We are used to seeing shots from TV news helicopters every day, they are part of the backdrop to life in the 21st century. But so often we hear them overlaid with studio commentary, so it’s interesting to hear that their raw audio contains telemetry. It caught the attention of [proto17], who took some audio pulled from a news helicopter video and subjected it to a thorough investigation to retrieve the data.
The write-up is at a very in-depth level, and while there’s an admission that some of the steps could have been performed more easily with ready-made tools, its point is to go through all steps at a low level. So the action largely takes place in GNU Radio, in which we see the process of identifying the signal and shifting it downwards in frequency before deducing its baud rate to retrieve its contents. The story’s not over though, because we then delve into some ASCII tricks to identify the packet frames, before finally retrieving the data itself. It still doesn’t tell you what the data contains, but it’s a fascinating process getting there nonetheless.
It’s easy to forget that GNU Radio has signal processing capabilities far beyond radio, but it was the subject of a fascinating Superconference talk. We even jumped on the bandwagon in the non-foolish part of our April Fool this year.
7 thoughts on “Pulling Data From News Feed Telemetry”
Reminds me of this older post:
Wow, that’s one of the best GNU Radio examples I saw so far. Step by step from acuiring the signal to drawing coordinates in Google map with every step clearly explained. I learned a lot from this, thanks very much!
Some older drones also encode downlink telemetry in the analog video. Usually at some non standard subcarrier frequency, using AFSK1200 with standard UART to generate the modulation
I can’t remember if multimon-ng allows specifying that subcarrier freq, but if it was standard AFSK1200 in this feed then you could have simply used that to decode, I think
Proto17 might be colorblind.
You got me! I knew someone was gonna call me out lol While writing the wiki I remember staring at the image and just guessing
It’s either myself or Proto17 that’s colorblind, as the “blue” baudline channel mentioned in the write-up is magenta for me!
Please be kind and respectful to help make the comments section excellent. (Comment Policy)