How to display the frequency/deviation when a signal is decoded

[obones]

Hello,

I have quite a few devices emitting around here and they are all decoded just fine by rtl_433 and my RTLSDR dongle.
As I'm trying to analyze the signals with other devices, I'd like to have a few more information about any given decode signal, in particular the center frequency (OOK/FSK) and the deviation (FSK).
Something like this output for instance

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
time      : 2023-03-10 11:22:46
signal    : 433.85MHz
model     : TFA-Pool     Id        : 166
Channel   : 2            Battery   : 1             Temperature: 11.1 C       Integrity : CHECKSUM
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
time      : 2023-03-10 11:24:20
signal    : 868.12MHz +- 10kHz
model     : LaCrosse-TX35DTHIT                     id        : 61
Battery   : 1            NewBattery: 0             Temperature: 11.2 C       Humidity  : 60 %          Integrity : CRC

I looked a the command line options, but I could not find anything obvious in this area.

Is this something that is already available in memory and just a matter of adding to the output via new code?

[merbanan]

[-M time[:] | protocol | level | noise[:] | stats | bits | help] Add various meta data to each output.

[merbanan]

I think the digits have been up for discussion before and the IIRC the tuner does not offer that good resolution. But maybe @zuckschwerdt knows better.

[zuckschwerdt]

Indeed, 100 kHz estimates are the best we get here, based on the instant phase angle used. It's not an FFT ;)

[obones]

Well, here I see that the second digit is actually of importance here, especially because of the rounding that is done with only one digit.
But I'll leave it as is and just keep a modified copy lying around whenever I need it.

[zuckschwerdt]

Expectation is the problem. People will see 433.85 and (rightfully) say the signal is at 433.85 MHz -- which it is most likely not.

[zuckschwerdt]

I already played with FFTW to good success in https://github.com/merbanan/rtl_433/tree/feat-beepdetect
It's planned to add a "true FFT" mode to get e.g. these stats at the expense of some cpu load.