Convert stereo to LCR - feasible or even desirable?

[Boye]

HiFi is an attempt to recreate an acoustic event. The means of creating an auditory illusion of this is traditionally a two-channel system.
The scientific approach in this forum is — for good reason — mainly concerned with reliable and correct reproduction of recordings.
However, stereo might not be the scientifically optimal way to achieve the desired illusion. It goes a long way, but we have learned some things since the introduction of surround sound and Atmos systems.
Do we even know what the best way to record and reproduce sound is? I suppose binaural recording and using headphones might be the simplest way, though it comes with some practical limitations.

This was a long introduction to get to the point.

Working with Jamo in the 90s, I learned about the advantages of using an additional center channel, especially for people sitting slightly off-center.
In a normal stereo setup, a listener sitting off-center will have the stereo balance skewed because the sound from the nearest loudspeaker will be louder and arrive earlier, both of which contribute to the nearest loudspeaker becoming dominant. As a result, even a mono signal will produce a phantom center image that appears off-center.
This problem can be mitigated by using a center channel for the mono part of the recording.

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?

 

[Thighsofsteel]

Meridian Audio did this 20 plus years ago with Trifield. It was excellent. In fact it was 2 channel to 5 channels but is was mainly the front three channels. I still miss it after replacing my Meridian processor with a Lyngdorf due to age more than anything

 

[theREALdotnet]

Any AVR will do it if you configure it as 3.x. The effect and how likeable it is depends on the source material and sound processing programs of the AVR. Sometimes it sounds ok, sometimes bad. I find that it can be beneficial for enhancing spoken narration (TV programs, documentaries, even movies), especially if there are household members with impaired hearing. For music playback I almost always prefer 2-channel playback.

 

[Thighsofsteel]

Here is an article on Trifield

https://www.meridian-audio.info/public/trifield%5B2563%5D.pdf

 

[EERecordist]

There is nothing wrong with trying it, the material, your room, and your subjective hearing will determine the result. It's like food or wine, if it tastes good it is good.

But generally it will not improve things and can make a mess.

In the mixing process the source microphones and processed material are deliberately placed into a 2 or 5.1, 7.1 playback format. The mix engineer is making artistic decisions listening through the release target speaker physical arrangement. You can't undo and redo that at home. You do find demixing usually attempting to separate drums, guitars and vocals.

The purpose of the center channel in movies is to make the dialog more audible. Generally the actors speaking will be in the center of the frame so it makes sense that they sound like they are in the center of the sound field.

If you have Atmos source material, Dolby has re-rendering. It is not my area of expertise and doubt that consumer can actively do it with released material. Maybe there is an ASR expert on that.

 

[90sNintendoKid]

I'm setting up a 3.0 with my old and yet trusty Marantz SR6012 in my Living Room and will be using Dolby Surround for stereo music.

I don't think the Marantz has center spread like the newer Denon/Marantz AVRs.

Gene from Audioholics has done a few YouTube videos on stereo music upmixing and center spread.

 

[Boye]

Here is an article on Trifield

https://www.meridian-audio.info/public/trifield%5B2563%5D.pdf

Thank you. It's a bit more complicated than I hoped for. But at least it does explain some of the theory and considerations in the system. Not somehting you just set up in your DSP

 

[Chris A]

...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...

A free JAES white paper--from 2002 (enclosed below).

Dave Griesinger (the real person that technically created Lexicon) has also written a few papers and PowerPoint presentations on the subject, albeit with particular emphasis on the surround channels (i.e., 5.1) to give depth/immersion (i.e., Logic7, et al.). A patent from 1998:

MULTICHANNEL ACTIVE MATRIX SOUND REPRODUCTION WITH MAXIMUM LATERAL SEPARATION


Chris

 

[olds1959special]

I have a Yamaha NS-C225 center speaker which I use with an Aiyama A70 Mono amp (LM4562 op-amp) through a stereo-mono converter box. I just use my computers mini jack audio output and then create a multi-output device with drift correction to get the center speaker to play along with the other speakers. Alternatively you can use splitters from your main output to do this. Then I adjust the volume to taste, and I think the SoundSource app can control the volumes of the different devices together - but for now I am doing it manually.

 

[Thighsofsteel]

Thank you. It's a bit more complicated than I hoped for. But at least it does explain some of the theory and considerations in the system. Not somehting you just set up in your DSP

No it isn’t. I still have my Meridian 861 in my cupboard. I am thinking of re- installing it for a couple of weeks to see if Trifield is just Rose tinted spectacles or as good as I remember

 

[MaxwellsEq]

Thank you. It's a bit more complicated than I hoped for. But at least it does explain some of the theory and considerations in the system. Not somehting you just set up in your DSP

This blunt issue is that most commercial music (e.g. in the charts) has only two channels of information encoded in release-medium (cassette, LP, CD). There is no 3rd channel, and storing the data on an LP would have required a matrix or carrier model.

