Hello and how do you do !
A few years ago I presented the theory and operation of matrixed single stereo loudspeaker
http://www.diyaudio.com/forums/multi-way/200040-stereophonic-sound-single-loudspeaker.html
Just recently I discovered that late Nagaoka has left us with a huge heritage of single stereo matrix loudspeakers
Basically he used two kind of matrixes. For 3 and for 4 elements.
The 4 element matrix looks like this:
Notice that this is MS matrix. The front elements sum acoustically to form L+R signal.
hkjunk0.web.fc2.com/hobby/audio/matrix_db/matrix02.jpg
And the 3 element matrix is like this:
We can see it is the famous x=0.5 matrix
(Notice though the left amplifier channel negative terminal is left unconnected, possibly to avoid a ground loop?)
hkjunk0.web.fc2.com/hobby/audio/matrix_db/matrix03.jpg
{HIRO}'s Website!! - matrix_db1
(edit: those hkjunk images are pain to see online, just copy the link to an empty window)
.
A few years ago I presented the theory and operation of matrixed single stereo loudspeaker
http://www.diyaudio.com/forums/multi-way/200040-stereophonic-sound-single-loudspeaker.html
Just recently I discovered that late Nagaoka has left us with a huge heritage of single stereo matrix loudspeakers
Basically he used two kind of matrixes. For 3 and for 4 elements.
The 4 element matrix looks like this:
Notice that this is MS matrix. The front elements sum acoustically to form L+R signal.
hkjunk0.web.fc2.com/hobby/audio/matrix_db/matrix02.jpg
And the 3 element matrix is like this:
We can see it is the famous x=0.5 matrix
(Notice though the left amplifier channel negative terminal is left unconnected, possibly to avoid a ground loop?)
hkjunk0.web.fc2.com/hobby/audio/matrix_db/matrix03.jpg
{HIRO}'s Website!! - matrix_db1
(edit: those hkjunk images are pain to see online, just copy the link to an empty window)
.
Attachments
Last edited:
Nagaoka MX-1
First such a matrix single stereo speaker was MX-1, using 4 element matrix
from here:
2009?8? - COLKID?????
Construction and the matrix connection:
from here:
ƒXƒs�[ƒJ�[�FƒI�[ƒfƒBƒI‘ÌŒ±�î•ñ
.
First such a matrix single stereo speaker was MX-1, using 4 element matrix
An externally hosted image should be here but it was not working when we last tested it.
from here:
2009?8? - COLKID?????
Construction and the matrix connection:
An externally hosted image should be here but it was not working when we last tested it.
from here:
ƒXƒs�[ƒJ�[�FƒI�[ƒfƒBƒI‘ÌŒ±�î•ñ
.
Looks like MX-1 was quite popular, at least in Japan.
from here:
’·‰ª“S’j‚³‚ñ
from here:
Whatnew�i‚«‚Ü‚®‚ê“ú‹L�j
from here:
ƒXƒeƒŒƒI
from here:
2????????? | ???????????? - ?????
from here:
http://itsouken.exblog.jp/19595281/
.
from here:
’·‰ª“S’j‚³‚ñ
from here:
Whatnew�i‚«‚Ü‚®‚ê“ú‹L�j
from here:
ƒXƒeƒŒƒI
from here:
2????????? | ???????????? - ?????
from here:
http://itsouken.exblog.jp/19595281/
.
Last edited:
in a book "Original speakers work of 45 Nagaoka Tetsuo" (google translation.)
Nagaoka MX-10
There was an improved version of MX-1 called MX-10, with more bass output apparently using some sort of reflex port extension hidden in the base.
The 4 element matrix was the same as in MX-1.
from here:
.page besides SPI sound plaza Ida / handmade speaker basic lecture (6)/wonderful speaker.page
from here:
�ì•i�W2
from here:
FE103-Sol???????? | Phile-web??????
.
There was an improved version of MX-1 called MX-10, with more bass output apparently using some sort of reflex port extension hidden in the base.
The 4 element matrix was the same as in MX-1.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
from here:
.page besides SPI sound plaza Ida / handmade speaker basic lecture (6)/wonderful speaker.page
An externally hosted image should be here but it was not working when we last tested it.
from here:
�ì•i�W2
from here:
FE103-Sol???????? | Phile-web??????
.
