Why do we need a center channel in a home theater system?

[JSmith]

Some light reading on problems with centre speakers;

PROBLEMS WITH HORIZONTALLY ORIENTED CENTER CHANNEL SPEAKERS

MTMs were designed to be vertically oriented because they have very good horizontal dispersion and less than ideal vertical dispersion.

When a center channel speaker that uses MTM orientation is laid on its side, you will get good vertical dispersion and a very poor and narrow horizontal dispersion. Although both should be wide, horizontal dispersion perhaps is more significant than vertical dispersion. Off-axis horizontal response is the sound that the listener hears while moving around the speaker.

https://forum.blu-ray.com/showthread.php?t=89614

https://www.audioholics.com/loudspeaker-design/center-channel-designs-1

http://techtalk.parts-express.com/f...spersion-of-center-channel-mtm-how-bad-is-bad



JSmith

 

[vavan]

Wow, live and learn! Just as I thought I've discovered and understood all the important settings of my AVR

another interesting option might be "dialog lift" (or whatever it's called) for those having their center channel low on the floor. I like what it does on my Pio with FH available

 

[JoachimStrobel]

Imagine the situation of listening to a recording where there is a trio playing - one in the center, one slightly to the left and one slightly to the right. With your "logic" we really should have two more speakers at the exact position of where the two side instruments should be because the waves coming from such a position if live would hit the ears differently (angle) than the proportionate mix of L and R to create that stereo image. Even if you had a center speaker where such an image would be created by a proportionate mix to L and C or R and C, it wouldn't be the same with the same logic because it would be still creating the "illusion" of a position in those two locations with two "artificial" sound sources.

The logical extension (to show the absurdity) of this is to have a zillion speaker array in front for every possible location of where the instrument might be on stage so that the wave coming from that instrument live would be exactly be the same as what is coming from a speaker corresponding to that position.

A moment of sanity would suggest that not only is this impractical but wholly unnecessary and that human hearing is perfectly capable of appreciating the content equally with far more imperfection than this theoretical artifact. Our entire visual depth perception is based on this "illusion" as well as making movies possible with small and finite frame rates.

Perhaps this is the difference between people who have an audio system primarily to appreciate the content that is being played vs people who want the system to "perfectly reproduce" something without which they are convinced they can not appreciate the content as well. Both have a point but only within limits, far from academic extremes.

Without trying to source and hence possibly citing it wrong, I believe those experiments have been made 90 years ago: How many front speakers would one need? And it turns out that three is a good number. Sure, more would be better, but one centre to complement the side ones is Ok. The HRTF is not that sensitive to need half rights and half lefts. But should we once turn to the subject of the rear speakers - there one would need zillions to re-image the rear sound field. But as we stay with the front sound field in this thread, the solution is simple: 3.

 

[JoachimStrobel]

As Sancus pointed out above, there is a "Center spread" option that yields a tremendous improvement, although it is still difficult to achieve soundstage that's as good as just the two main speakers provide.
Based on my own experience so far - mind that my center speaker is not perfectly matching and is not located at the perfect ear-level height - I would argue that for one seat only, and even for two closely located seats (e. g. you and your significant other on a couch), with a less than perfect center channel speaker, one is better off without a dedicated center speaker. But I'm now curious about a perfect center channel. I also barely listened to any DVD-Audio mch music yet, which I'm going to listen to as time allows.

I am not a good cook, but when I cook I only use the best ingredients just to be sure that the fault then was elsewhere (me). And with the centre, I use identical ones, all upright with the same amp and I Dirac the whole setup. It still is not perfect, and that is then the fault of the recording or my hearing, but nothing else (OK, could be Dirac filter issues trying to harmonize 5 speakers to a common room curve, but that is another issue). And based on that, I can analyse those recordings or my hearing.

