The opening post is now updated with the images!
Thanks! We are doing hardly any promotion, because right now demand exceeds supply. We are however investing in production capacity and we expect to be able to increase supply considerably in the near future. Then we will also start active promotion. Expect to be seeing several reviews in highly regarded magazines coming out in the coming months!
Great to hear you like it!
I developed a method to optimize the cardioid. I'd love to share it, but unfortunately we decided to keep it a trade secret. We patented the specific configuration of our cardioid midrange enclosure. That document should at least help in the design of a cardiod. If you are working on a project, you could start a thread here and I would love to suggest specific changes you could experiment with.
Directional loudspeaker
EP 3018915 A1
It works really well indeed. They sound good even in very reverberent rooms, such as my own (we moved recently and I will improve the acoustics in the coming months). What most people really like is that - unlike conventional speakers - the 8c can be placed very close to the front wall without any detrimental effect. Actually, it is designed to be close to the wall.
From 100 hz up the 8c has a very constant, unidirectional radiation pattern (see datasheet http://dutchdutch.com/assets/images/8c-spec-sheet.pdf). That means it has constant directivity down to well below the Schroeder frequency in most rooms.
We considered making the bass cardioid too. The reasons we decided not to, are that it costs output, it requires a larger enclosure and that in 'small' domestic rooms a cardioid does not perform better per se. Bass performance is very room and position dependent. Standing waves are the dominant factor. The biggest advantages of a cardioid are in the low midrange and upper bass, where discrete reflections cause distinct peaks and dips in the response.
Inspired by the works of Roy Allisson for Acoustic Research I came up with the idea to place subwoofers on the back of the enclosure and thus close to the front wall. The wall functions as a springboard for the bass, so the speaker and wall basically become one system and their radiation pattern is a hemisphere. Therefore the directivity index in the bass is very similar to the directivity at higher frequencies, which ensures the 8c has a neutral tonal balance in both small and very large rooms.
Thanks! We are doing hardly any promotion, because right now demand exceeds supply. We are however investing in production capacity and we expect to be able to increase supply considerably in the near future. Then we will also start active promotion. Expect to be seeing several reviews in highly regarded magazines coming out in the coming months!
Great to hear you like it!
I developed a method to optimize the cardioid. I'd love to share it, but unfortunately we decided to keep it a trade secret. We patented the specific configuration of our cardioid midrange enclosure. That document should at least help in the design of a cardiod. If you are working on a project, you could start a thread here and I would love to suggest specific changes you could experiment with.
Directional loudspeaker
EP 3018915 A1
Dang.
It works really well indeed. They sound good even in very reverberent rooms, such as my own (we moved recently and I will improve the acoustics in the coming months). What most people really like is that - unlike conventional speakers - the 8c can be placed very close to the front wall without any detrimental effect. Actually, it is designed to be close to the wall.
From 100 hz up the 8c has a very constant, unidirectional radiation pattern (see datasheet http://dutchdutch.com/assets/images/8c-spec-sheet.pdf). That means it has constant directivity down to well below the Schroeder frequency in most rooms.
We considered making the bass cardioid too. The reasons we decided not to, are that it costs output, it requires a larger enclosure and that in 'small' domestic rooms a cardioid does not perform better per se. Bass performance is very room and position dependent. Standing waves are the dominant factor. The biggest advantages of a cardioid are in the low midrange and upper bass, where discrete reflections cause distinct peaks and dips in the response.
Inspired by the works of Roy Allisson for Acoustic Research I came up with the idea to place subwoofers on the back of the enclosure and thus close to the front wall. The wall functions as a springboard for the bass, so the speaker and wall basically become one system and their radiation pattern is a hemisphere. Therefore the directivity index in the bass is very similar to the directivity at higher frequencies, which ensures the 8c has a neutral tonal balance in both small and very large rooms.
Last edited:
The 8c looks like an incredible speaker. Amazing job on the resistive midrange enclosure to get a cardioid response like that.
I do have a question though. As I understand, cardioid speakers should be -6dB down at 90 degrees. However your speaker is around -10dB down at 90 degrees. How are you getting the higher directivity? I'm currently trying to design a speaker that uses multiple drivers and DSP to achieve the cardioid response, but seeing what you've done is making me rethink that.
I do have a question though. As I understand, cardioid speakers should be -6dB down at 90 degrees. However your speaker is around -10dB down at 90 degrees. How are you getting the higher directivity? I'm currently trying to design a speaker that uses multiple drivers and DSP to achieve the cardioid response, but seeing what you've done is making me rethink that.
