Itinerary: Difference between revisions
No edit summary |
No edit summary |
||
| Line 2: | Line 2: | ||
Nelson: Bob's method of interpreting frequencies | Nelson: Bob's method of interpreting frequencies | ||
We think it'll be best to just sample more than twice per period (once per up section, and once per down section) and then accumulate the abs of first difference. I think that'll be really effective and just incredibly inexpensive on compute -- probably best we could possible get on that front. | |||
Also gives a confidence score per bit of the thing, So if the high frequency should accumulate arround 20, and the low frequency arround 10, and we get an 11 then things are good, a 14 and things are bad | |||
Revision as of 18:09, 5 August 2026
Nelson: Bob's method of interpreting frequencies
We think it'll be best to just sample more than twice per period (once per up section, and once per down section) and then accumulate the abs of first difference. I think that'll be really effective and just incredibly inexpensive on compute -- probably best we could possible get on that front.
Also gives a confidence score per bit of the thing, So if the high frequency should accumulate arround 20, and the low frequency arround 10, and we get an 11 then things are good, a 14 and things are bad