ISS (ZARYA): automatic audio recording and Deepgram transcription example (v0.8.7)

[sgoudelis]

This example schedules suitable passes of the ISS (ZARYA), records audio from its UHF FM voice repeater, and transcribes received speech with Deepgram. The screenshots use the active 437.800 MHz voice-repeater transmitter listed by SatNOGS at the time of writing.

The receiver, gain, antenna, and local RF conditions will differ between stations. Treat the USRP B210 and 15 dB gain below as an example, then tune those values for your own hardware.

Before configuring the observation

1. Use Ground Station v0.8.7 or newer and configure the station's correct location.
2. Connect an SDR that can receive around 437.800 MHz and an antenna suitable for the 70 cm satellite band. Confirm that the SDR appears in Ground Station before scheduling a pass.
3. Open Satellites → Orbital Data Sources and synchronize the satellite data. Confirm that 25544 - ISS (ZARYA) and Mode V/U FM - Voice Repeater CTCSS 67.0 Hz - 437.800 MHz are available.
4. Open account menu → Preferences → Integrations, configure a Deepgram API key, and save the preferences. Deepgram transcription requires internet access and sends the received audio to Deepgram's service.
5. Check the current ISS repeater status and transmitter details before scheduling. A pass can produce a silent recording and no transcript when the repeater is inactive or nobody is speaking.

1. Add the monitored satellite

Open Scheduler → Monitored Satellites → Add, then set:

• Enabled: on
• Satellite: 25544 - ISS (ZARYA)
• Minimum Peak Elevation: 15°
• Task Start Elevation: 10°
• Lookahead Window: 24 hours
• Rotator: select the rotator attached to your station, or None for a fixed antenna
• Unpark before observation start: off in the pictured setup
• Park rotator after observation end: off in the pictured setup

The 15° minimum excludes very low passes, while the tasks begin when the ISS reaches 10° elevation. The screenshot uses an S.A.T. Hamlib rotator; choose your own configured rotator or None.

2. Configure the SDR session

Select Add SDR and configure the session:

• SDR: select the receiver attached to your deployment
• Sample Rate: 1 MHz
• Gain: 15 dB in this USRP B210 example; adjust for your receiver and noise level
• Antenna Port: UHF (TX/RX) in this example
• Auto-calculate Center Frequency: on
• Center Frequency: 438050000 Hz in the pictured setup

With automatic center-frequency calculation enabled, Ground Station places the 437.800 MHz downlink away from the SDR's DC spike while keeping it inside the sampled bandwidth. Let the application calculate the center frequency for the selected transmitter instead of copying 438.050 MHz blindly to another receiver.

The screenshot shows a second available SDR session, but only one receiver is required for this example. Both tasks below use the selected B210 session.

3. Add audio recording and transcription

Choose Audio Recording, then set:

• Transmitter: Mode V/U FM - Voice Repeater CTCSS 67.0 Hz - 437.800 MHz
• Demodulator: FM (Frequency Modulation)

Ground Station records the demodulated audio as a 16-bit PCM, mono, 48 kHz WAV file.

Next, choose Transcription and set:

• Transmitter: Mode V/U FM - Voice Repeater CTCSS 67.0 Hz - 437.800 MHz
• Modulation: FM (Frequency Modulation)
• Provider: Deepgram
• Source Language: English
• Translate To: No Translation

Choose the source language that matches the speech you expect to receive. Translation is unnecessary for this example; using Deepgram translation also requires a configured Google Translate API key.

The two tasks run during the same observation. Audio Recording preserves the received WAV audio, while Transcription streams the demodulated audio to Deepgram and saves recognized text.

4. Save and verify

Save the monitored satellite. Ground Station should create observations for matching passes inside the 24-hour lookahead window. If needed, select the monitored satellite and regenerate its observations.

Before the pass:

• Confirm that an upcoming observation appears in the Scheduler.
• Make sure another session is not using the selected SDR.
• Confirm that the Deepgram API key is configured and the backend has internet access.
• Check the antenna, SDR access, current orbital data, and rotator connection if one is configured.

After the pass, open the observation and the File Browser. The WAV recording and saved transcription should appear with the observation's output.

If the WAV file contains only noise or silence, confirm that the 437.800 MHz repeater was active during the pass, the FM signal was visible, and the gain and antenna were suitable. If audio was recorded but no text appears, verify the Deepgram key, backend internet access, selected source language, and whether the received speech was clear enough to recognize. Radio noise, overlapping stations, Doppler error, and weak signals can reduce transcription accuracy.