So there's no 3rd channel to extract. And there are no rules for how the data can be reconstructed. Given these constraints, any 3rd channel you create will effectively be wrong. And it will be differently wrong for every different recording (and potentially every different remaster). That doesn't mean to say you won't enjoy it, and perhaps prefer it to the phantom centre which IS encoded in the media.

 

[Boye]

A free JAES white paper--from 2002 (enclosed below).

Dave Griesinger (the real person that technically created Lexicon) has also written a few papers and PowerPoint presentations on the subject, albeit with particular emphasis on the surround channels (i.e., 5.1) to give depth/immersion (i.e., Logic7, et al.). A patent from 1998:

MULTICHANNEL ACTIVE MATRIX SOUND REPRODUCTION WITH MAXIMUM LATERAL SEPARATION


Chris

Now that is really complicated. But I'll try and see if I can make sense of it, even though some of the math is far beyond my grasp.

 

[kyuu]

Given these constraints, any 3rd channel you create will effectively be wrong. And it will be differently wrong for every different recording (and potentially every different remaster). That doesn't mean to say you won't enjoy it, and perhaps prefer it to the phantom centre which IS encoded in the media.

Sure, but some might say that the entire premise of recording and playing back in two channels is wrong (or at least deficient) to begin with.

 

[DVDdoug]

The old Dolby Pro Logic did a pretty-good job of "steering" centered dialog to the center. But it steers everything toward the center when left and right are identical (or mostly identical) so it tends to mess-up music and it's unpredictable with movies that aren't Pro Logic encoded.

There are Pro Logic "music" and "matrix" modes that don't dynamically steer the sound and they tend to work better with music but you may or may not like the results and different music will be "decoded" differently.

I don't think modern AVRs have Pro Logic but they usually have some up-mixing options.

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.

Way back in the original quadraphonic days I couldn't afford quad (and quad sources were limited) but I built a L-R circuit (easy with an op-amp) and ran the L-R to an amplifier for the rear channels. There are phase issues because "L-R" means that the "L" is in-phase with the front and "-R" is out-of-phase with the front; But you can "randomize" that by flipping polarity of one rear speaker. That setup didn't have a center speaker. Original quad didn't have a center.

I feel like L+R to a center speaker mostly just deteriorates the stereo image because full-left and full-right panned signals now go to the center.

 

[kyuu]

I feel like L+R to a center speaker mostly just deteriorates the stereo image because full-left and full-right panned signals now go to the center.

Huh? Wouldn't anything panned fully to the left or right explicitly not go to the center?

In any case, the stereo image only applies in a single listening position. Anyone who isn't listening alone or who doesn't like to be tied to a specific spot, as the OP alluded to with people sitting "slightly off-center", seem like they could only benefit.

 

[Chris A]

Now that is really complicated. But I'll try and see if I can make sense of it, even though some of the math is far beyond my grasp.

Paul Klipsch papers are much older...but much more tractable:

 

[MaxwellsEq]

Sure, but some might say that the entire premise of recording and playing back in two channels is wrong (or at least deficient) to begin with.

Yes, I wasn't defending stereo. 3 channel, if introduced at the start, would probably have been be superior.

But it wasn't, and so we only have 2 channels, and no built-in encoding for centre data.

 

[terryforsythe]

Never mind.

 

[Matias]

Syn was released and quickly discontinued. Apparently not a big seller.

New Schiit SYN - Analog Surround Sound Processor

From https://www.schiit.com/products/syn Imagine surround sound that’s easy, fun, and works with everything. That’s Syn. Well, Syn is also a DAC, remote control preamp, and headphone amp as well. Oh yeah, and it won’t become a paperweight when standards change. Head exploded? That’s OK. We’ll...

www.audiosciencereview.com

 

[Chris A]

Perhaps you might think about using one of the better SINAD preamp/processors (i.e., AVP), but buy used to get the price way down. Here is one that still has good performance by Amir's standards (I currently use one for my setup):

https://www.hifishark.com/model/emotiva-xmc-1

Its surround processor choices seem to sound better than most off-the-shelf AVRs. I also think the bass management is nice to have if you think about adding a subwoofer.

You'll need three amplifier channels for L/C/R, but I find it's nice to not have to use built-in AVR amplifiers--which I find often are a little flat sounding.

Chris