I guess you mean at a lower layer, like everything below the transport layer, including packet framing, byte sequences, and radio modulation techniques. That's a good point.

And that would be true for virtually all "encrypted" human wireless communications today -- that we could see the structure at all of the layers below where the encryption is applied -- but isn't it still possible with some spread-spectrum techniques to make a meaningful radio signal look quite a bit like noise?