The simplest but most explanatory definition I have seen Plagiarized from another site.
DSMX is basically two layers of multi-access techiques.
DSM2 is based on something called CDMA (Code Division Multiple Access). In simple term, this means two things: 1) The signal is spread out over a wider frequency band and 2) each transmitter/receiver pair uses it's own coding scheme to "scramble" the signal. The coding schemes are designed in such a way that even if two stations are transmitting on the same frequency, the respective signals can still be isolated by the receivers, thanks to the coding schemes.
CDMA is great, because it's resilient to static, interference and other transmitters transmitting on the same frequency. However, there's a limit to how many stations can transmit on the same frequency before the signal gets too diluted to be successfully isolated by the receiver.
So DSM2 employs a simple scheme to try to avoid this. When starting up, the transmitter tries to find two free frequencies and starts transmitting. The receiver is scanning the frequencies for the unique signature of its paired transmitter. Once found, the receiver locks in and you have a working link. DSM2 uses two separate channels, so if one gets knocked out by interference, the other channel may still be usable. The problem is that the channel allocation happens on transmitter startup, so if both channels become unusable at some later point in the flight, you may still lose the link.
And that's why DSMX was designed. DSMX essentially uses the same encoding scheme as DSM2, but the frequency changes thousands of times per second according to a pseudo-random sequence negotiated between the transmitter and receiver. Even if some channels used become completely saturated and unusable, it would normally only result in cut-outs a couple of milliseconds, which would be too short to even notice (in theory).
...and that's pretty much how it works!