 

[Thomas_A]

To a first approximation (neglecting reflections and other room modes), the distance of the perceived point away from any speaker is the weighted ratio of the corresponding volumes for the same sounds from each speaker reaching the ear. This is how balance (and individual channel balance settings in contemporary equipment) works for imaging. But it also depends on the frequency of the sound.

Lower frequencies are much less localized to the ear so other cues "trick" the mind into thinking it is clumped with the related sounds. This is how subs are incorporated into the system. Play the sub on its own, and you can fairly reliably locate the sub which can be totally off-center. Play it as part of the content and it will look like the bass parts are coming from wherever the related higher frequency sounds are coming from. So psycho-acoustics are also involved here even with only audio cues.

If you play a pure tone and move around, you will much more easily notice the change in center depending on your relative position to the speaker. Play it with mixed content and your mind plays a part in that perception, including an expectation that the dialog is in the center even without visual cues.

And if you have actual stereo content in which the far speaker may be carrying different content with stronger audio cues (frequencies for localization), your perception may still be swayed in that direction even if you are sitting closer to the other speaker.

And set-up? My speakers are almost 45 degree toe-in. While sitting in front of the left speaker the right one is on-axis.

 

[Vasr]

Without trying to source and hence possibly citing it wrong, I believe those experiments have been made 90 years ago: How many front speakers would one need? And it turns out that three is a good number. Sure, more would be better, but one centre to complement the side ones is Ok. The HRTF is not that sensitive to need half rights and half lefts. But should we once turn to the subject of the rear speakers - there one would need zillions to re-image the rear sound field. But as we stay with the front sound field in this thread, the solution is simple: 3.

You are mixing up two different things here.

If you are referring to the Bell Labs study in the 1930s that suggested a center channel, the suggestion had to do with localizing the center image and syncing with the visual cues on screen not audio on its own (the original thing that was being discussed here before it got thread-jacked into multi-channel music). My very first post in this thread justified a reason for needing the center channel to be able to better sync with the visual cues of a HT set up. That rationale is still valid and more real than any HRTF reasons. So, bringing in HRTF esoterica in that context is irrelevant and moot.

In that Bell Labs proposal, the center channel was proposed to be required ONLY for people that were off-center (since most of the people in a a theater would be off center) and not required if you happened to be in the center of L and R so that they didn't lose the sync with the screen. It has nothing to do with HRTF. In fact, asserting that the person in the center doesn't need a center channel speaker discounted the effect of HRTF (not that HRTF was being studied then). So, you cannot use that study's rationale for having a center channel in an audio only listening in the center MLP and pull out a "3 is enough" because of that study. That particular study is not relevant to your context.

There is no study I am aware of that talks about the requirement of a center speaker for HRTF reasons.

The HRTF phenomena is real. The difference in localization perception depending on source(s) are well documented. The theory is sound. But its practical applications are more to do with how to simulate a specific location without a speaker in that location via the speakers you have, not to suggest there has to be one in the center. Some of the synthesized surround processors do this assuming just two speakers. A typical center speaker is not a point source in the center anyway to be nitpicky precise.

I would even venture to say that if anyone did a blind test of the same content for audio only (without visual cues) via a L-R, a L-C_R and a L-Plantom C-R and set just the level balance and mix to have the resulting image sweep similarly between the speakers at the MLP, most people wouldn't be able to differentiate between them spatially or tonally. Even if one could measure the difference.

 

[Vasr]

And set-up? My speakers are almost 45 degree toe-in. While sitting in front of the left speaker the right one is on-axis.

Yes, that would help with off-center viewing and one of the advantages of a phantom center like the wall mounted MMG-Ws. They are on a swivel and you can adjust the angle to optimize the center localization as widely as possible for a given room.

The problem in doing this with a toe-in with just L and R is that a large toe-in narrows the sound stage considerably for stereo content from the center listening position. So a separate phantom C with two speakers right next to the screen would help in syncing the screen stage to the audio stage as well as compensate for off-center viewing if the toe in can be adjusted without affecting the 2.0 sound stage or introducing gaps between L and R and surround speakers with a narrow soundstage in 5.x+ configurations. A more flexible solution in this compromised world of trade-offs.

