Fast-scan television (FSTV, or analog ATV) is real-time moving video, sent the way broadcast TV was: a new picture many times a second. It looks like television, and you watch it on a television, but the signal is enormously wider than any voice mode.
Why it is so wide
A picture contains a lot of detail. SSTV sends it slowly through a narrow channel; fast-scan sends a fresh frame every 1/30 of a second, so the same amount of detail has to go through thousands of times faster, and the channel has to be proportionally wider. Speed is paid for in bandwidth.
Rounded, illustrative values. Picture detail differs between the two, but the trade is the same: speed costs bandwidth.
A 6 MHz signal also lets in 6 MHz of noise, roughly 34 dB more than a 2.4 kHz SSB signal, so a fast-scan picture needs a much stronger signal at the receiver than speech does. See Noise and signal-to-noise ratio.
How the picture is made
The video is scanned in lines. In North American NTSC, 525 lines make a frame, sent as two interlaced fields, odd lines first and then even, to reduce flicker.
Drawn with 15 lines instead of 525. Same pattern.
To save bandwidth, the video is sent as vestigial sideband AM: one full sideband and a stub of the other. The low video frequencies are kept, and the whole thing is narrower than ordinary double-sideband AM.
Video signal frequency runs outward from the carrier. VSB trims the lower sideband's far side.
Sound travels on a separate subcarrier in the same channel.
Where and how
A 6 MHz signal won't fit in the HF or low VHF bands, which are narrow and shared with many modes. Fast-scan therefore lives on 70 cm and higher bands, where there is room. On 70 cm the channels overlap some cable-TV channels, so a cable-ready television can receive it directly. On higher bands a converter is needed first.
A camera's video output goes into the transmitter and the audio into its sound input.