Your first SSTV decode

Watch the recording become a picture

The VIS header identifies Robot 36, then the image builds line by line. The frame is 320 × 240 pixels. The complete WAV lasts 38.315 seconds, including its header and surrounding silence.

The green trace is the live audio FFT; orange is peak hold. The waterfall shows the most recent 10 seconds. The display narrows to a 6 kHz audio span, using a 1,024-point FFT and a Hann window. Levels are dBFS, not calibrated RF power. These are measurements of the file input, not an antenna signal.

At the end, the receiver returns to Idle and Images reads 1. The completed picture, peak hold and last waterfall remain visible.

Before decoding an audio file

A filename ending in .wav identifies a container, not necessarily the signal inside it. Begin with the unchanged practice file so format conversion, a missing header and radio setup are not mixed into the same experiment.

Use the original PCM file as a baseline instead of re-recording the browser's speaker output or making a lossy conversion. The recorded checksums and conversion results identify the exact source used in our controlled file check.

Recover your first picture

1. Select SSTV receiving. Use the SSTV signal category, not SSDV, DRM or Weather Fax. Keep automatic mode detection enabled so the file's VIS header can select Robot 36.
2. Import the downloaded WAV. On iPhone or iPad, tap Open audio; the Open Audio File picker lets you select robot36.wav. File decoding starts after selection. On Mac, use Open Replay…, choose the WAV, set playback to 1× with Loop off, then start the receiver.
3. Let the whole frame finish. Look for Robot 36 and the image progress. The signal lasts about 38 seconds at normal playback speed. Do not restart it just because the first lines are black or grey.
4. Compare your image with the card above. You should see the circle, colour bars and Robot 36 label, with sstv.org.uk visible along the bottom. Slightly softened edges are normal for this analog mode.

The audio plays, but no SSTV picture appears

1. Separate playback from import. Hearing this page's audio does not establish that the app's selected source is receiving samples. Import the WAV through the file path described above.
2. Check the signal category and the beginning. Select SSTV and include the VIS header. SSDV packets, DRM waveforms and radio I/Q are different inputs.
3. Compare the original and edited file. Keep normal playback speed and compare recorded duration and pitch. Use the actual sample-rate comparison before treating 44.1 kHz itself as the cause.
4. Separate file success from live reception. If the original works, investigate the radio/audio route with a retained recording. If the original also fails, report the app version, device, selected input, filename and detected mode.

Use the iPhone and iPad import walkthrough or the Mac input-routing guide for platform-specific steps. A local file result is not proof of microphone, receiver or RF performance.

A stereo recording sounds loud, but the decoder receives silence

Check the left and right channels separately before adding gain. Some file-input paths average stereo into mono. If the two channels contain the same signal with opposite polarity, their average cancels to silence—even though either channel is audible on its own. This is one specific routing failure, not a reason to reject every stereo WAV.

We checked this with generated Robot 36 audio and the documented average-to-mono loader in Open-SSTV 0.6.18. The rate, duration and original picture stayed fixed; only the channel polarity or channel selection changed.

1. Keep the original. Inspect or listen to each channel separately in an audio editor. A channel can contain the useful signal, silence or a different signal.
2. Make a separate mono copy of the useful channel. These FFmpeg examples select a channel explicitly; they do not average both together:

ffmpeg -i recording.wav -af 'pan=mono|c0=c0' -c:a pcm_s16le left.wav
ffmpeg -i recording.wav -af 'pan=mono|c0=c1' -c:a pcm_s16le right.wav

Use the copy containing the intact SSTV header and image. Do not mix left and right again if they cancel. Selecting one channel can also discard useful content, so choose it from the recording rather than assuming that left is always correct. The FFmpeg pan documentation describes channel mapping.

Audio only: these steps are for ordinary stereo audio, not raw SDR I/Q. I/Q channels have a different role and need an appropriate demodulation path first.

Download the actual results and checksums, reproduction script and setup and scope notes. The script starts from our unchanged known mono practice WAV and creates all three cases; it does not use a radio.

Does SSTV audio need to be loud?

Keep file replay and live input separate. A quiet but intact recording can work; turning up an already distorted recording is not a general repair. Before changing gain, try the unchanged practice WAV, check the selected input and inspect the original recording for a complete beginning and end.

We varied only the amplitude of the same 38.315-second mono PCM file, keeping its rate, header and picture fixed. In this Open-SSTV 0.6.18 file test, all three inputs produced a 320 × 240 Robot 36 image:

The last column averages absolute channel differences against the image decoded from the unchanged file; it is not a perceptual quality score. Clipping percentages include the entire file, including its quiet sections. The recovered pictures were similar, but neither altered file produced pixel-identical output.

This does not establish a safe clipping level or a weak-signal limit. We added no receiver noise, interference, speaker or microphone. The quiet result therefore says little about a noisy live signal, and a decoded picture does not prove that its source waveform was undamaged. These are not new SSTV Signal Desk or Lite app measurements.

1. If the practice file fails too, check import, signal category and source selection before adding gain.
2. If only your recording fails, keep the original and compare its header, duration, channels and waveform. A loud recording can still have the routing or timing problems described above.
3. For live reception, use the selected input's meter and avoid sustained overload. If an earlier stage already clipped the signal, lowering the later playback volume cannot restore the samples that were lost.

Download the actual results and checksums, reproduction script and setup notes. Inspect the actual decoded quiet-file picture and clipped-file picture. Start the reproduction from the original practice WAV linked above.

After your first decode

App Store

Get SSTV for iPhone, iPad and Mac

Start with free practice in SSTV Lite, or choose the full SSTV Signal Desk (Pro) edition. Both editions are available for iPhone, iPad and Mac.

Lite includes a successful-reception allowance and optional in-app purchases. Pro is a separate one-time purchase. See your local price on the App Store, or compare the editions.