Reading about cardioid speakers and all the benefits got me thinking of experimenting with my next speaker. I have a speaker in design with front and back woofers crossed over to a midrange at 400Hz. I want to achieve cardioid pattern from 400Hz down to at least 70-80Hz for the woofer section with FIR filtering.
A few questions that I can't seem to find answers to.
Since the rear radiation is mostly cancelled out, how much maximum output do I lose with a cardioid speaker vs both woofers playing in phase? I'm hoping I don't lose 6dB of bass from the loss of one woofer (being delayed to cancel rear radiation) by going cardioid.
Does it work in a room? Everything I've read says cardioid is great for outdoor concerts so the performers aren't blasted by all the bass. What about indoors?
Thanks
A few questions that I can't seem to find answers to.
Since the rear radiation is mostly cancelled out, how much maximum output do I lose with a cardioid speaker vs both woofers playing in phase? I'm hoping I don't lose 6dB of bass from the loss of one woofer (being delayed to cancel rear radiation) by going cardioid.
Does it work in a room? Everything I've read says cardioid is great for outdoor concerts so the performers aren't blasted by all the bass. What about indoors?
Thanks
How much you lose depends on the separation distance between the woofers. Cardioid is the sum of 1/2 dipole and 1/2 monopole. Calculate the maximum distance allowable for a 400 Hz crossover and then calculate how much cancellation you get in front of the speaker. Expect something in the range of 10 dB.
Sorry if I wasn't clear before. I wasn't wondering on the amount of cancellation on the rear. I'm wondering how much output in the front direction I'll lose. I'll normally gain 6dB of maximum output with a second woofer. But since the second woofer is now used to turn the woofer region into a cardioid, I'm wondering how much output I'll lose compared to 2 normal woofers playing. Or, how much output I'll gain (if any) over one woofer.
A cardioid woofer equalized to the same axial SPL as a monopole will radiate 4.77dB less acoustic power. However, a cardioid also has an efficiently that drops off 6dB/octave relative to the frequency where the unequaled cardioid has the same axial response as a monopole. Thus, an unequaled cardioid will radiate 4.77dB less power than a monopole at the frequency where they have the same axial SPL, 10.77 dB less power than the monopole an octave lower, and 16.77 db less power 2 octave lower, etc. This is compared to a single monopole woofer. Of course, if you build the cardioid using two identical woofers you will necessarily be supply 3db more power to the unequaled cardioid to start with, so add another 3dB to those power looses.
Last edited:
The 6dB/octave roll off sounds like the behavioural of a dipole, but I remember reading that the cardioid has more output than a dipole. What's the reason behind having the 6dB/octave roll off for a cardioid?
Thanks
Last edited:
The cardioid and dipole have exactly the same directivity index, meaning for the same axial SPL they both radiate 4.77 dB less power than a monopole. Dipole and cardioid are both gradient systems and roll off at 6dB/octave. However, due to the delay added to the rear source of a cardioid the frequency at which the unequalized dipole and cardioid achieve the same axial amplitude and the single monopole source is one octave lower for a cardioid compared to a dipole when the separation between the two sources is the same. For this reason, a cardioid, with the same separation as a dipole, is 6dB more efficient than the dipole.
Thank you for the explanation.
Do you have a good source of where I can learn a lot more about cardioid speakers? I'm more and more interested of this concept.
But before I do, this question will determine how much effort I want to put into this, because my baffle is extremely small at 6.25" x 11". This means I'll start experiencing the 6db/oct roll off somewhere around 400Hz by a rough estimate. If I want to maintain cardioid pattern to 100Hz (and go back to bipole below 100Hz), I'll lose up to 12dB of efficiency, which means my speaker is limited to about 99dB of max SPL around 100Hz in cardioid mode, which isn't very high.
