Level adjustability for the rear channel / channels is useful, but not completely critical -most never bother but just drive them from the existing stereo amplifier, providing it can handle the extra load & isn't an internally bridged type. When I run those setups, I just take the amplifier pre-out into a second cheap stereo integrated to drive the rear channels.
So actually I do have enough amp channels to try it. I just have to turn my side stereo channels into difference channels, which is easy, and then place them behind me. I don't really have much room behind me but I'll give it a try. I'll see how that compares to the delayed, attenuated widely spaced stereo thing I'm doing now. If it works well I'll probably have to mount the speakers up high, near the ceiling to prevent blocking a path to the garage door.
That was actually one of the more popular configurations if people were running a pair of rear speakers playing the difference channel (mounted high up on the rear wall).
Your other option, which Michael Gerzon was mostly responsible for is a variation on this & width stereo. Standard Hafler configuration with a pair of rear speakers with their negative terminals tied together. The negative between the two is then grounded to the negative amplifier terminals through a variable resistor. At zero nominal value, you've got width / 'spacial stereo' -basically just an extra stereo pair of speakers behind you. At high levels (at least 1.5x the nominal rated impedance of the rear speakers) you've regular Hafler / stereo difference. Somewhere lurking in between, you've got a basic 4.0 quadraphonic setup (this is why it's best done with a separate amplifier for the rear channels, so you've more control) where you've got a hybrid, so there's some stereo effect, but some bleed-through from the opposing channel. It's a bit like a difference signal that has some extra degree of directionality & this is roughly the basis for some of the original quadraphonic matrix systems -basically what ended up being called Dynaquad. And I suppose, since it was Gerzon who arguably did some of the spadework, you could even call it a very distant ancestor of ambisonics. You can replace the variable resistor with a fixed, as was usually done with Dynaquad, but having the ability to adjust it to taste can be useful. If you liked it & found you didn't adjust anything beyond the basic setup, you can always measure off the resistance value & replace with a fixed value.
Your other option, which Michael Gerzon was mostly responsible for is a variation on this & width stereo. Standard Hafler configuration with a pair of rear speakers with their negative terminals tied together. The negative between the two is then grounded to the negative amplifier terminals through a variable resistor. At zero nominal value, you've got width / 'spacial stereo' -basically just an extra stereo pair of speakers behind you. At high levels (at least 1.5x the nominal rated impedance of the rear speakers) you've regular Hafler / stereo difference. Somewhere lurking in between, you've got a basic 4.0 quadraphonic setup (this is why it's best done with a separate amplifier for the rear channels, so you've more control) where you've got a hybrid, so there's some stereo effect, but some bleed-through from the opposing channel. It's a bit like a difference signal that has some extra degree of directionality & this is roughly the basis for some of the original quadraphonic matrix systems -basically what ended up being called Dynaquad. And I suppose, since it was Gerzon who arguably did some of the spadework, you could even call it a very distant ancestor of ambisonics. You can replace the variable resistor with a fixed, as was usually done with Dynaquad, but having the ability to adjust it to taste can be useful. If you liked it & found you didn't adjust anything beyond the basic setup, you can always measure off the resistance value & replace with a fixed value.
I got a phone call today at work from someone who says they invented a new way of using multiple speakers for stereo playback. Hmm. Funny I would be called, of all people, a guy who's been going on for months now on this forum about using multiple speakers for stereo playback. I've got a funny feeling... If it's a prank, that's funny!
For awhile I ran the old Yamaha DSP-100, which derives 4 ambient channels from the main stereo channels. It has dozens of setting, many of them taken from famous concert halls. But I always came back to the simple difference signal from left and right. That's what consistently sounded the most natural.
Looking at the Gerzon arrangement here: https://www.michaelgerzonphotos.org...d a 'sum-and,between' the two stereo channels.
One concern I have is the difference channel in the rear. Do you invert the left or the right? One is going to be in phase with the front and one side channel, the other isn't so it seems assymetrical. I wouldn't have even tried it because of that. Or am I missing something? Maybe it doesn't really matter. Setting up like that is something I might have to do in another room.
One concern I have is the difference channel in the rear. Do you invert the left or the right? One is going to be in phase with the front and one side channel, the other isn't so it seems assymetrical. I wouldn't have even tried it because of that. Or am I missing something? Maybe it doesn't really matter. Setting up like that is something I might have to do in another room.
Not at all (you're not missing anything as such).
