For my 21st loudspeaker build, I would like to implement Elias Pekonen's SSS.
The concept is very unique and interesting.
dl.dropbox.com/u/2400456/html/Elias_Pekonen/SingleSpeakerStereo.html
I have ordered 3x HiVi B3N / B3S for this endeavour !
I will compare it with my personal reference:
gainphile.blogspot.com.au/2012/03/s19-4-way-dipole-radiator.html
Fun times ahead
2 speaker stereo is dead
The concept is very unique and interesting.
dl.dropbox.com/u/2400456/html/Elias_Pekonen/SingleSpeakerStereo.html
I have ordered 3x HiVi B3N / B3S for this endeavour !
I will compare it with my personal reference:
gainphile.blogspot.com.au/2012/03/s19-4-way-dipole-radiator.html
Fun times ahead
21 is a popular number nowadays
If you got any experience of matrix stereo, you'll already have a hint what to expect. Single speaker stereo is a simulation of three speaker matrix stereo. The room side walls are an essential part of the system:
An externally hosted image should be here but it was not working when we last tested it.
See also my FDTD soundfield simulations of SSS. They'll serve as a proof of it's achieved intended functionality.
Elias Pekonen Home Page - FDTD simulations on Single Speaker Stereo SSS
Yes, fun you'll not be lacking
- Elias
It is a simulation, however, that depends on a directivity of the L and R speakers that is not clearly in evidence, and which is also dependent on a high degree of symmetry in sidewall reflections. An alternate scheme (which would be easy to implement and test) is suggested by your "M/S" example (at you web site). A pair of Magnaplanar MMG dipoles could easily provide a M/S source of well defined directivity . . .
I found an old figure of mine from the Stereolith thread describind MS stereo playback of two crossed dipoles:
An externally hosted image should be here but it was not working when we last tested it.
What matters is the response to the 45 angle direction to generate the side wall reflections towards listening position. We need angle independent summation from the two dipoles. I don't think wide ESL is any good in this regard due to their beaming nature at high freqs.
Anyway, as you know the matrix coefficient of 0.5 is the optimal one, so MS stereo is then.. suboptimal
- Elias
Last edited:
In the two different systems of Elias' post #4, an essential difference is that with the single speaker stereo system, the (L-0.5R) and (R-0.5L) signals are time-delayed in arriving to the listener, relative to the (0.5L + 0.5R) signal. That is, for single speaker stereo set-up, the summed signal has precedence over the difference signals. If the time delay is several milliseconds, then the listener recognizes the sound reflected off of the side walls as non-direct ambient sound and there is no stereo imaging.
It would be great if this really did work, but I don't think so.
Possibly a solution would be to use signal processing to delay the summed signal so that all three signals arrive to the listener simultaneously....
Regards,
Pete
It would be great if this really did work, but I don't think so.
Possibly a solution would be to use signal processing to delay the summed signal so that all three signals arrive to the listener simultaneously....
Regards,
Pete
Last edited:
Why don't you try it and you will change your mind.
There are several aspects you neglect.
First the precedence is not simply an one dimensional phenomena but it includes both time and amplitude, and usually known as time intensity trading. For this reason there is a psychoacoustic filter involved in the speaker which steers the high freq energy from the center to the sides.
More shocking aspect is that precedence effect (Haas) is only valid with one reflection. In a small room acoustic space a cluster of reflections eminating from about the same lateral angle will override Haas effect completely.
There are issues with current psychoacoustic theories within small rooms that need some rethinking. There would be a good opportunity for someone research minded to write a couple of academic papers on these issues (hint)
- Elias
Elias, your post #9,
Colloms cites precedence only of the direct sound in allowing the listener to hear a stereo image (in his book High Performance Loudspeakers). He says that if the listener is 3 to 5 meters distant from stereo speakers, then "most of the sound energy heard is reverberant". The arrival time of the direct "focused" sound ahead of the "diffused" reflected sound he claims is the only thing that allows for the perception of stereo.
I don't understand why the precedence effect should be valid for only a relatively focused reflection where all of it arrives to the listener at about the same time. Even if the arrival times of reflections from a side wall are spread out over a fairly large time interval, still the direct sound of the summation signal arrives prior to any of the reflections off of the side wall.
In the patent literature I've seen a proposal at least similar to your SSS from some time ago. Certainly the fact that your SSS will only work in some listening rooms reduces its commercial viability. But as far as I know a system like your SSS has never been commercialized, suggesting that it is lacking as a method of stereo reproduction.
One other critique that I can make is that reflectivity of the walls of the room would need to be taken into account. That is, I think that the level of the directly radiated summation signal should be reduced relative to the level of the difference signals reflected off of the side walls.
Regards,
Pete
Colloms cites precedence only of the direct sound in allowing the listener to hear a stereo image (in his book High Performance Loudspeakers). He says that if the listener is 3 to 5 meters distant from stereo speakers, then "most of the sound energy heard is reverberant". The arrival time of the direct "focused" sound ahead of the "diffused" reflected sound he claims is the only thing that allows for the perception of stereo.
I don't understand why the precedence effect should be valid for only a relatively focused reflection where all of it arrives to the listener at about the same time. Even if the arrival times of reflections from a side wall are spread out over a fairly large time interval, still the direct sound of the summation signal arrives prior to any of the reflections off of the side wall.