It is still unintuitive to me why there exists a 6dB/oct roll off below "Fequal". I'm wondering, is there any way to get around the 6dB/oct efficiency loss? Could the use of a carefully designed variable delay FIR filters possibly negate the 6dB/oct roll off? The Kii Audio Three holds cardioid pattern down to 40Hz. I can't imagine them throwing away 15dB or so of bass output just to hold the cardioid pattern. That's a huge sacrifice better suited with subwoofers than a smallish speaker with 6" woofers
Last edited:
An externally hosted image should be here but it was not working when we last tested it.
There are four drivers in Kii Audio's cardioid array.
There's a thousand watts spread out over those four drivers.
If I'm not mistaken, the following speakers all work the same way, at least on the horizontal axis:
1) Beolab 90
2) Kii Audio Three
3) EAW Otto
4) Lexicon Soundsteer
Arguably, the Lexicon is the most sophisticated, as it controls directivity horizontally AND vertically, using two different technologies.
I did a thread about them here: http://www.diyaudio.com/forums/multi-way/302839-beer-budget-beam-steering.html
I need to clarify something:
Is the 6dB/oct roll off simply less efficient, or is the maximum xmax limited output reduced at 6dB/oct as well? I.e., If Fequal is 100Hz for a dual opposed sealed subwoofer, and the maximum Xmax limited output at 50Hz is 100dB at 100W. If the subwoofer is in "cardioid mode", does that mean the *Xmax* limited output is 94dB at 50Hz, or 100dB at 50Hz, but requires 400W?
Is the 6dB/oct roll off simply less efficient, or is the maximum xmax limited output reduced at 6dB/oct as well? I.e., If Fequal is 100Hz for a dual opposed sealed subwoofer, and the maximum Xmax limited output at 50Hz is 100dB at 100W. If the subwoofer is in "cardioid mode", does that mean the *Xmax* limited output is 94dB at 50Hz, or 100dB at 50Hz, but requires 400W?
Yes. If you know the max or displacement limited SPL at Fequal then below that the displacement limited SPL will also drop at 6dB/octave compared to the monopole. You can not avoid the roil off as it is a consequence of phase difference between the the sources resulting from the path length differences and the necessary delay. That is why dipoles and cardioids are referred to a 1 st order gradient systems. The axial response rolls off at 6dB/octave.
If it is only phase, doesn't that mean we can correct it with FIR filters? I can design a FIR filter where the phase is progressively increased as frequency decrease with just the right phase difference for every frequency due to the path length differences.
There's no free lunch.
The phase difference between the two radiators is what creates the cardioid response.
If you want directivity control with maximum output, you want to use waveguides.
The phase difference between the two radiators is what creates the cardioid response.
If you want directivity control with maximum output, you want to use waveguides.
Aww, I like free lunches! 😛
If that's the case, then the Kii Three is indeed throwing away around 10dB or so of output. It basically means the 4 woofers really only have the output of about 1 woofer. That's a huge sacrifice! However, the reviews say the bass output and maximum SPL of the entire speaker is very high, doesn't seem like that's achievable with just the output of one sealed 6'' woofer.
Last edited:
Looks like the Kii Three only holds cardioid pattern to 80Hz, which is a lot more reasonable. The 40Hz is an erroneous figure from a reviewer.
OK, I'm going to try cardioid on my speaker. Now the hard part: I can hardly find any information on making a cardioid speaker. I'm reading conflicting things on how to make cardioid speakers. Linkwitz says just 2 monopoles back to back where the rear is delayed by the period of their distance D. But others say that is only cardioid at a small range of frequencies and a variable delay in order to achieve cardioid at a wide range of frequencies. I can't find more information to figure which is correct myself.
I can do both methods above since I have a FIR capable DSP. I want to have the woofers radiating cardioid from 100Hz and up, say 1000Hz. How should I approach making the woofers radiate cardioid?
I can do both methods above since I have a FIR capable DSP. I want to have the woofers radiating cardioid from 100Hz and up, say 1000Hz. How should I approach making the woofers radiate cardioid?
- Status
- Not open for further replies.