At its most basic, the original difference signal configuration simply used a single rear speaker, its terminals wired to the L & R + terminals on the amplifier. That's also what Brian Eno advocated about a dozen years later. If you want to run a pair of rear speakers, most of the time it's just done with their + posts connected to the corresponding + on the amplifier & the - posts on the rear speakers strapped together. But if yo feel inclined to experiment, you can flip their relative polarity to the fronts -just wire the amplifier + outputs to the corresponding - binding posts on the rear speakers, then strap the + binding posts on the rear speakers together. Job-done. If you find the stereo image slightly degraded in either (some do, some don't) then Gerzon's refinement is probably the ultimate version of this configuration -technically better than Dynaquad as it gives a bit of (variable) stereo separation to the rear channels (technically making it a 4.0 setup, although unlike Dynaquad, ambiance extraction is the primary goal).
At its most basic, the original difference signal configuration simply used a single rear speaker, its terminals wired to the L & R + terminals on the amplifier. That's also what Brian Eno advocated about a dozen years later. If you want to run a pair of rear speakers, most of the time it's just done with their + posts connected to the corresponding + on the amplifier & the - posts on the rear speakers strapped together. But if yo feel inclined to experiment, you can flip their relative polarity to the fronts -just wire the amplifier + outputs to the corresponding - binding posts on the rear speakers, then strap the + binding posts on the rear speakers together. Job-done. If you find the stereo image slightly degraded in either (some do, some don't) then Gerzon's refinement is probably the ultimate version of this configuration -technically better than Dynaquad as it gives a bit of (variable) stereo separation to the rear channels (technically making it a 4.0 setup, although unlike Dynaquad, ambiance extraction is the primary goal).
Well, that's fairly straightforward. Matrixing a centre channel usually doesn't help in that sense; it can help dialogue imaging for off-axis viewers, but if you're hoping for enhanced stereo (as in lateral) it will tend to have the opposite effect, whereas some version of a rear difference signal will usually expand things. Technically speaking, on paper it too will reduce the stereo soundstage, usually by about 2dB, but the precedence effect kicks in, especially if you have independent level adjustment of the rear, and the advantage which isn't considered in the basic comparison is that a much broader range of positions in the room will give you a reasonable (even if it isn't technically accurate) stereo image.
If you want a 'one-box' solution though, the easiest is a quasi soundbar type arrangement with an additional lateral (can be angled), rear or up-firing pair of drivers (ditto) at the opposing ends playing the same stereo signal. It isn't as effective as some of the EQ'd / processed variations, but the increased ratio of reflected to direct radiation can provide some of that effect. You can experiment with the effects with a version of the old width stereo: just set up a 2nd pair of speakers laterally to the first, but reasonably close to them, preferably playing at a slightly lower level. You can expand that further with a 3rd pair of speakers. Some prefer them all firing directly ahead, but no harm in experimenting, like a laterally curved concave or convex array, or placing the additional speakers up-firing, laterally or even reversed. The latter can provide some interesting effects if there are decently reflective surfaces available: having speakers facing 'the wrong way' was one of the cultist tweaks in some circles a couple of decades ago -might still be. Easy & flexible way to temporarily see the effects, anyway, before actually committing to a build.
If you want a 'one-box' solution though, the easiest is a quasi soundbar type arrangement with an additional lateral (can be angled), rear or up-firing pair of drivers (ditto) at the opposing ends playing the same stereo signal. It isn't as effective as some of the EQ'd / processed variations, but the increased ratio of reflected to direct radiation can provide some of that effect. You can experiment with the effects with a version of the old width stereo: just set up a 2nd pair of speakers laterally to the first, but reasonably close to them, preferably playing at a slightly lower level. You can expand that further with a 3rd pair of speakers. Some prefer them all firing directly ahead, but no harm in experimenting, like a laterally curved concave or convex array, or placing the additional speakers up-firing, laterally or even reversed. The latter can provide some interesting effects if there are decently reflective surfaces available: having speakers facing 'the wrong way' was one of the cultist tweaks in some circles a couple of decades ago -might still be. Easy & flexible way to temporarily see the effects, anyway, before actually committing to a build.
The array I started this thread on fits the bill for a single console that will throw a wide soundstage, despite having a center channel. The issue is it only works for listeners on-axis. Off axis it collapses down to mono, but at least if they're watching TV, the sounds will stay close to the center. Adding more speakers in various ways can add some vibrance to the room and maintain some sense of stereophonic sound for people sitting off axis.
After listening for a week or two to the enhancements from extra speakers, recursive crosstalk reduction, and some messing with the range above 10k to try to optimize stereo difference, I'm back to just the straight 3 speaker array from 220 Hz on up. It really sounds best without all that extra stuff going on. The reason I was motivated to try this in the first place was to get a pure, clean sound, especially from the center of the sound stage, which I think it does extremely well. Too much messing around for extra width, more stereo effects for off axis, etc. has drawbacks. I noticed I wasn't loving it as much, so I put it back to where I had it last time I loved it, and I love it again! Keep it simple as possible.
I want this "console" for the foot of my narrow bed! Please oh please give me a clear recipe with wiring/circuit diagram (can't Youtube). I followed/searched/read many many threads and got increasingly confused. The simplest experiment without cross-wiring that I can think of uses an array of 6 identical fullrange drivers (no I haven't tried it yet)
R -L L R -R L
As a matter of fact I use a vintage Celestion "console" mono-pair of 8" midwoofer/midtweeter for foot-of-bed stereo. The drivers sounded identical 120-1200hz; the midtweeter was then BSC-EQ filtered (9ohm || 1.5mH@533hz) to match the midwoofer, quite acceptably. Now I have scrounged a stereo fullrange pair of the same "family" and would like to use them together, or at least L C R.
Very much appreciated.
The six identical full range drivers can work. I've tried it putting the drivers on top each other, so L on top left, -R on bottom left, L on top center, R on bottom center, and -L on top right and R on bottom right.
It looks like this
+L +L -L
-R +R +R
So you can see it's all left channels on top with the furthest to the right out of phase, and all right channels on bottom, with the furthest to the left out of phase. As long as you are on-axis and your ears are centered between the top and bottom rows of drivers you'll get a good stereo effect.
That prevents any cross wiring, and doesn't require any line level or digital signal mixing. The downside is you don't benefit from the perfect cancelation you get when you pre-mix the signal before reaching the amplifiers. The side speakers end up working furiously against each other in the bass, so it's a waste of your amplifier's energy.
For line level signal mixing, I think the easiest solution involves using a computer to pre-mix the channels. The way I do it is to use a Mac Mini running Audio Hijack. If you do that, you only need three 8 ohm bookshelf speakers that are all the same, and a regular 2 channel integrated amp that can handle a 4 ohm load.
Use Audio Hijack channel mixers to convert the normal L,R stereo signal to L+R, and L-R signal. Feed L+R to your amplifier's right speaker channel, and L-R to your amplifier's left speaker channel. Run the amplifier right channel to the center speaker, and run the left to the side channels - wired in parallel, with the left side speaker wired in phase and the right side speaker wired out of phase.
The hardest part is getting the channels mixed. I think that's best done digitally. It can also be done very cleanly using an analog mixing circuit but you have to either make one or find some pro audio mixing board that will let you do it. The other hard part is finding 3 identical speakers that are appropriate for the job. If you plan on going DIY, then I do recommend putting all 3 drivers on a single baffle. It keeps them equally spaced apart and lets the whole thing be a single unit.
It looks like this
+L +L -L
-R +R +R
So you can see it's all left channels on top with the furthest to the right out of phase, and all right channels on bottom, with the furthest to the left out of phase. As long as you are on-axis and your ears are centered between the top and bottom rows of drivers you'll get a good stereo effect.
That prevents any cross wiring, and doesn't require any line level or digital signal mixing. The downside is you don't benefit from the perfect cancelation you get when you pre-mix the signal before reaching the amplifiers. The side speakers end up working furiously against each other in the bass, so it's a waste of your amplifier's energy.
For line level signal mixing, I think the easiest solution involves using a computer to pre-mix the channels. The way I do it is to use a Mac Mini running Audio Hijack. If you do that, you only need three 8 ohm bookshelf speakers that are all the same, and a regular 2 channel integrated amp that can handle a 4 ohm load.
Use Audio Hijack channel mixers to convert the normal L,R stereo signal to L+R, and L-R signal. Feed L+R to your amplifier's right speaker channel, and L-R to your amplifier's left speaker channel. Run the amplifier right channel to the center speaker, and run the left to the side channels - wired in parallel, with the left side speaker wired in phase and the right side speaker wired out of phase.
The hardest part is getting the channels mixed. I think that's best done digitally. It can also be done very cleanly using an analog mixing circuit but you have to either make one or find some pro audio mixing board that will let you do it. The other hard part is finding 3 identical speakers that are appropriate for the job. If you plan on going DIY, then I do recommend putting all 3 drivers on a single baffle. It keeps them equally spaced apart and lets the whole thing be a single unit.
Last edited:
https://www.schiit.com/products/syn...twzRWsAF2aR8vW7RNoYemJpWsUlLYskkaAhEcEALw_wcB
Schitt has a cheap new device that creates a center and rear feeds from stereo. Might be an affordable way to experiment. They don't really explain how they're doing it, but I guess it's just a variation of the old school techniques discussed here
Schitt has a cheap new device that creates a center and rear feeds from stereo. Might be an affordable way to experiment. They don't really explain how they're doing it, but I guess it's just a variation of the old school techniques discussed here
Turn the bed around, use the headboard for an OB panel ;')
Unless you were to cross them above the bass, so they're at least neutral. No load at the bass frequencies, no energy consumption.
Hello there ...
Coming in on the 'tail end' of this thread, I am likely to repeat some things already said.
The 'original old matrix' of L+R & L-R is antiquated today and could only serve as low level 'fill-in speakers' if at all.
Before Dolby Digital, one of the greatest achievements of Dolby Labs. was the ProLogic extraction of a center channel.
A lot of very thoughtful & clever research went into it. (before P.L. center the maximum separation was only 6dB)
However, the (bandwidth limited) rear channel left a lot to be desired - but necessary to reduce analogue distortion.
Depending on how many 'pre processors' and amps you want to use, ProLogic center can still play a role in your system.
However, once again, great research has gone into high-end surround equipment to 'bridge the gap' between >
Music vs Cinema via selectable functions.
So, I say, save your penny's and buy the very best Digital gear you can afford
PS.
Be prepared for speakers to 'soak up' a lot of money.
Coming in on the 'tail end' of this thread, I am likely to repeat some things already said.
The 'original old matrix' of L+R & L-R is antiquated today and could only serve as low level 'fill-in speakers' if at all.
Before Dolby Digital, one of the greatest achievements of Dolby Labs. was the ProLogic extraction of a center channel.
A lot of very thoughtful & clever research went into it. (before P.L. center the maximum separation was only 6dB)
However, the (bandwidth limited) rear channel left a lot to be desired - but necessary to reduce analogue distortion.
Depending on how many 'pre processors' and amps you want to use, ProLogic center can still play a role in your system.
However, once again, great research has gone into high-end surround equipment to 'bridge the gap' between >
Music vs Cinema via selectable functions.
So, I say, save your penny's and buy the very best Digital gear you can afford
PS.
Be prepared for speakers to 'soak up' a lot of money.
PS. PS.
I forgot to mention >
Get one of these while they are still cheap & available
https://hifi-wiki.com/index.php/Yamaha_DSP_E580
I forgot to mention >
Get one of these while they are still cheap & available
https://hifi-wiki.com/index.php/Yamaha_DSP_E580
I looked in to that. They do some EQ on the side channels so it's not optimized for the setup I started this thread about.
The surround matrixes are quite a bit different approach than what I've been proposing on this thread. My setup is more of a crosstalk reduction scheme than a channel up-mixing scheme. The Dolby ProLogic is excellent, but very different. Great for increasing the listening area and keeping center vocals in the center. I've worked with it, and concluded that I often prefer stereo upmixed with Dolby ProLogic to discrete channel Atmos mixes, even though the discrete multi-channel mixes are inherently superior, I often don't like the choices that are made when they do the mixing, which ends up being more important to me.Hello there ...
Coming in on the 'tail end' of this thread, I am likely to repeat some things already said.
The 'original old matrix' of L+R & L-R is antiquated today and could only serve as low level 'fill-in speakers' if at all.
Before Dolby Digital, one of the greatest achievements of Dolby Labs. was the ProLogic extraction of a center channel.
A lot of very thoughtful & clever research went into it. (before P.L. center the maximum separation was only 6dB)
However, the (bandwidth limited) rear channel left a lot to be desired - but necessary to reduce analogue distortion.
Depending on how many 'pre processors' and amps you want to use, ProLogic center can still play a role in your system.
However, once again, great research has gone into high-end surround equipment to 'bridge the gap' between >
Music vs Cinema via selectable functions.
So, I say, save your penny's and buy the very best Digital gear you can afford
PS.
Be prepared for speakers to 'soak up' a lot of money.
THANKS for clarifying for me. I can start this "project" by playing with six drivers L-R | L+R | R-L (or eight 2L-R | L+R | 2R-L per "trinaural" link below); then three speakers cross-wired a la Hafler (diagram posted by MITsound in the full-range gallery); then a simple mixer/circuit; then some or other inexpensive Dolby ProLogic II decoder (overkill Arcam AVR, big Yamaha DSP, or smaller but less mainstream device). The Schiit SYN hasn't come to China yet; possibly a crossover to headphone stereo (another project of mine).
https://www.diyaudio.com/community/threads/open-wing-headphone-crossfeed-stereo-sound.391630/
https://www.diyaudio.com/community/...r-3-speaker-stereo-matrix.222881/post-3231194 (Elias)
https://www.diyaudio.com/community/threads/full-range-speaker-photo-gallery.65061/post-7546258 (MITsound)
https://www.diyaudio.com/community/threads/full-range-speaker-photo-gallery.65061/post-7536518 (dcoffin)
I guess, "how to enhance stereo using three contiguous speakers" is a FAQ, but really of little practical/popular/market interest :-( Because most people don't know stereo sound (3D localization), and increasingly so given smartphones & earbuds. Has there ever been a three-speaker LCR boombox? What about driver-array soundbar? Tech probably isn't the biggest challenge....