Preparing pictures for SSTV — what survives the mode

Know your canvas

The mode fixes the resolution, and it is smaller than almost anything else you work with:

Every one of those is 4:3. Crop to 4:3 yourself and you control what gets kept; leave it and the software decides, which usually means bars or a stretch.

Note the airtime column while you are here. PD290 is nearly five minutes of continuous transmission for one picture — a real commitment of both your transmitter and the frequency.

What survives

• Large, bold shapes. A subject that reads as a silhouette will always arrive recognisable.
• Saturated colour. Strong primaries survive; subtle tonal gradations band and blotch.
• Faces. Unreasonably robust — we are very good at recognising them from very little information.
• High contrast. Contrast is what noise eats first, so start with more than you think you need.

What does not

• Fine texture. Foliage, gravel, brickwork and hair become noise — 320 pixels cannot resolve them, and what fills the gap is speckle.
• Small or thin text. Anything below roughly a twelfth of the frame height is a gamble.
• Dark images. Noise is most visible in shadow. A low-key photograph arrives dramatically worse than a bright one over the identical path.
• Fine gradients. Skies especially — expect visible banding.

Text like a road sign

Most SSTV pictures carry a callsign, and most of them are harder to read than they need to be. The rules are the same as signage:

• Big. Callsign text should occupy a tenth of the frame height at least.
• Bold. Heavy weights survive; thin and light strokes disappear into noise.
• Contrasting. White on a dark solid band, or black on a light one. Never text directly over a busy photograph.
• At the top. Convention, and practical: if the receiving station catches the transmission partway through, the top is the part they missed — but if they lose the signal partway, the top is the part they kept.

Martin M1 leaves about 16 pixels of banner room at the top, which is exactly what the convention grew up around. Open Signal can stamp de YOURCALL onto outgoing pictures automatically — see the callsign stamp guide.

A working recipe

1. Crop to 4:3, choosing the part of the image that works as a bold shape.
2. Lift the shadows. Raise the black point so nothing important sits in the darkest tenth.
3. Add contrast and a little saturation — more than looks right on screen. It will look right after transmission.
4. Add the banner: solid band, bold callsign, high contrast.
5. Check it small. Scale the image down to 320 pixels wide and look at it. If you cannot read the callsign now, nobody will read it through noise.

That last step is the one that matters and the one everyone skips. Your picture will never look better than it does at 320 pixels in a quiet room.

When the mode is the wrong tool

Sometimes the picture genuinely needs the detail — a QSL card with an address on it, a schematic, a photograph you want kept. Squeezing it into 320×256 is the wrong answer.

That is what byte-exact file transfer and hybrid mode exist for: the real file, at full quality, rather than a re-scanned approximation. The trade-offs are set out in analogue versus digital pictures.