelievably better than the living room setup with 4-way mains and 2-way C/RL/RR. Fortunalely I'm not a movie freak, I mostly listen to stereophonic music.
All speakers should have similar crossover topology to keep phase through spectrum identic. Localization just doesn't happen if this rule is broken. Delay/distance adjustment cannot correct this.
Early stereo test was to change polarity of the other speaker's wires and listen how image was lost. But nowdays the most common fault I see is to have 3-way mains and a 2-way in others. As well LR2 vs. LR4 turns phase/polarity differently. My HT system has identical coaxial L/C/R but different 2-way rears, and it sounds unbelievably better than the living room setup with 4-way mains and 2-way C/RL/RR. Fortunalely I'm not a movie freak, I mostly listen to stereophonic music.
I second that, resistive cardioid (only usable for midrange) however might allow to reduce the dispersion towards the wall, causing less peaks and cancellations.
A separate bass unit with long enough wavelengths will not be affected very much by path length difference.
Yes. That is what highly regarded D&D 8C does, except the bass unit is two rear facing woofers in the same box with the resistive cardioid mid and waveguided tweeter.
That is an interesting perspective... I had not thought of that but it does make sense.
I AM a movie freak, but I watch films with simple stereo 2-channel sound, the same system I use for music.... I am unusual in that I feel that dialogue, plot, narrative, mood, character development, and visual presentation are FAR more important than a surround sound effect.
I reduced the baffle width and chamfered the driver hole, to allow better air/pressure circulation.
Again, as before, I put some polybatting behind the driver and covered the sides and top side with melmine foam, width 4,5 cm, thickness 3 cm.
This is the result:
and for comparison once again the previous version, posted above #124 with wider baffle, no chamfered driver hole:
the better circulation and narrow baffle have (mostly positive) effects:
- increased, more even directivity below 500 Hz (comes with the price of slightly reduced SPL)
- slightly increased directivity above 1 kHz, up to about 5 kHz, where the diaphragm is beaming already.
- the "open baffle" peak at 1 kHz is stronger. as far as I understand this is the increased output because of the coinciding rear pressure wave.
Last edited: