About a year ago, I bought five of the Monoprice speakers because they were on sale for 60% off. I ran them for a while as a bizarre ambio setup in my office.
I can post docs on that.
When Erin H posted the amazing measurements of the Kef Blade Meta 2 a few weeks ago, I decided to mess around a bit with making active cardioids using "off the shelf" speakers.
This is what a cardioid radiation pattern looks like.
These are the speakers: "Monolith by Monoprice Encore T5 Tower Speakers." The pics in the photo are the Encore "T6" which is 20% bigger than the T5s that I have. In other respects, they're very similar. Here's some measurements from Erin: https://www.erinsaudiocorner.com/loudspeakers/monoprice_encore_t6/
There are a couple of (obvious) ways to make an active cardioid:
1) You can put two speakers back-to-back, separate them by one quarter wavelength, put them in the same polarity, and create a Gradient Cardioid
2) You can put two speakers back-to-back, separate them by one quarter wavelength, put them in opposite polarity, and create a End Fire Cardioid
I can post docs on that.
When Erin H posted the amazing measurements of the Kef Blade Meta 2 a few weeks ago, I decided to mess around a bit with making active cardioids using "off the shelf" speakers.
This is what a cardioid radiation pattern looks like.
These are the speakers: "Monolith by Monoprice Encore T5 Tower Speakers." The pics in the photo are the Encore "T6" which is 20% bigger than the T5s that I have. In other respects, they're very similar. Here's some measurements from Erin: https://www.erinsaudiocorner.com/loudspeakers/monoprice_encore_t6/
There are a couple of (obvious) ways to make an active cardioid:
1) You can put two speakers back-to-back, separate them by one quarter wavelength, put them in the same polarity, and create a Gradient Cardioid
2) You can put two speakers back-to-back, separate them by one quarter wavelength, put them in opposite polarity, and create a End Fire Cardioid
I went with a Gradient cardioid. I need to run the calculator from here: https://www.merlijnvanveen.nl/en/calculators/28-sad-subwoofer-array-designer-en
You'll have to play around with the calculator, but as I understand it:
1) end-fire cardioids have better rejection to the back, but higher efficiency overall
2) gradient cardioids have worse rejection to the back, but higher efficiency overall
This is because the gradient cardioid is in-phase while the end-fire cardioid is out-of-phase.
You'll have to play around with the calculator, but as I understand it:
1) end-fire cardioids have better rejection to the back, but higher efficiency overall
2) gradient cardioids have worse rejection to the back, but higher efficiency overall
This is because the gradient cardioid is in-phase while the end-fire cardioid is out-of-phase.
Yes, the cardioid pattern only works for about one octave. Once I got this thing set up, it became fairly clear that the 'trick' to these things is that you need cardioid speakers for both the woofers and the midbass. Basically use two additional woofers to create a cardioid pattern, and one of them is used for the midbass, and the other is used for the woofer.
the problem you have, other than the mess in your living room 
, is that when you put two of those back to back, you get a huge separation between drivers which will limit how high you can get a cardioid response.
You might better put them side by side with one facing back, or front to back. polarity adjusted, with a small gap for slot loaded equivalen
, is that when you put two of those back to back, you get a huge separation between drivers which will limit how high you can get a cardioid response.You might better put them side by side with one facing back, or front to back. polarity adjusted, with a small gap for slot loaded equivalen
I've kinda been wondering what sort of interesting things could be done with a similar setup. I imagine building a bipole, with separate inputs for the front and the rear. Which is sorta what you did. And then you have DSP to do kooky things. E.G.
- high pass? low pass?
- notches for front wall reflections?
- phase manipulation via all-pass, or subdivision into bands and inverted, then recombined
I read someone's post once, that they were learning to measure and they set up the windowing wrong, instead of setting the window to end at the first reflection they mistakenly set it to start there. So instead of a measurement of the direct sound, they measured the room sound. I wonder, could you EQ that separately from (without screwing up) the direct sound, by using DSP on the rear side? (would that even be useful?)
Anyway, not that these are necessarily good ideas, but it seems like there are many things you could experiment with. I'm interested to hear about your experience with the cardioid setup, but also any other interesting tricks you tried.
- normal bipole
- dipole
- cardioid? (I didn't know there were two kinds!)
- have the rear play L minus R (high passed?)
- add delay?
- kooky EQ?
- high pass? low pass?
- notches for front wall reflections?
- phase manipulation via all-pass, or subdivision into bands and inverted, then recombined
I read someone's post once, that they were learning to measure and they set up the windowing wrong, instead of setting the window to end at the first reflection they mistakenly set it to start there. So instead of a measurement of the direct sound, they measured the room sound. I wonder, could you EQ that separately from (without screwing up) the direct sound, by using DSP on the rear side? (would that even be useful?)
