Are three channels better than two for stereo reproduction?

[terryforsythe]

How would I extract the exclusively mono content and only have that play from the center though?

I'm not sure about that. It may take some research.

Nonetheless, for the center channel, both the EP Optimal and Gerzon algorithms merely sum the left and right channels, and then reduce the gain, e.g.,

EP Optimal: Cs = 0.5270(L+R)
Gerson: Cs = 0.707(L+R)

The miniDSP products discussed and CamillaDSP will do that.

 

[NTK]

How to make the center channel only play mono content, as in content that is present in both channels minus difference content.

Not simple. See Michael Gerzon's paper.

http://decoy.iki.fi/dsound/ambisonic/motherlode/source/7038.pdf

 

[Kal Rubinson]

Is 3 channel mixing a thing?

Rarely but it exists. The most obvious instances are the Mercury Living Presence and RCA Living Stereo recordings that were made with a 3 microphone array. Originally, mixed down to 2 channels for LP and CD releases, many have been made available in valid 3 channel releases on SACD.

(Note that this has been stated earlier, yet the question is still asked.)

 

[terryforsythe]

With regard to imaging, psychoacoustics plays a significant role.

This morning, the boss (a.k.a., my wife) is working in the office, so I have the house to myself. I pulled my speakers out into my family room to their optimal listening positions, about 1m from the back wall and toed out with respect to my listening position. I then started listening to music.

Initially my speakers were not imaging as well as they do normally. Immediately I thought this thread and all of the discussion about imaging must have gotten into my head.

To check the imaging, I put on "Chocolate Chip Trip" by Tool, closed my eyes, and focused on the drums panning around the sound stage. The imaging came back. This must have conditioned by brain. I replayed music I had played earlier and the rock solid phantom images were back.

 

[steve59]

Stereo is stereo and 3 channel is a home theater option, like asking if a car is a better bike than a bike, similar purpose different method of delivery.

 

[terryforsythe]

Stereo is stereo and 3 channel is a home theater option, like asking if a car is a better bike than a bike, similar purpose different method of delivery.

There is a fair amount of scientific literature on the subject, some of which has been discussed herein, that present a different perspective, backed up by scientific studies.

 

[Robin L]

Stereo is stereo and 3 channel is a home theater option, like asking if a car is a better bike than a bike, similar purpose different method of delivery.

Stereo means "solid" and an array of multiple speakers fed separate sonic perspectives results in more solid sonic imagery than just two channels. This applies to both music and video. The question is then how complex and expensive an audio system the market can bear. For video, 5.1 is pretty common. For music, not so much.

 

[terryforsythe]

The part I am not able to figure out though is how to make the center channel only play mono content, as in content that is present in both channels minus difference content.

Giving this more thought, in CamillaDSP you can create multiple mixers, and then put those in series in the Pipeline. You could experiment with that.

For example, in the first mixer you can sum the left and right channels.

In the second mixer you again can sum the left and right channels, but this time invert either the left channel or the right channel.

Inverting the left channel in the second mixer would be equivalent to (L+R) - (L-R). Inverting the right channel in the second mixer would be equivalent to (L+R) - (R-L).

 

[LowShelfFilter]

Giving this more thought, in CamillaDSP you can create multiple mixers, and then put those in series in the Pipeline. You could experiment with that.

For example, in the first mixer you can sum the left and right channels.

In the second mixer you again can sum the left and right channels, but this time invert either the left channel or the right channel.

Inverting the left channel in the second mixer would be equivalent to (L+R) - (L-R). Inverting the right channel in the second mixer would be equivalent to (L+R) - (R-L).

Would it matter which channel I inverted in the second mixer? I saw someone talking about this same methodology in a reddit post.

 

[terryforsythe]

Would it matter which channel I inverted in the second mixer? I saw someone talking about this same methodology in a reddit post.

I'm not sure. Also, the gain of the channels in the first mixer and second mixer would need to be different.

If not, (L+R) - (L-R) = 2R and (L+R) - (R-L) = 2L. This would defeat the purpose.

I think I would just leave the center channel as L+R, with the gain selected per one of the previously discussed algorithms.

 

[goat76]

The part I am not able to figure out though is how to make the center channel only play mono content, as in content that is present in both channels minus difference content.

The thing is that even if you figure out a way to extract the dead-centered phantom content from a 2-channel stereo file and play that through a center speaker, it's not necessarily something you would prefer the sound of.

A few years ago, I watched a three-hour-long video about mixing for Dolby Atmos, and Andrew Scheps and his fellow mixing engineers all came to the same conclusion that they found most sound objects when just panned to the center speaker to not sound right. They thought most sound objects sounded way too isolated coming from just the center speaker, so they often chose to let those sound objects be played as phantom-centered sounds by the main left and right front speakers instead, just as they are in a regular 2-channel stereo mix.

I have also examined how things are usually done with Atmos mixes by listening closely to the individual speakers, and it's way less common than one would think that they mix sound objects exclusively to the center channel, other than stuff like a high-hat or the dry track of a kick drum. The voice of the singer is usually either about equally as much panned to all three front speakers, or played as a phantom-centered sound by just the front left and the front right speaker.

 

[LowShelfFilter]

I'm not sure. Also, the gain of the channels in the first mixer and second mixer would need to be different.

If not, (L+R) - (L-R) = 2R and (L+R) - (R-L) = 2L. This would defeat the purpose.

When looking at it algebraically, it doesn't seem possible to only extract mono information, in which case, gain adjustments shown earlier in this thread would be ideal.

The thing is that even if you figure out a way to extract the dead-centered phantom content from a 2-channel stereo file and play that through a center speaker, it's not necessarily something you would prefer the sound of.