 

[Vasr]

Can you think of a particular scene(s) that would be good for testing for this problem? Would make for a great tool to tune or check your audio setup!

The discussion above triggered recollection of this video. Pretty crude but OK if you want to see the relationship between your video screen stage and audio screen stage at the edges.

 

[Thomas_A]

Yes, that would help with off-center viewing and one of the advantages of a phantom center like the wall mounted MMG-Ws. They are on a swivel and you can adjust the angle to optimize the center localization as widely as possible for a given room.

The problem in doing this with a toe-in with just L and R is that a large toe-in narrows the sound stage considerably for stereo content from the center listening position. So a separate phantom C with two speakers right next to the screen would help in syncing the screen stage to the audio stage as well as compensate for off-center viewing if the toe in can be adjusted without affecting the 2.0 sound stage or introducing gaps between L and R and surround speakers with a narrow soundstage in 5.x+ configurations. A more flexible solution in this compromised world of trade-offs.

What is the reason for narrowing the soundstage? Isn’t this dependent of the dispersion characteristics of the speaker and wall reflections?

 

[abdo123]

It’s funny how you think a center image in stereo is ‘fine’

You basically have to have your setup in an underground bunker (no windows, no walls with different densities, furniture symmetry) and have both your ears’ hearing equal.

I have never experienced proper center (in stereo) in my life as my left ear is more dominant. For my partner, outside of headphones and near field speakers, he never heard it in our apartment either.

 

[Alexium]

It’s funny how you think a center image in stereo is ‘fine’

You basically have to have your setup in an underground bunker (no windows, no walls with different densities, furniture symmetry) and have both your ears’ hearing equal.

I have never experienced proper center (in stereo) in my life as my left ear is more dominant. For my partner, outside of headphones and near field speakers, he never heard it in our apartment either.

I understand that physiological peculiarities, if any, are a huge factor, but I am in a small untreated room, with the audio system located asymmetrically (50 cm from one side wall, 2 meters from the other), and imaging in stereo configuration (only L and R playing) is quite good. I'm probably missing a lot of subtle details due to both cheap speakers and poor soeaker location in an untreated room, there is no depth to the scene, but left to right positioning, including centering, is very good.

In fact, I have never heard a really good soundstage in headphones, despite the fact that my headphone gear is much better than speaker gear (I even had Sennheiser HD800 for a while). Centering is solid, yes, by everything else is moot. I think it's because most recordings are mixed for loudspeakers, only binaural recordings are designed to give you proper imaging with headphones.

 

[abdo123]

I understand that physiological peculiarities, if any, are a huge factor, but I am in a small untreated room, with the audio system located asymmetrically (50 cm from one side wall, 2 meters from the other), and imaging in stereo configuration (only L and R playing) is quite good. I'm probably missing a lot of subtle details due to both cheap speakers and poor soeaker location in an untreated room, there is no depth to the scene, but left to right positioning, including centering, is very good.

In fact, I have never heard a really good soundstage in headphones, despite the fact that my headphone gear is much better than speaker gear (I even had Sennheiser HD800 for a while). Centering is solid, yes, by everything else is moot. I think it's because most recordings are mixed for loudspeakers, only binaural recordings are designed to give you proper imaging with headphones.

and how far do you sit from these two speakers? are you sitting close to the wall infront of these two speakers? is your room regularly shaped?

 

[Vasr]

What is the reason for narrowing the soundstage? Isn’t this dependent of the dispersion characteristics of the speaker and wall reflections?

Here is one explanation of this:

http://www.acousticfrontiers.com/speaker-off-axis-understanding-the-effect-of-speaker-toe-in/

 

[Robin L]

I had a 5.1 system in the garage, a dedicated Infinity Primus set, with two Towers and a center speaker up front, two mini towers [the 250s] in the rear. Those were used from Amvets, set me back all of $80.