Last edited:
The directivity is a combination of a cardioid pattern and the directivity you would normally get with a speaker this size. Therefore the directivity is a little bit higher than just a cardioid.
Both paths can lead to a great sounding speaker. Have fun!
Thanks, youknowyou!
keyser, your webpage shows specifications of 8c even if I click for the 8m specs.
Please can you show us the difference of 8c and 8m? obviously the directivity was measured without bass driver (xo 100Hz)
http://dutchdutch.com/assets/images/8c-spec-sheet.pdf
Please can you show us the difference of 8c and 8m? obviously the directivity was measured without bass driver (xo 100Hz)
http://dutchdutch.com/assets/images/8c-spec-sheet.pdf
The opening post is now updated with the images!
Thanks! We are doing hardly any promotion, because right now demand exceeds supply. We are however investing in production capacity and we expect to be able to increase supply considerably in the near future. Then we will also start active promotion. Expect to be seeing several reviews in highly regarded magazines coming out in the coming months!
Great to hear you like it!
I developed a method to optimize the cardioid. I'd love to share it, but unfortunately we decided to keep it a trade secret. We patented the specific configuration of our cardioid midrange enclosure. That document should at least help in the design of a cardiod. If you are working on a project, you could start a thread here and I would love to suggest specific changes you could experiment with.
Directional loudspeaker
EP 3018915 A1
It works really well indeed. They sound good even in very reverberent rooms, such as my own (we moved recently and I will improve the acoustics in the coming months). What most people really like is that - unlike conventional speakers - the 8c can be placed very close to the front wall without any detrimental effect. Actually, it is designed to be close to the wall.
From 100 hz up the 8c has a very constant, unidirectional radiation pattern (see datasheet http://dutchdutch.com/assets/images/8c-spec-sheet.pdf). That means it has constant directivity down to well below the Schroeder frequency in most rooms.
We considered making the bass cardioid too. The reasons we decided not to, are that it costs output, it requires a larger enclosure and that in 'small' domestic rooms a cardioid does not perform better per se. Bass performance is very room and position dependent. Standing waves are the dominant factor. The biggest advantages of a cardioid are in the low midrange and upper bass, where discrete reflections cause distinct peaks and dips in the response.
Inspired by the works of Roy Allisson for Acoustic Research I came up with the idea to place subwoofers on the back of the enclosure and thus close to the front wall. The wall functions as a springboard for the bass, so the speaker and wall basically become one system and their radiation pattern is a hemisphere. Therefore the directivity index in the bass is very similar to the directivity at higher frequencies, which ensures the 8c has a neutral tonal balance in both small and very large rooms.
Gotta admit, I'm kinda stumped on this one. I usually have a knack for reverse engineering patents, but I'm drawing a blank on this. (Way back in 2003-2004, when I was reverse engineering the Unity horn on a car audio form, that was largely based on reading the patents and some 'hints' that Tom Danley posted on the bass list and audio asylum. Plus, the work of John Sheerin.)
An externally hosted image should be here but it was not working when we last tested it.
Okay, here's what I know:
1) From about 1000Hz and up we can control the directivity using the baffle itself, basically it will constrain the radiation to 180 degrees due to it's width. (1000Hz is 34cm long)
2) As the wavelengths exceed 1000Hz the sound will start to wrap around the baffle, but the transition will be fairly gentle due to the roundover
3) In a dipole, the radiation to the sides will cancel because the front and the back are out-of-phase.
OK, so here's my hunch:
When you put a coupling chamber in front or behind a woofer, that coupling chamber forms a low pass filter.
Anyone who's built a Unity or Synergy horn knows this well. The coupling chamber rolls off the high frequencies. But it also introduces a delay that has the effect of moving the midranges and woofers BACKWARDS towards the apex of the horn.
That's why Synergy horns don't need an all-pass filter in the crossover. The acoustic filter introduces a delay.
OK, with me so far?
So in the Dutch & Dutch speaker, if the back chamber was very large the corner of that filter would be very low.
IE, if you put an 18" woofer into a three cubic foot wood box and you stuff it with fiberglass insulation, you're going to get a filter with a corner that's very low and a very long delay with it.
So...
I think that the key to the Dutch & Dutch speaker is juggling the volume of stuffing in the enclosure, and more importantly, the volume of the enclosure.
You basically want to make the volume of the midrange enclosure small enough so that it adds a little delay at 1000Hz, and then more delay at 900Hz, and then more at 800Hz, etc...
The idea is that the stuffing creates a physical low pass filter, similar to an electrical low pass filter, and that filter adds a delay that's frequency dependent.
