Flatscreen between speakers

Phantom center is a long used term that basically describes imaging between the speakers from standard 2 channel stereo. Derived center is a common term for a synthetic channel that is electronically extracted from the left and right channels of a 2 channel stereo recording to make a center channel.

It's interesting to think that before the age of electronics it would have been pretty much impossible to pull of the phantom center image trick because there were no multiple, near identical, time aligned sound sources available. 

@asctim 

So phantom centre describes imaging where the subject is dead centre!  Are there terms for where all the other sound sources might be located?

In the old days of classical stereo, recordings were made using a very small number of microphones (two or three being common) and the imaging "trick" was pulled off to great effect - no need for "multiple, near identical, time aligned sound sources" to be electronically processed

It’s not necessarily just dead center. Phantom imaging is any image generated some place between the speakers when they both play the same thing in phase. Or if they play the same thing with a time delay on one of the speakers. It’s only dead center if both speakers are at the exact same level and in phase, which means the comb filtering is at it’s worst. (There are potential combinations of level difference and phase/timing difference that can bring the phantom image back to dead center.) A lot of content has a lead singer or instrumentalist panned dead center, so it makes sense in a lot of ways to move that to a center speaker. The devil is in the details of how to do that, so it’s no surprise that audiophiles have often not been satisfied with the results of a center channel. It depends a lot on what characteristics of the sound they are most sensitive to.

If two or more omnidirectional microphones that are not coincident in location are used to record an orchestra, there will be delay characteristics introduced into the recording that may be considered technically less than optimal. I personally think a lot of the Mercury Living Presence recordings sound great despite this issue. Blumlein demonstrated that two coincident microphones could produce a coherent stereo image that was correct in both amplitude and phase at each of the listener’s ears for imaged sound objects placed anywhere between the speakers, with no delay echoes introduced as the microphones are right on top of each other, just pointed in different directions. Unfortunately it only works perfectly when sounds are hard panned to one speaker or the other. Anything in between will sound like it’s coming from the correct direction, but the tone will be increasingly corrupted by interference patterns, with it being at its worst when the imaging is dead center. The interference patterns are not caused by the microphones, but by the 2 speakers when they play the recording back, and this happens with 2 speakers regardless of how the mics were set up.  Fortunately this doesn’t sound as bad as it looks like it should. But still, it doesn’t sound as clear and pure of timbre as does a hard panned sound. So, center channels, and perhaps a couple more speakers between the center and side speakers have desirable potential. The problem is how to make the recording, or how to upmix a good 2 channel recording to more channels without doing more harm than gain. 

I learned recently that there are some French multi-channel recordings that were made with 5 microphones spaced apart in a row in front of the orchestra. The intent is to play the recording back with 5 speakers arranged in your room with the same spacing as the microphones. This allows for a proximation of wave front reproduction that a 2 speaker system cannot reproduce at all. Unfortunately it also generates lots of comb filtering/interference from delays when played back, which could be alleviated to various degrees with more microphones and more speakers. 100 channels would probably do a marvelous job of pushing comb filtering/interference patterns above the critical range, but would also be ridiculous to pull off. 100 speakers and 100 amp channels. Yikes! I’ve thought about ordering some of these 5 channel recordings and having a listen, but honestly I’m enjoying 2 speakers well enough and it’s considerably more convenient. 

@asctim 

When thinking about cancellation and reinforcement of sound, I would encourage you to think about the wavelengths of sound waves in the audible spectrum - Google AI thinks:

The human audible spectrum encompasses frequencies from approximately 20 Hz to 20 kHz. Correspondingly, sound wavelengths within this range in air at standard temperature and pressure vary from about 17 meters (56 feet) at the low end (20 Hz) to 17 millimeters (0.67 inches) at the high end (20 kHz). 

So there is usually lots of distance between speakers and listeners to hold many complete waves for most frequencies, and even between the drivers in one speaker!  There are pictures in this thread where the distance from a main speaker to one driver in the centre channel is about the same as the distance between the outer drivers in the centre channel.  No wonder it sounded better with a smaller centre speaker.

I would also encourage you to explore multi-channel recordings from France and other parts of Europe mainly to the North.  Several thousand classical SACDs are available, for example from Presto.  Most include a CD layer, so you can compare two-channel CD quality with DSD, which is usually on the disk both as two-channel and multi-channel.  SACD has now been around for almost a quarter of a century and these days most cost the same as a CD.  It is hard to find a streaming service that offers multi-channel DSD?

Studios have offered more than two channels since the 1970s, both on tape and quadraphonic records. The CD standard mentions expansion to 4 channels.

Today Dolby Atmos offers up to 32 channels - something being used by some engineers.  Dark Side of the Moon is quite something.  But the best exponent in my opinion is from Norway - Morten Lindberg and his label 2l.no.  I remembered rave reviews in Gramophone for the classic recording Reflections which was released in 2016 in a pack containing a SACD plus a Blu-ray audio disk with many options including Dolby Atmos and 9.1-channel Auro-3D.

Morten Lindberg is willing to leap on anything new - he uses 64-bit formats and floating-point numbers, not the whole numbers we are used to with PCM.  For a lot more depth see Merging Technologies - Use Cases

@richardbrand. Thanks for your thoughtful comment. My personal opinion is that stereophonic was designed to be sufficient for the reproduction of live music. By definition 
“Stereophonic sound, commonly shortened to stereo, is a method of sound reproduction that recreates a multi-directional, 3-dimensional audible perspective”

When done correctly, two speakers can put you right in the venue where the performance is recorded. If you listen to Charles Mingus, Miles Davis and most recordings from that era, you will realize that the weakness in the reproduction chain is the album itself or more precisely, the recording. Take two CDs of the same performance, one from Japan and the other from the USA and you will quickly realize that a lot can be missed from the same performance. If you’re not familiar with BACCH DSP, do a quick search and see how Dr Choueiri was able to restore a lot of that magic back. The problem with multichannel is that it requires a much higher expenditure in money and space. Stereo on the other hand seems simple and clever.