The Primus C 25:

Primus 360:

Primus 250:

Also had three powered subs, also from Amvets. Didn't have it for video-my video was a 13" Sony Trinitron, classic old-school for $1, also Amvets. Enough for me to scroll through menus, I wanted a surround system. Center images from all recordings tightened up, but the 5.1 recordings imaged best. 2-channel Stereo is a set of tricks, fragile ones, which is why most stereo imaging fails apart outside of the "sweet spot".

 

[Thomas_A]

Here is one explanation of this:

http://www.acousticfrontiers.com/speaker-off-axis-understanding-the-effect-of-speaker-toe-in/

The problem with reflections from closest wall (including front wall) in small rooms is that they may blur details and destroy the soundstage. The delay needs to be long enough and/or level needs to be low enough to be beneficial for the "envelope" effect. So I am not convinced that the soundstage is enhanced. It also depends on the directivity, wide or narrow dispersion speakers. You can have wide dispersion in the HF region which makes the reflected vs direct sound still quite similar over a wide frequency range, even with heavy toe-in (but at a lower level). My speakers start to have significant deviations around 8-9 kHz at large angles, so the reflected sound in the important 1-6 kHz range are quite similar to the direct sound.

 

[Vasr]

Sure, there are a lot of variables in room and speaker design as well as distance between speakers and to the MLP that would affect the impact of a toe-in but in general, higher the toe-in, narrower the sound stage perception unless you have some kind of omni-directional speakers. I think these are fairly well documented including, for example, the pros and cons of the Tannoy method, so I don't think this is controversial. You aren't trying to enhance the sound stage as much as minimize the narrowing with too much toe-in.

In an individual specific case, you can do your own experiments to see what works best. Since, it is a trade-off between a number of things, you land up optimizing for the best compromise depending on requirement. Vertical drivers like the Magnepans show even more pronounced effects depending on whether tweeters are placed outside or inside.

Using a visual cue in front like the video above playing full screen between the speakers will likely make it easier to judge any shift in sound stage with toe-in since it is rather difficult to measure the sound stage perception.

 

[Alexium]

Here is one explanation of this:
http://www.acousticfrontiers.com/speaker-off-axis-understanding-the-effect-of-speaker-toe-in/

Thank you, this is a brief, but quite a good explanation. There's also an interesting discussion on extreme toe-in in the comments there.

 

[Alexium]

and how far do you sit from these two speakers? are you sitting close to the wall infront of these two speakers? is your room regularly shaped?

About 2.6 meters between me an each of the speakers. The room is rectangular, with a normal ceiling (parallel to the floor), and yes, the speakers are close to the front wall (40 cm between the wall and the back of the speaker). This is because my small home theater zone is set up across the shorter side of the room rather than the longer side (and this is also why it's hugging one side wall while being quite away from the other).

Don't get me wrong, I'm not claiming I have awesome soundstage - in fact, I'm starting to think I have never in my life heard the proper stage/imaging. But when I fire up a live jazz recording I usually can easily locate different instruments at different positions off-center. And if something is dead center, I can identify it as such as well. Not with stunning pin-point laser clarity, but the general idea gets across.

 

[JoachimStrobel]

You are mixing up two different things here.

If you are referring to the Bell Labs study in the 1930s that suggested a center channel, the suggestion had to do with localizing the center image and syncing with the visual cues on screen not audio on its own (the original thing that was being discussed here before it got thread-jacked into multi-channel music). My very first post in this thread justified a reason for needing the center channel to be able to better sync with the visual cues of a HT set up. That rationale is still valid and more real than any HRTF reasons. So, bringing in HRTF esoterica in that context is irrelevant and moot.

