I wake up this morning with that question in my mind, anyone knows?
Prob not the final answer but…
http://www.vinylrecordscollector.co.uk/text/vinylhist14b.html
In 1958 we were introduced to stereo effects on phonographic records. The stereo technology consists of two separate channels of sound recorded in the same groove given the name of stereophonic of which the effect still resident in the vinyl records of today. Just as the use of two eyes creates a perception of depth, so can the effect of musical presence be achieved by stereophonic? The recording sounds with two or more appropriately positioned microphones playing it back into two separate Hi-Fi loudspeakers facilitate this in the recording and provide two separate signal channels.
Well it doesn´t explain how the signal from a single groove is split into two channels for the stereo sound.
I think, but not 100%, that one channel is recorded in the groove’s sidewall, giving lateral movement to the stylus, and the other channel is in the base of the groove giving up-and-down movement.
IIRC, the sum of the Left & Right channel signals (L+R) is etched into the carrier groove as a lateral deviation from centerline. An (L-R) signal is etched into the groove as a width variation which is sensed by the stylus as vertical motion, e.g. as the groove narrows, the pointed stylus rides upwards, and as it widens, the stylus rides downwards.
In the player, simple circuitry can retrieve the L channel by adding the two raw sylus signals and retrieve the R channel by subtracting the two raw signals.
Why use this apparently complicated system? Backwards compatibility. A mono player could play a stereo record and it’d retrieve only the L+R signal, which is exactly what you’d want it to sound like.
The problems of serving the installed user base & backwards compatibility issues didn’t start with the advent of the PC or the VCR. It’s been in Engineering of all sorts for centuries.
I remember that at the time stereo(phonic) music was first becoming popular two local radio stations each broadcast (simulcast?) one channel of the music so that we could experience stereo sound on two seperate radios. It was actually pretty cool. 
Crude, but cool.
These were AM stations, FM hadn’t really come into it’s own yet.
Peace,
mangeorge
The BBC did this as well, in the 1960’s, but they used an FM radio station for one channel and a TV station for the other. These experimental broadcasts were usually carried out on a Saturday morning when the TV station was not being used for regular programmes. This made more sense because back then many people only had one radio set , so using the TV station meant that more people could hear the broadcasts.
Come to think of it, Rayne Man, they did that radio/tv thing here also. Usually for a music show IIRC. It only happened a couple of times though. I think the AM/AM radio broadcasts preceeded that by a little. Most likely due to competition between stations. There was AM stereo for a short time, but FM took over the airwaves very quickly.
I was 15 in 1960, so you can imagine how exciting this was to all us teeny-boppers.
AM/FM simulcast wasn’t done often, but did take place during the 1950s, and 60s. My father was a fan of classical music, and relied on the local Phila station that had both AM and FM transmitters for this occasional treat. My first stereo tuner, a second hand tube-type Harman-Kardon had the AM/FM simulcast selection, a feature that I’d sometimes mess with to see how two different stations would sound at the same time.
Here is someone’s recollection of those BBC experiments.
Stereo radio began in the UK with Saturday morning experiments in 1958 and I remember lugging my parents large valve radio and placing it six feet from the family television to listen to them. One channel was carried via Third Programme transmitters and the other by TV sound, neither network then having normal programmes on a Saturday morning. As we lived on Tyneside and one signal arrived via VHF and the other from the local medium wave repeater, lets just say that the fact that the speakers were not a matched pair did not affect the value of the experiment!
And a longer article here.
http://www.bbctv-ap.co.uk/pawl01ster.htm.
Of course this system was a dead-end and soon after, the BBC adopted the American Zenith-GE pilot tone system which is now used all over the world , using just one transmitter.
Although you’re right about the need to provide mono compatibility, vinyl record grooves are no longer cut using sum and difference techniques. Since the process was standardized in 1958, stereo cutting heads have been of the 45/45 type. As recording engineer Roger Nichols explains:
Playback cartridges use a similar 45/45 degree pickup. The coil arrangment for a moving-magnet cartridge can be seen here. The outputs of the electromagnetic coils (or piezo transducers in the case of a ceramic device) are left and right channel directly - no electronic sum and difference manipulation need be performed.
[Sum and difference channels are use in stereo radio transmission, however. The carrier is Frequency Modulated with the L+R signal up to 15kHz, which all FM radios (even mono ones) can pick up. The L-R signal amplitude modulates a carrier (38kHz IIRC), and this difference channel can be detected and demodulated by FM stereo radios (giving rise to L and R channels by the “simple circuitry” that LSLGuy mentioned. The higher bandwidth and amplitude modulation of the L-R signal means that it’s noisier than the L+R signal, which is why distant FM stations often sound cleaner in mono than in stereo.]
Well, that seems to settle it; odd enough this morning I woke up* thinking that the explanation given by LSLGuy looked a lot like a stereo FM signal.
Seems that my brain shuts up 15 minutes after I woke up lately 