Anyway, not that these are necessarily good ideas, but it seems like there are many things you could experiment with. I'm interested to hear about your experience with the cardioid setup, but also any other interesting tricks you tried.
lol I'm pretty sure that was me, and I found the same thing when I measured my room this time around:
Basically, the reflection off of the back wall is so strong, it's nearly identical to the source. The only real issue I can see is that the treble is rolled off above 10khz.
Basically, the reflection off of the back wall is so strong, it's nearly identical to the source. The only real issue I can see is that the treble is rolled off above 10khz.
"For Dummies" note, you have the two mixed up, the gradient reverses the polarity of the rear speaker.
Since the reflection off of the back wall is nearly identical to the source, neither an end fire or gradient low frequency configuration is going to do anything useful, and adding more HF reflections won't reduce your acoustic nightmare.
Oh well..
"What interstings things could be done...", you say?
I have project inspired by recent Genelec offerings. A lot of people point out that it looks like a cardiod subwoofer, but in my testing I did not find any benefits from configuring it as a cardiod sub. The reflections ruin it. All I could do is make really nice nulls on either side!
But yeah, if they can be bi-amped, throw some DSP at them and see what happens. I have had good luck using delays to even out low frequency response. I do this instead of placing multiple subwoofers around the room.
Of course these speakers don't seem to extend much below 100hz, but it's worth experimenting.
https://www.diyaudio.com/community/...on-of-integrating-multiple-subwoofers.401343/
Some example images and data:
End fire cardioid example (setup described here) 4 subs spaced 1m apart in a line. Each sub delayed by multiples of 2.9ms. No phase inversion or other manipulation. Akabak simulation file in zip.
End fire cardioid setup with only 2 subs (just disabled two in above sim):
End fire cardioid example with 2 subs as above, but spaced only 300mm apart with 0.87ms delay (we loose low frequency directivity):
Gradient cardioid example. 2 subs spaced 1m apart. Rear sub has 2.8ms delay and phase inversion. Although we only have 2 subs, the low frequency directivity is maintained but higher frequency control is lost:
Gradient cardioid example same as above but only 300mm spacing and 0.87ms delay. High frequency directivity is maintained as well as low frequency. What we do not see here is that efficiency is lost in the low frequency because cancellation from the out of phase sub is more 'complete'. There is always a trade-off!
Another option is to put the cancellation woofers on the sides of the speaker like the Kii Three does. I'm guessing these simply work like a 'virtual' larger driver and are in phase, possibly with the front woofer delayed. The directivity control would only work in the range where the side woofers are basically omni-directional and able to wrap around the cabinet, likely hence the slim and curved design. There would also be no control below a certain wavelength where they would simply sum nor above a certain wavelength where lobbing would get messy. So the sides would only be bought in at the lower range of the front woofer to support it and increase directivity, a bit like a 2.5-way then switch to dedicated subs once the directivity is lost.
End fire cardioid example (setup described here) 4 subs spaced 1m apart in a line. Each sub delayed by multiples of 2.9ms. No phase inversion or other manipulation. Akabak simulation file in zip.
End fire cardioid setup with only 2 subs (just disabled two in above sim):
End fire cardioid example with 2 subs as above, but spaced only 300mm apart with 0.87ms delay (we loose low frequency directivity):
Gradient cardioid example. 2 subs spaced 1m apart. Rear sub has 2.8ms delay and phase inversion. Although we only have 2 subs, the low frequency directivity is maintained but higher frequency control is lost:
Gradient cardioid example same as above but only 300mm spacing and 0.87ms delay. High frequency directivity is maintained as well as low frequency. What we do not see here is that efficiency is lost in the low frequency because cancellation from the out of phase sub is more 'complete'. There is always a trade-off!
Another option is to put the cancellation woofers on the sides of the speaker like the Kii Three does. I'm guessing these simply work like a 'virtual' larger driver and are in phase, possibly with the front woofer delayed. The directivity control would only work in the range where the side woofers are basically omni-directional and able to wrap around the cabinet, likely hence the slim and curved design. There would also be no control below a certain wavelength where they would simply sum nor above a certain wavelength where lobbing would get messy. So the sides would only be bought in at the lower range of the front woofer to support it and increase directivity, a bit like a 2.5-way then switch to dedicated subs once the directivity is lost.
I don't think I posted any subjective descriptions of the difference. Here goes:
First think I noticed was that the bass seemed to go away, it was like I'd applied a high pass filter on the system. Then I walked to the back of the room, and realized what the issue was; I was listening just 2-3 feet away from the wall where the speakers are located, and because I was so close to the back wall, the amount of bass radiated into the back wall was reduced. I wasn't standing "in a null" but I WAS standing in a zone where the low frequencies were reduced.
I went to the back of the room, where my couch is, about three meters from the speakers. The bass was still quite thin. I made some tweaks to the EQ settings.