In that Bell Labs proposal, the center channel was proposed to be required ONLY for people that were off-center (since most of the people in a a theater would be off center) and not required if you happened to be in the center of L and R so that they didn't lose the sync with the screen. It has nothing to do with HRTF. In fact, asserting that the person in the center doesn't need a center channel speaker discounted the effect of HRTF (not that HRTF was being studied then). So, you cannot use that study's rationale for having a center channel in an audio only listening in the center MLP and pull out a "3 is enough" because of that study. That particular study is not relevant to your context.

There is no study I am aware of that talks about the requirement of a center speaker for HRTF reasons.

The HRTF phenomena is real. The difference in localization perception depending on source(s) are well documented. The theory is sound. But its practical applications are more to do with how to simulate a specific location without a speaker in that location via the speakers you have, not to suggest there has to be one in the center. Some of the synthesized surround processors do this assuming just two speakers. A typical center speaker is not a point source in the center anyway to be nitpicky precise.

I would even venture to say that if anyone did a blind test of the same content for audio only (without visual cues) via a L-R, a L-C_R and a L-Plantom C-R and set just the level balance and mix to have the resulting image sweep similarly between the speakers at the MLP, most people wouldn't be able to differentiate between them spatially or tonally. Even if one could measure the difference.

As I said, I have not googled the work I cited but will do so eventually, it tried to ask the question how many speaker are needed for a crude soundfield reproduction.
Since watching Starwars I (now IV) I have no doubt that decoupling voices from (battlefield)sound is a must. Hence, I am not sure why there is such a huge discussion about it. This has nothing to do with HRTF, which is a nice extra benefit a centre speaker has. Using a phantom instead of a true centre might work, and 30 years ago, as Mch was new often a necessity as the focus was on Stereo only then. If video is not so important then a phantom is still fine. As are many choices in audio. But I still try to understand why this is such a strong issue. And that human hearing is able to localize sound by its timbre is rather trivial and I would call this esoteric and moot ( had to google that word...)

 

[JoachimStrobel]

You are mixing up two different things here.

If you are referring to the Bell Labs study in the 1930s that suggested a center channel, the suggestion had to do with localizing the center image and syncing with the visual cues on screen not audio on its own (the original thing that was being discussed here before it got thread-jacked into multi-channel music). My very first post in this thread justified a reason for needing the center channel to be able to better sync with the visual cues of a HT set up. That rationale is still valid and more real than any HRTF reasons. So, bringing in HRTF esoterica in that context is irrelevant and moot.

In that Bell Labs proposal, the center channel was proposed to be required ONLY for people that were off-center (since most of the people in a a theater would be off center) and not required if you happened to be in the center of L and R so that they didn't lose the sync with the screen. It has nothing to do with HRTF. In fact, asserting that the person in the center doesn't need a center channel speaker discounted the effect of HRTF (not that HRTF was being studied then). So, you cannot use that study's rationale for having a center channel in an audio only listening in the center MLP and pull out a "3 is enough" because of that study. That particular study is not relevant to your context.

There is no study I am aware of that talks about the requirement of a center speaker for HRTF reasons.

The HRTF phenomena is real. The difference in localization perception depending on source(s) are well documented. The theory is sound. But its practical applications are more to do with how to simulate a specific location without a speaker in that location via the speakers you have, not to suggest there has to be one in the center. Some of the synthesized surround processors do this assuming just two speakers. A typical center speaker is not a point source in the center anyway to be nitpicky precise.

I would even venture to say that if anyone did a blind test of the same content for audio only (without visual cues) via a L-R, a L-C_R and a L-Plantom C-R and set just the level balance and mix to have the resulting image sweep similarly between the speakers at the MLP, most people wouldn't be able to differentiate between them spatially or tonally. Even if one could measure the difference.

Steinberg, J. C., and Snow, W. B. (1934). Auditory perspective-physical fac- tors. Transactions of the American Institute of Electrical Engineers 53(1), 12-17. (Behind paywalls), cited, amongst others in https://www.harman.com/documents/HowManyChannels_0.pdf