Interesting is the Nagaoka MX-15 speaker with 3 element matrix. Matrix is x=0.5.
http://stat001.ameba.jp/user_images/20140212/00/jawapannacotta/62/a3/j/o0800059812843418391.jpg
from here:
???????(MX-15)???????????????
hkjunk0.web.fc2.com/audio/speaker/mx-15/mx-15_03.jpg
from here:
{HIRO}'s Website!! - mx-15
from here:
‚l‚w�|‚P‚T‚à‚Ç‚«‚R–{ƒ}ƒgƒŠƒNƒXƒXƒs�[ƒJ�[‚Ì�»�ì�@‚S / ‰Á“¡‚¿‚á‚ñ‚Ì“ú‹L
.
http://stat001.ameba.jp/user_images/20140212/00/jawapannacotta/62/a3/j/o0800059812843418391.jpg
from here:
???????(MX-15)???????????????
hkjunk0.web.fc2.com/audio/speaker/mx-15/mx-15_03.jpg
from here:
{HIRO}'s Website!! - mx-15
from here:
‚l‚w�|‚P‚T‚à‚Ç‚«‚R–{ƒ}ƒgƒŠƒNƒXƒXƒs�[ƒJ�[‚Ì�»�ì�@‚S / ‰Á“¡‚¿‚á‚ñ‚Ì“ú‹L
.
This does look very interesting and easy to try. I finally have the "lunch money" class A/B TDA7297 amp and a bunch of full range drivers lying around. Just need a box - foam core will do. Can they all share the same volume?
Can someone describe qualitative listening impressions of what they sound like?
Can someone describe qualitative listening impressions of what they sound like?
Yes the elements share the same volume.
I recommend using the 3 element version. That is what I've had very good results.
For starters any box will do. Here's my first prototype some years ago
An externally hosted image should be here but it was not working when we last tested it.
It was a sort of "foam core" as well
An externally hosted image should be here but it was not working when we last tested it.
In the past I wrote some listening impressions on 3 element matrix (then using Alpair 5 full range driver with additional mono bass). See here somewhere in the middle of the page "Subjective Listening Impressions of this Prototype of SSSx5"
Elias Pekonen Home Page - Stereophonic Sound from a Single Speaker
.
Do you mean the 3 element matrix ?
6 dB is the channel separation of this passive matrix, and it is also the maximum electrical separation that can be achieved with passive matrix.
The speaker signals are
Ls = L - 0.5*R
Cs = 0.5*(L + R)
Rs = R - 0.5*L
where Ls, Cs, and Rs are the left, center and right speaker signals. And L and R are the left and right stereo signals.
For mono signal we have L=1 and R=1
in this case speaker signals will be
Ls = 0.5
Cs = 1
Rs = 0.5
In this case center element is 6dB louder than side elements, and sound is being directed straight towards listener.
For left side panned signal we have L=1 and R=0
in this case speaker signals will be
Ls = 1
Cs = 0.5
Rs = -0.5
In this case the left element signal is 6dB louder than center and right element signals, and sound is directed towards left side of the room.
Of course this kind of single speaker stereo does not work without room side wall reflections ! That is clear. But usually rooms have walls ...
See here for FDTD wave field simulations of different scenarios
Elias Pekonen Home Page - FDTD simulations on Single Speaker Stereo SSS
For example, here left panned stereo signal at 1000Hz:
An externally hosted image should be here but it was not working when we last tested it.
.
Yes the elements share the same volume.
I recommend using the 3 element version. That is what I've had very good results.
For starters any box will do. Here's my first prototype some years ago
An externally hosted image should be here but it was not working when we last tested it.
It was a sort of "foam core" as well
An externally hosted image should be here but it was not working when we last tested it.
In the past I wrote some listening impressions on 3 element matrix (then using Alpair 5 full range driver with additional mono bass). See here somewhere in the middle of the page "Subjective Listening Impressions of this Prototype of SSSx5"
Elias Pekonen Home Page - Stereophonic Sound from a Single Speaker
.
Great website on this! Thanks for putting it all together and after reading your listening impressions and seeing the ability to get a stable stereo phantom image with 90 deg of head turning, I am convinced that I have to try it. I will make it with a 15 liter box and qnty 3 Vifa TC9FD's. Need to find a capacitor and resistor for the psychoacoustic filter.
I have several small speakers already in their own boxes, could I not place three boxes arranged in same manner and connect them electrically in the same way to achieve this? Two Karlsonator 0.4x boxes for the side firing and a sealed spiral TL for the center. Hmm? This would just make it an electrical connection exercise to start out with.
Last edited:
Initial test using 2x 0.4x Karlsonator & Nautaloss II
As an inital test to hear how this sounds, I decided to use existing small speakers arranged in LCR at 90 deg to each other.
I am using a 0.4x scale Karlsonator for the sides and the Nautatloss II (dual driver) for the center. For the divider network, I am using this circuit which has a psychoacoustic filter for the center channel:
with a 5 ohm and a bunch of caps added together to make 4 uF. I used a small PC board and soldered the connections so that flying leads can be easily routed to the three speakers. Glad I used the PC board - there are quite a lot of connections even though it looks simple. I connected everything together and am driving it with the "Lunch Money" class AB chip amp - the TDA7297.
Well, I hooked it all together and fired it up with some test tracks...
It sounds very nice! The imaging of the instruments in the Dave Brubeck band song were very convincing, although in an A/B test with true stereo separate speakers, that was better. However... I have to say that it is great if you want to move about in the room. The sweet spot is HUGE!
