I wanted an answer to a boringly practical question: what if the network is gone, but the speakers and microphones still work?
SignalHop is an acoustic modem and mesh prototype that moves tiny messages over ultrasound instead of Wi‑Fi or Bluetooth.
What it actually does
The current build uses:
- FSK tones at 18 kHz and 20 kHz
- 48 kHz sample rate
- a 4 up-chirp sync preamble
- frames with payloads up to 255 bytes
- a WAV/audio runtime for encode/decode and round-trip testing
That makes it useful for emergency text, sensor payloads, and low-bandwidth status updates. Not glamorous. Very useful.
The measured part
From the current proof pack:
| Metric | Value |
|---|---|
| Encode latency | 16.634 ms mean |
| Decode latency | 713.028 ms mean |
| Round-trip | true |
| Payload tested | 77 bytes |
| Max payload | 255 bytes |
The decode path is slower because it scans the full signal with correlation. That is fine for v1. The important part is that the numbers are measured, not imagined.
The honest limitation
The core modem path is field-deployable today, but the full multi-node mesh still needs hardware validation across real devices. I’m keeping that line explicit on purpose.
That means this is a real protocol stack, not marketing fog.
Proof pack
Links
- GitHub: https://github.com/AmSach/SignalHop
- Bench JSON: https://raw.githubusercontent.com/AmSach/SignalHop/master/media/bench.json
- Summary JSON: https://raw.githubusercontent.com/AmSach/SignalHop/master/media/summary.json
Sound is the oldest protocol. We just updated the spec.



Top comments (0)