What is SSTV, and how does it actually work?

The core idea

Television as it was broadcast needs several megahertz of bandwidth, because it sends twenty-five complete frames every second. An amateur voice channel is about three kilohertz wide — roughly a thousand times too narrow.

SSTV resolves this by giving up on motion entirely. Send one frame, take as long as you need, and the bandwidth problem disappears. A picture that takes two minutes instead of a fortieth of a second fits comfortably into the space of a human voice, and anything that carries a voice worldwide will carry the picture too.

Brightness becomes pitch

The encoding is almost disarmingly simple. Each pixel’s brightness is represented by an audio frequency:

The transmitter sweeps along one row of the image, varying the tone as the brightness changes, then drops to 1200 Hz to mark the end of the line. The receiver hears the sync pulse, starts a new row, and paints what follows.

That warbling sound is literally the picture: rapid variation means fine detail, long steady tones mean flat areas of colour.

Getting colour

There is only one tone at a time, so colour has to be sent in sequence. The transmitter sends the red content of a line, then the green, then the blue, and the receiver reassembles them.

Different families made different choices here, which is most of why there are so many modes. Martin and Scottie send separate RGB scans. The PD modes send luminance and chrominance instead, sharing colour information between two lines to buy resolution — which is why PD290 manages 800×616 where Martin M1 manages 320×256.

The VIS header

With dozens of incompatible modes in use, the receiving station has to know which one is arriving. That is the job of the VIS code — Vertical Interval Signalling — a short digital burst at the very start of the transmission.

It is genuinely digital, in the middle of an otherwise analogue mode: a leader tone, a break, then seven data bits and a parity bit sent as two distinct frequencies. Those eight bits identify the mode.

This is why modern software never asks you what is coming. Open Signal reads the VIS header, selects the right decoder from six families, and starts painting — the mode appears on screen because the transmission announced it.

No error correction, on purpose

Classic SSTV has no forward error correction, no checksums and no retransmission. A burst of noise on the path becomes a burst of speckle in the picture, at the exact position it occurred.

This turns out to be the most important property it has.

• A weak signal produces a noisy but recognisable picture. You can see who it is from, read the callsign in the banner, and file a report.
• A digital mode on the same path produces either a perfect picture or nothing whatsoever. There is no partial credit below the threshold.
• Because there is nothing to negotiate, any SSTV receiver anywhere can decode any SSTV transmitter, and could in 1975.

Which of those behaviours you want depends entirely on the path and the day — the trade is laid out properly in analogue versus digital pictures.

Why it survived

Sixty-odd years is a long time for a mode to remain in daily use. SSTV survived because it is simple enough to implement from a published specification, degrades gracefully instead of failing, and needs nothing but a radio and a sound card.

What has changed is the software around it. The signal on the air in 2026 is the same signal; the difference is that decoding it no longer requires a dedicated monitor, a long-persistence phosphor, or an afternoon of patience.