What's the difference between Stereo, Joint Stereo, and Dual

[Pianoman]

Anyone knows ?

[KoalaBear]

A stereo recording is one which is designed to reproduce the phenomenon of acoustic parallax, in which your brain can determine the direction a sound is coming from based on the difference between what's heard by either ear. The two channels are only slightly different from one other, but that difference is enough to give the sound the illusion of depth.

In the context of MPEG, a "stereo" recording is one in which each channel holds a complete copy of the soundtrack, one for each ear. This is expensive in terms of bitrate, so "joint stereo" was developed in which the two channels are combined for the purposes of storage, but the differences appropriate to the left and right ear are reproduced when the recording is played back.

"Dual channel" means exactly that -- the left and right channels can hold the same, similar or different soundtracks depending on what's needed. You can have two copies of a mono soundtrack, a real or perhaps simulated stereo soundtrack, or one mono soundtrack in two different languages.

[JHebert]

Here is an explanation of Stereo vs Joint Stereo copied from the guides at AnandTech:

Another basic option determines how stereo music is encoded. Stereo is just as it sounds - each channel is allocated exactly half the available bit stream, whether it needs it or not (this applies to both CBR and VBR). Joint Stereo (aka MS Stereo or M-S Stereo) is more complicated. Joint Stereo takes advantage of the fact that oftentimes there is only a small or moderate difference between the left and right channels. By encoding the common signal plus the difference signal (and letting the player reassemble them), a significant reduction in required bit rate for a given quality can be achieved.

[vitualis]

I may be mistaken, but my understanding is as follows:

In the context of MPEG-1 audio,

"Stereo" or or "MS Stereo" is sort of what JHerbert just described. The MPEG audio encoder can adjust how much of the total bitrate is given to the left and right channels depending on... "need" I suppose, and the relative complexity of the left and right channels. For example, if the left channel is silent while there is audio on the right channel, the majority of the total bitrate can be allocated to the right channel.

"Joint stereo" is where the audio can be/is encoded as L+R and L-R channels. That is, it looks at the redundancy between the left and right channels. As most stereo audio tends to have the left and right channels actually quite similar, joint stereo is relatively much more efficient at encoding MPEG audio, especially at lower bitrates. I believe in the "joint-stereo" scheme of encoding, it can fluctuate between JS and MS depending on the acutal audio (e.g., if there is a large difference between L and R channels, it may be better to encode as MS rather than JS for that section).

Although JS audio tends to yield better results (e.g., at 128 kbit/s MP3), it can also lead to some loss of stereo separation so it is probably not recommended if you are encoding to a high bitrate MP2/3 where fidelity will be good. Also, some encoders inappropriate switch between JS and MS when encoding and this can (apparently) lead to audible artifacts.

"Dual channel" is exactly what KoalaBear described. Exactly half the total bitrate is given to each channel (i.e., each channel is completely independent to the other). This method would probably be preferable if you actually have two mono tracks. In the context of a "stereo" soundtrack, this is the least efficient encoding method of the three.

Regards.