Juggle the volume of the enclosure and the density of the stuffing and you'll get a delay that gets longer and longer as you go lower and lower in frequency and THAT is how you get these beautiful polar response plots without resorting to six channels of amplification and expensive DSP.
Or at least that's my guess lol
Here's a diagram that illustrates my point from the last post.
In the last post, I suggested that the 'trick' to the Dutch & Dutch speaker is to juggle the volume of the chamber behind the woofer and the density of the stuffing. The idea is to create a low pass filter with a corner that's relatively high.
The reason that we want a low pass filter is because a low pass filter will create a delay that gets longer and longer as the frequencies get lower and lower.
The reason that you want THAT is because if you get the filter shape right, there will be a delay between the front and the back that attenuates the response.
The reason that this is so magical is because it eliminates all the issues you see with dipole speakers from Linkwitz and Kreskovsky. Their speakers are very nice, no doubt, but they require many channels. The reason that the require many channels is because the dipole works over a narrow band. The reason that the dipole works over a narrow band is because the delay between the front and the back of the baffle is FIXED. It's fixed because the DISTANCE between the front and the back is fixed.
An acoustic low pass filter eliminates this drawback. The delay gets longer as the frequencies get longer.
So far so good?
Now if you look at the diagram that I posted above, you can see that a low pass filter creates a delay that is almost constant, when measured in wavelengths.
IE, the delay starts at zero and it grows to 0.3 milliseconds, but if you evaluate that length in WAVELENGTHS the length is almost constant.
Neat, huh?
There's still a lot of work to do though. You would need to figure out the correct volume of the enclosure, what type of stuffing to use, how much to use, the correct corner frequency for that low pass filter, and the correct slope. (I think you want a first order lowpass btw)
I think I'm on the right track here...
By the way, the cardioid speakers from Lexicon, B&O, and Kii Audio, they all work like this. The difference is that Dutch & Dutch came up with a way to do it that doesn't require twenty seven channels of amplification and DSP.
In other words, you could achieve this same type of response by using two woofers in a sealed box and appylying filters to the woofers to control their frequency response and phase.
High five to keyser .... the D&D products looks amazing
Yes very interesting, and especially for us DIY'ers.
Guess it is the same e.g. Amphion and others have been working on for years, but for sure this implementation seems to really excell.
I'm thinking that the side "vents" at the lowest frequencies work much as the Aperiodic Vents from ScnaSpeak, from older days.
Trick is to make a too small blosed box, where you'd get a high Q and a peak at the tuning freq. The vent will then lower the Q by lowering the pressure and thereby also lowering the peak.
Done corectly you can get the perfect 2 order roll off without the peak.
For something like the SB17NRXC35-8 you could go with 3 liter, getting a 3 db peak at 150 Hz. Adding an aperiodic vent you get a perfect roll off at aroud 110 Hz.
For higher frequencies you'll have sound escaping the vent, out of phase with the front, creating the higher directivity (cardioid).
The effect will be less and less as you go higher in frequency as the damping material will have higher and higher effect. The low pass filter effect.
So this must be tried out ... when I get time
Another thought I had was placing two broad band units, onon each side, e.g. Vifa TG9. These could then be both filtered and dealyed to control directivity, and here you would be able to make corection up to a few kHz I would guess.
I guess this is what is done in the Kii. Using the side woofer/subs to correct the directivity of the front firing mid-woofer. But additionally together with the back firing units acting as a cardioid system at very low freqs also ....or this is what I'm guessing.
But if it can work mechanically/acoustically, it is much cheaper and much less complex.
Yes very interesting, and especially for us DIY'ers.
Guess it is the same e.g. Amphion and others have been working on for years, but for sure this implementation seems to really excell.
I'm thinking that the side "vents" at the lowest frequencies work much as the Aperiodic Vents from ScnaSpeak, from older days.
Trick is to make a too small blosed box, where you'd get a high Q and a peak at the tuning freq. The vent will then lower the Q by lowering the pressure and thereby also lowering the peak.
Done corectly you can get the perfect 2 order roll off without the peak.
For something like the SB17NRXC35-8 you could go with 3 liter, getting a 3 db peak at 150 Hz. Adding an aperiodic vent you get a perfect roll off at aroud 110 Hz.
For higher frequencies you'll have sound escaping the vent, out of phase with the front, creating the higher directivity (cardioid).
The effect will be less and less as you go higher in frequency as the damping material will have higher and higher effect. The low pass filter effect.
So this must be tried out ... when I get time
Another thought I had was placing two broad band units, onon each side, e.g. Vifa TG9. These could then be both filtered and dealyed to control directivity, and here you would be able to make corection up to a few kHz I would guess.
I guess this is what is done in the Kii. Using the side woofer/subs to correct the directivity of the front firing mid-woofer. But additionally together with the back firing units acting as a cardioid system at very low freqs also ....or this is what I'm guessing.
But if it can work mechanically/acoustically, it is much cheaper and much less complex.
An externally hosted image should be here but it was not working when we last tested it.
Here's another idea I had about this box. This idea is simpler than what I posted five months ago.
If you look at the original box and the new box, there are a couple of changes:
1) There are far fewer ports
2) The ports are relatively close to the woofer.
I'd invested a lot of time tinkering with this idea in Hornresp, but Hornresp can't measure a fundamental aspect of Keyser's design:
Not all of the sound radiated in the box will escape from the port immediately.
For instance, in the Dutch and Dutch speaker, the pathlength from the front of the cone to the back of the cone is about 60cm.
In a conventional dipole, the pathlength difference will have a couple of effects. First, a big peak that occurs when the front and back are in phase, "the dipole peak." And then a rolloff below the peak.
One of the things that I couldn't figure out about this passive design was how is it working across a broad range of frequencies? And I think that the answer is "because not all of the sound inside the box escapes immediately."
To get good cardioid behavior, we need the sound emitted from the ports to be delayed with frequency. IE, it's not sufficient if the delay is one millisecond across the board; with every octave lower that it plays, the delay must grow longer.
But since the sound can't escape immediately, that may be a passive way to generate the delay. Some sound escapes the box immediately, some sound escapes the box after a single reflection, some takes a few reflections.
BTW I can't think of any easy way to simulate this! We're basically talking about a complex interaction of the following:
1) the resistivity of the stuffing
2) the location of the ports
3) the diameter of the ports
4) the distance from the front to the back of the enclosure
High five to keyser .... the D&D products looks amazing
Thanks Baldin!
So this must be tried out ... when I get time
It's a lot of fun and a great challenge to design a good acoustic/mechanical cardioid, but I have to say it's quite difficult to get it really right. You have to realize that in order to make a proper cardioid, you need to match both the amplitude and phase of the front-wave and rear-wave. It's very difficult to do that with a single driver.
When you do get it right, its cardioid performance is similar to an active solution, but you have the added benefit of properly dealing with the back-wave. With an active cardioid you're putting a lot of acoustic energy into a sealed enclosure (two or more drivers instead of one), energy that you have to deal with or it will cause colorations.
This is the approach I would recommend for a DIY project. It's still a good challenge, but easier to pull off.
There's acoustic/damping material inside the enclosure. In case of the Olympus speaker (this thread) it was rockwool. At low frequencies / long waves the sound pressure kind of takes the shortest route out of the enclosure. Above a couple of hundred hertz almost all sound is damped.
This is indeed the greatest challenge. Basically, in the lower end of its passband the 8c midrange is a pure cardioid. Then as the directivity of the driver and baffle edge diffraction starts to increase, the output of the slots requires a low-pass filter that matches the decrease in output to the rear. At the top of the midrange drivers's passsband the directivity is no longer the result of the cardioid configuration, but just the driver itself and edge diffraction.
Actually, a theoretical cardioid consists of two point sources spaced some distance apart. The point source in the rear has inverted polarity and a delay equivalent to the propagation delay between the two sources. So no frequency-dependent delays.
The internal dimensions of the cardioid enclosure are much smaller than the wavelengths by the way, so it's difficult to think in terms of 'reflections' inside.
This is the first cardioid-like speaker I ever built. It really was a matter of trial & error in the beginning. I made holes in the sides, top and back panel. Later I replaced the back panel with a panel with no holes. That improved things considerably. As you know, the latest designs have a single port closer to the front. In this 6moons review you can find some photo's of the construction.
There are some independent measurements of 8c over at Soundstage:
SoundStageNetwork.com | SoundStage.com - NRC Measurements: Dutch & Dutch 8c Loudspeakers
While everything measures more or less the way i expected (very good), backfiring subs puzzle me. What's with all that hash only 15dB bellow the midrange level that spans 500-1200Hz and then drops to -20dB and covers rest of the frequency range ?
SoundStageNetwork.com | SoundStage.com - NRC Measurements: Dutch & Dutch 8c Loudspeakers
While everything measures more or less the way i expected (very good), backfiring subs puzzle me. What's with all that hash only 15dB bellow the midrange level that spans 500-1200Hz and then drops to -20dB and covers rest of the frequency range ?
Last edited:
Don’t forget you a get a high quality DAC, DSP and 6 channels of amplification included in that price
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.