I use an EQ curve that's a little bit like what Andy Wehmeyer from Audio Frog recommends for cars; he basically promotes a curve where there's a verrrrrry gentle decrease in output, beginning at the bass and going all the way up to 20khz. I'm doing this off the top of my head, but IIRC, Andy recommends a slope of about 0.5-1.0dB per octave, iirc. This yields a curve where the midrange is zero, 20khz is -5dB, and 20hz is +5db. (This math assumes a slop of 1dB per octave, adjust accordingly if you're using 0.5dB per octave.)
With that curve - which is generally what I run - this "poor man's cardioid" sounded too thin.
So I increased the low frequencies and that helped quite a bit.
This is the part where someone will want to see measurements, and this is a tricky one. Since the "array" is two conventional speakers back to back, in order to measure it outside, it would be tricky. (Would have to match the spacing and the angles 100%.) Also, the back wall and the furniture adjacent to each cardioid pair will change the response.
So I came to the conclusion that I'd just measure the speaker in the room itself, and take a series of measurements in various locations to get an idea of what the in-room response average looks like.
That seemed to work pretty well.
This isn't my measurement. It's from ASR, from a forum member. It's consistent with what I saw when I measured my room, except I didn't have any 'bump' in the low frequencies. (The volume of my room is huge.) I did have a rolloff of the highs, which I corrected with EQ.
Some additional subjective observations:
First think I noticed was that the bass seemed to go away, it was like I'd applied a high pass filter on the system. Then I walked to the back of the room, and realized what the issue was; I was listening just 2-3 feet away from the wall where the speakers are located, and because I was so close to the back wall, the amount of bass radiated into the back wall was reduced. I wasn't standing "in a null" but I WAS standing in a zone where the low frequencies were reduced.
I went to the back of the room, where my couch is, about three meters from the speakers. The bass was still quite thin. I made some tweaks to the EQ settings.
I use an EQ curve that's a little bit like what Andy Wehmeyer from Audio Frog recommends for cars; he basically promotes a curve where there's a verrrrrry gentle decrease in output, beginning at the bass and going all the way up to 20khz. I'm doing this off the top of my head, but IIRC, Andy recommends a slope of about 0.5-1.0dB per octave, iirc. This yields a curve where the midrange is zero, 20khz is -5dB, and 20hz is +5db. (This math assumes a slop of 1dB per octave, adjust accordingly if you're using 0.5dB per octave.)
With that curve - which is generally what I run - this "poor man's cardioid" sounded too thin.
So I increased the low frequencies and that helped quite a bit.
This is the part where someone will want to see measurements, and this is a tricky one. Since the "array" is two conventional speakers back to back, in order to measure it outside, it would be tricky. (Would have to match the spacing and the angles 100%.) Also, the back wall and the furniture adjacent to each cardioid pair will change the response.
So I came to the conclusion that I'd just measure the speaker in the room itself, and take a series of measurements in various locations to get an idea of what the in-room response average looks like.
That seemed to work pretty well.
This isn't my measurement. It's from ASR, from a forum member. It's consistent with what I saw when I measured my room, except I didn't have any 'bump' in the low frequencies. (The volume of my room is huge.) I did have a rolloff of the highs, which I corrected with EQ.
Some additional subjective observations:
- Once the speakers were set up as a poor man's cardioid, AND once addition EQ was applied to the low frequencies, the low frequencies were noticeably more "articulate." Bass guitar riffs are clearer, drums and beats are better defined.
- It was quite shocking how loudly I could play the speaker facing BACKWARDS. I started out with the rear facing speaker in the pair playing about 10dB quieter than the front speaker in the pair, and I was able to increase the volume of the rear-facing speaker higher and higher, with no obviously bad effects. This was a bit of a shock; you'd think a strong reflection off of a very speculative back wall would turn things into an acoustic mess, but it really didn't seem to be the case. I know this gets into the territory of "Bose Direct / Reflecting Speakers," but the effect was not unpleasant. One of the weirdest aspects was that the reflected energy seemed to make the overall sound MORE clear not LESS clear. From what I've read about Geddes and diffraction, I understand that psychoacoustically, early reflections are B-A-D while late reflections can be euphonic. Off the top of my head, I do not recall if there's a consensus when "early reflections" end and "late reflections" begin. 10+ years ago I built some horns that had a LOT of high order modes(1) and my reaction to those HOMsters was that listening to them felt "tense." Very early reflections make the sound reproduction sound abrasive in such a way, that you want to turn down the volume knob or turn the speaker off entirely. This poor man's cardioid DIDN'T do that, and my hunch is that it's because the reflections from the very speculative back wall, they arrive at least 3-5 milliseconds later than the radiation from the front speaker. In my HOMster experiments, the early reflections and diffraction happen in a fraction of a millisecond, just centimeters away from the diaphragm of the tweeter.
What is the front speaker's delay time?
What is the backward speaker spacing from the front, polarity, distance from the wall it faces?
Broadly speaking, when done right this doesn't create a reflection so much as it becomes the reflection, and is no longer a problem (below a certain frequency). I've used subs facing the wall with good results.