A few years ago, I watched a three-hour-long video about mixing for Dolby Atmos, and Andrew Scheps and his fellow mixing engineers all came to the same conclusion that they found most sound objects when just panned to the center speaker to not sound right. They thought most sound objects sounded way too isolated coming from just the center speaker, so they often chose to let those sound objects be played as phantom-centered sounds by the main left and right front speakers instead.

I have also examined how things are usually done with Atmos mixes by listening closely to the individual speakers, and it's way less common than one would think that they mix sound objects exclusively to the center channel, other than stuff like a high-hat or the dry track of a kick drum. The voice of the singer is usually either about equally as much panned to all three front speakers, or played as a phantom-centered sound by just the front left and the front right speaker.

I think this has more to do with the directional properties of the loudspeaker used, than anything. With an omnidirectional center channel, for example, this would certainly not be an issue.

Which brings up another point: the directional properties of a loudspeaker are not necessarily representative of a real singer or instruments, which further confuses the question of what you are supposed to do for the most realistic sound reproduction.

 

[terryforsythe]

what you are supposed to do for the most realistic sound reproduction.

With current technology, there is no "most realistic sound reproduction." There are different strategies for obtaining good sound reproduction, but none are truly realistic, and which strategies are considered best may vary from person to person, e.g.:

open baffle vs closed baffle
concentric drivers vs separate drivers vs full range drivers
cone drivers vs electrostatic vs ribbon vs AMT
paper cones vs aluminum cones vs composite cones
horn tweeters vs waveguided tweeters vs flush tweeters
narrow directivity vs wide directivity
tower speakers vs standmount/bookshelf speakers with subwoofer(s)
etc.

All of the above have their advantages and disadvantages. I do not consider any strategy to be best. I have found what strategies currently work well for me in my listening environments. But, I remain open to other strategies.

 

[LowShelfFilter]

After getting the insight I desired from this thread, I decided not to add a center channel for now. My reasons are as follows:
1. Proper implementation is significantly more complex/costly than I had imagined
2. The only place I could put the center channel is above my TV, which may damage imaging coherency (but could also increase soundstage height)
3. I do not need to account for multiple listening positions as this is mainly a personal listening space
4. The center channel would still play information that is exclusive to the left and right channels
5. I do not have much space to widen my left and right speakers to compensate for the narrower center image caused by using a center channel
6. The loudspeaker does not know the dispersion characteristics of the recorded source (singer, instrument, etc.), so it is questionable what improvements even will be brought

Thanks to those who helped.

 

[dlaloum]

Stereo is stereo and 3 channel is a home theater option, like asking if a car is a better bike than a bike, similar purpose different method of delivery.

Magnepan’s “Tri-Center” Concept: Does Stereo Sound Better with Three Channels? - The Absolute Sound

Magnepan’s “Tri-Center” Concept: Does Stereo Sound Better with Three Channels? -

www.theabsolutesound.com

 

[theREALdotnet]

Magnepan’s “Tri-Center” Concept: Does Stereo Sound Better with Three Channels? - The Absolute Sound

Magnepan’s “Tri-Center” Concept: Does Stereo Sound Better with Three Channels? -

www.theabsolutesound.com

Betteridge's Law?

 

[Holmz]

Rarely but it exists. The most obvious instances are the Mercury Living Presence and RCA Living Stereo recordings that were made with a 3 microphone array. Originally, mixed down to 2 channels for LP and CD releases, many have been made available in valid 3 channel releases on SACD.

(Note that this has been stated earlier, yet the question is still asked.)

Thanks @Kal Rubinson.


I can only envision taking one’s favourtite motorcycle, and making it into a Robin Reliant.

 

[LowShelfFilter]

As an update to this thread, I repositioned my speakers such that they are spread apart more (actually equilateral monitoring triangle now). The reason for this is that I was comparing my system to a JBL IMAX theater when watching Deadpool and Wolverine and realized I needed more soundstage width. The main issue now is that the sweet spot is +/- 2 inches and unusable for multi-person listening. The root cause of this also results in timbre of the phantom images being degraded further. I am fully convinced now that a center channel is needed for the highest quality sound reproduction.

 

[Miguelón]

Considering that two speakers cannot accurately reproduce a 3D acoustic event in the first place, would using a center channel summed to mono with LR playing normally provide for a superior/more realistic listening experience than just using LR? Please note that I am not referring to listening to three channel recordings, but rather up-mixing two channel recordings to three channels.

I am considering routing audio as such with a MiniDSP 2x4 product:
Output 1: Right Speaker (Right input channel)
Output 2: Left Speaker (Left input channel)
Output 3: Center Speaker (Right + Left input channels)
Output 4: Subwoofer(s) (Right + Left input channels)

From Floyd Toole's book, there are significant acoustic advantages to using a center channel. Are there any downsides to using LCR for stereo reproduction instead of LR?
View attachment 382242 View attachment 382243

I don’t know if Floyd tested his own scheme, but I listened once at a friend’s home and sounded awful, was an experiment.

Stereo is recorded for two channels exclusively, time delay between left and right is clue to good representation of object position in the brain, and the center speaker destroys this part.

Even inter ears loudness will be altered by the center speaker, the thing we listened for was somehow a massive sound bar: totally center focused with some little side to side sensations.

I called a “monostereo” setup

 

[goat76]

As an update to this thread, I repositioned my speakers such that they are spread apart more (actually equilateral monitoring triangle now). The reason for this is that I was comparing my system to a JBL IMAX theater when watching Deadpool and Wolverine and realized I needed more soundstage width. The main issue now is that the sweet spot is +/- 2 inches and unusable for multi-person listening. The root cause of this also results in timbre of the phantom images being degraded further. I am fully convinced now that a center channel is needed for the highest quality sound reproduction.

How far apart do you have your speakers now?