In the patent literature I've seen a proposal at least similar to your SSS from some time ago. Certainly the fact that your SSS will only work in some listening rooms reduces its commercial viability. But as far as I know a system like your SSS has never been commercialized, suggesting that it is lacking as a method of stereo reproduction.
One other critique that I can make is that reflectivity of the walls of the room would need to be taken into account. That is, I think that the level of the directly radiated summation signal should be reduced relative to the level of the difference signals reflected off of the side walls.
Regards,
Pete
Well if I'm lucky the three B3Ns will arrive today or tomorrow
Elias, do you have any guidelines about the front and side baffle dimensions? My plan is basically to come up with whatever aesthetically pleasing (it will stand directly in front of me!) ... and equalise accordingly.
for my part I would say that 23x23x23 cm are minimum dimensions - such is my experience with a simple stereo bipole (that is L and R back-to-back only, without any center or any matrix filtering) with speakers slightly bigger than B3N
I've tried front baffle widths between 20-25 cm. Depth could be a little more than width, currently mine is 33 cm but I suppose 25-35 should be fine. The height has settled automatically for around 40 cm.
Usually I cross around 200-300 Hz to stereo dipole basses. The bass reproduction of the box is not a design goal.
- Elias
Usually I cross around 200-300 Hz to stereo dipole basses. The bass reproduction of the box is not a design goal.
- Elias
Does it prevent you from trying it yourself ?
The dimensions of my listening area aren't compatible for your proposed type of reproduction. Also I assume that any furnishings in a room obstructing sound radiated by the speaker towards the side walls would not be a good thing, making this type of system not so practical for a lot of people.
I'm not suggesting that nobody would like the type of reproduction produced by your SSS. For example, a fair number of people apparently like the reproduction of the Bose 901. As you probably know, in the Bose 901, eight of the nine drivers of the system face away from the listener, while the remaining one faces toward the listener.
Regards,
Pete
This is my thinking of the design currently.
The reason enclosures so thin is to allow addition of 8" woofers (also 3 of them on the front and sides) in case SPL is not enough to transition to the subs.
Subwoofers would be dipoles at 200hz, but eventually will be true cardioid.
gainphile.blogspot.com.au/2012/11/ep21-single-stereo-loudspeakers.html
The reason enclosures so thin is to allow addition of 8" woofers (also 3 of them on the front and sides) in case SPL is not enough to transition to the subs.
Subwoofers would be dipoles at 200hz, but eventually will be true cardioid.
gainphile.blogspot.com.au/2012/11/ep21-single-stereo-loudspeakers.html
Interesting thread.
This PDF of a 1971 article from Ted Jordan may be of interest. It uses a reflector system with two drivers back to back in a central box. The end result is broadly similar to the SSS but with a greater roll-off at HF from the center 'speaker'.
I tried it many years ago and thought it worked quite well, although only with single cone drivers.
http://www.ejjordan.co.uk/PDFs/Jordan_WW_Feb_71.pdf
Using a pair of reflectors would broaden the use of the SSS to rooms which don't have two adjacent parallel walls and the article gives some guidelines re size and materials for the reflectors.
This PDF of a 1971 article from Ted Jordan may be of interest. It uses a reflector system with two drivers back to back in a central box. The end result is broadly similar to the SSS but with a greater roll-off at HF from the center 'speaker'.
I tried it many years ago and thought it worked quite well, although only with single cone drivers.
http://www.ejjordan.co.uk/PDFs/Jordan_WW_Feb_71.pdf
Using a pair of reflectors would broaden the use of the SSS to rooms which don't have two adjacent parallel walls and the article gives some guidelines re size and materials for the reflectors.
Last edited:
There is a thread "discussing" that approach.
All of these unconventional approaches to stereo reproduction can (not necessarily do) sound more realistic when compared to the typical reflection-controlled approach but do they sound better?
In my opinion they do not sound better because they completely ignore how most recordings are created. Just like upmixing stereo for multichannel playback can create a very realistic presentation with certain recordings, it can also completely fail.
Instead of forcing stereo to be what it is not, it probably makes more sense to move on to multichannel which supports a controlled way of delivering a more realistic spatial reproduction.
All of these unconventional approaches to stereo reproduction can (not necessarily do) sound more realistic when compared to the typical reflection-controlled approach but do they sound better?
In my opinion they do not sound better because they completely ignore how most recordings are created. Just like upmixing stereo for multichannel playback can create a very realistic presentation with certain recordings, it can also completely fail.
Instead of forcing stereo to be what it is not, it probably makes more sense to move on to multichannel which supports a controlled way of delivering a more realistic spatial reproduction.
Ah, that's an endless debate -- the feel of realistic vs real precision.
For me, if the presentation is reasonable and life-like, I don't really care it's a copy of the recording scene (or master room) or not.
I believe real multi-channel can be excellent. However, to complete the whole thing -- from recording to playback, there'd be so many more possible mistakes than plain stereo. We can't stop that. (no one can)
In comparison, it's nothing wrong keeping the simple stereo format and improving it somewhat in the range we can.
For me, if the presentation is reasonable and life-like, I don't really care it's a copy of the recording scene (or master room) or not.
I believe real multi-channel can be excellent. However, to complete the whole thing -- from recording to playback, there'd be so many more possible mistakes than plain stereo. We can't stop that. (no one can)
In comparison, it's nothing wrong keeping the simple stereo format and improving it somewhat in the range we can.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.