I really am enjoying this and can highly recommend this to others to try. Now that I know the topology works I will make a dedicated box. I actually like having the dual drivers in the front - they are running in parallel so have +6dB boost over the sides but I actually like that given the filter values I am using.
One thing that is a little pesky is that large percussive transients generate a click - almost like the amp is clipping although I know I am not driving it hard. need to debug that...
A note on a possible enclosure would be to make a Karlsonator scaled for qnty 4 Vifa drivers, and this will make large amounts of bass. Mount two drivers facing the front like the 0.53x scale Karlsonator but with two drivers side facing. It will probably be about 0.65x scale but that should give a big sound and no need for a sub as it will reach 50 Hz.
Thanks for the tips Elias.
As an inital test to hear how this sounds, I decided to use existing small speakers arranged in LCR at 90 deg to each other.
I am using a 0.4x scale Karlsonator for the sides and the Nautatloss II (dual driver) for the center. For the divider network, I am using this circuit which has a psychoacoustic filter for the center channel:
An externally hosted image should be here but it was not working when we last tested it.
with a 5 ohm and a bunch of caps added together to make 4 uF. I used a small PC board and soldered the connections so that flying leads can be easily routed to the three speakers. Glad I used the PC board - there are quite a lot of connections even though it looks simple. I connected everything together and am driving it with the "Lunch Money" class AB chip amp - the TDA7297.
Well, I hooked it all together and fired it up with some test tracks...
It sounds very nice! The imaging of the instruments in the Dave Brubeck band song were very convincing, although in an A/B test with true stereo separate speakers, that was better. However... I have to say that it is great if you want to move about in the room. The sweet spot is HUGE!
I really am enjoying this and can highly recommend this to others to try. Now that I know the topology works I will make a dedicated box. I actually like having the dual drivers in the front - they are running in parallel so have +6dB boost over the sides but I actually like that given the filter values I am using.
One thing that is a little pesky is that large percussive transients generate a click - almost like the amp is clipping although I know I am not driving it hard. need to debug that...
A note on a possible enclosure would be to make a Karlsonator scaled for qnty 4 Vifa drivers, and this will make large amounts of bass. Mount two drivers facing the front like the 0.53x scale Karlsonator but with two drivers side facing. It will probably be about 0.65x scale but that should give a big sound and no need for a sub as it will reach 50 Hz.
Thanks for the tips Elias.
Attachments
Last edited:
Ask Wesayao to build a third tower and connect them together in a central Tri-Tower-of-Power in his house.
These look like they would work very well in a line array: Peerless TC6WD02-04 Silver Series 2" Midrange
You would need 32 of them to go floor to ceiling at CTC of 3 in x 3 or about 100 drivers, $500 total in drivers. It would indeed sound very room-filling.
These look like they would work very well in a line array: Peerless TC6WD02-04 Silver Series 2" Midrange
You would need 32 of them to go floor to ceiling at CTC of 3 in x 3 or about 100 drivers, $500 total in drivers. It would indeed sound very room-filling.
This setup reminds me of the pseudo quadro from the 70ies/80ies
sounded nice but channel separation was crap.
Well of course it's not going to be like two independent speakers which can achieve nearly infinite separation in an anechoic chamber. However, with room walls reflecting and mixing the left and right channels in a typical room I doubt that the separation is much better than 20 dB.
In the simulations of a hard left pan, Elias shows that you can get some pretty clean separation - about 10 to 15 dB it looks like I I am reading the contour of correctly.
An externally hosted image should be here but it was not working when we last tested it.
It's not for critical listening but for some applications like an mp3 kitchen speaker or clock radio - this is the way to do it and still keep a slim footprint. It certainly works better than trying to get stereo imaging with 2 conventional speakers spaced 2 ft apart like in a typical boombox.
Last edited:
Von Ah,
My guess is that this is a class D output stage from the internal photo showing 12 coil chokes and filter capacitor banks. Most class D are bridge tied loads and will not work. I had the same problem since I was extensively using the TPA3116 amp. I ended up buying a couple of the TDA7297 chip amps (15 watts/ch) for $5 ea. They sound very nice and are class AB and can be tied into a matrix as described above. Even if they burnt out $5 is not a big deal but I am finding lots of uses for these amps.
X
Lunch money amp: http://m.aliexpress.com/item/1590579320.html?tracelog=wwwdetail2mobilesitedetail
My guess is that this is a class D output stage from the internal photo showing 12 coil chokes and filter capacitor banks. Most class D are bridge tied loads and will not work. I had the same problem since I was extensively using the TPA3116 amp. I ended up buying a couple of the TDA7297 chip amps (15 watts/ch) for $5 ea. They sound very nice and are class AB and can be tied into a matrix as described above. Even if they burnt out $5 is not a big deal but I am finding lots of uses for these amps.
X
An externally hosted image should be here but it was not working when we last tested it.
Lunch money amp: http://m.aliexpress.com/item/1590579320.html?tracelog=wwwdetail2mobilesitedetail
Last edited: