Convert stereo to LCR - feasible or even desirable?

[FreakyKiwi]

There is an easier way.

I 'm using some spare infinity LCR speakers from a HT system for my PC and have directly wired all three to a stereo amp.

L+ and L to the left speaker
R+ and R- to the right speaker
L+ and R+ to the centre speaker

This means that the centre speakers plays the difference between the left and right so fills in the sound nicely.

I'm did the same in my wife's Mercedes SLK when I replaced the broken factory 7 channel amp with an after market head unit and separate 4 channel amp in order to connect the tiny centre speaker on the dash.

There's a name for this solution but I can't remember it.

 

[gnarly]

My question now is: Could it be a good idea — and even possible — to convert recorded stereo (L/R) signals into three signals (L/C/R)? And how would you even do that? I am mainly thinking about processing digital signals in real time.
My initial idea was to send L+R to the center, L−R to the left, and R−L to the right, but I suspect there might be issues with phase and levels.

What is your take on this?

My take is it can work very well with the right speaker setup and the right DSP to implement.
For me, the right speaker setup is 3 identical speakers each on its own sub. Essential to have speakers at an equidistant arc from listening position, or with center delayed to accomplish same thing.
DSP is a drag and drop prosound unit that various LCR matrices can be placed using summing mixers. Such as Biamp, BSS, Symmetrix, QSYS.

Here's a good site on LCR matrices...http://elias.altervista.org/html/3_speaker_matrix.html
With snips from that site showing the matrices I use most often, given the angles my speakers are at. Versions of the Trifield that's been mentioned.

You can see how the coefficients vary depending on listening angle. It's what makes the matrices work much better than simpler L & R summing without weightings.

It's always a try and see, in terms of does LCR sound better than stereo. I'd say 2/3 of the time it does.
But really well done stereo is hard to beat imo/ime...problem is not all stereo is so really well done!! lol

 

[Boye]

My take is it can work very well with the right speaker setup and the right DSP to implement.
For me, the right speaker setup is 3 identical speakers each on its own sub. Essential to have speakers at an equidistant arc from listening position, or with center delayed to accomplish same thing.
DSP is a drag and drop prosound unit that various LCR matrices can be placed using summing mixers. Such as Biamp, BSS, Symmetrix, QSYS.

Thanks! This is what I was looking for. I didn't imagine you could improve much on listening to a good stereo recording sitting in the sweet spot.
(The finest experience was a stereo setup with Quad ESL 63s with dipole subwoofers. All phase corrected and filterede with Acourate. That was the most holographic sound image I have ever experienced. The downside was a very small sweet spot.)
My idea was to stop the center image from moving to the side for listeners sitting off-center, thereby retaining a stable phantom image of instruments placed in the middle. I have no idea of how this might change the perceived width and depth of the sound stage as well as location of instrument groups, though.

 

[Boye]

There is an easier way.
I 'm using some spare infinity LCR speakers from a HT system for my PC and have directly wired all three to a stereo amp. [...]
This means that the centre speakers plays the difference between the left and right so fills in the sound nicely.

That reminds me of Brian Eno's suggestion for how to listen to his album "Ambient 4: On Land". Hovever, the difference signal was supposed to go to a rear speaker.
It doesn't seem to me that a difference signal would improve the positioning of the (mono) phantom center. I mean, if there is a singer placed in the middle, it would be a mono signal, but no sound would be sent from a difference signal. My idea seems to be somewhat the opposite: extracting the mono part and send that to the center, and the let left and right speaker add the necessary stereo information.

 

[john61ct]

Yes I believe how Schiit SYN works

from my past notes:

"ambiance extraction" "Hafler matrix circuit" as implemented by Jason Stoddard

merges the in-common content of the L+R channels to get the mono center channel

subtract the channels' center (esp vocals) to extract natural ambience for rear surround

SYN surround left really is just "L minus R" and SYN surround right really is just "R minus L"

rears have the difference signal of the left and right fed to them.

left and right rear speakers out of phase with each other

delivers 5.1 output: front left and right, center (vocals centered FRONT), rear left and right, and subwoofer

when no rear surround required, just use front system, main pair + center, 3.0 or with subwoofer for 3.1

improves MANY video games, also use gaming headphones with an electret microphone


using digital input, standard USB from PC better than analog input

does not do virtual surround in 2-channel speaker mode or via headphones

other channels seem to be balanced but not taking focus away from the center.


Center: simulated single mono channel, can be configured with the Width and Presence controls to vary perception of central soundstage

direct manual gain control, separate center channel level

all use Alps pots

Process mode, doesn't change the main pair channels at all, activates center and surround, extra ambience, without affecting the main channels' stereo imaging

simulated Surround sound processor

no need for high SQ except main L/R front, nor high power SPL except on center