Read_e2ds should (optionally) retrieve uncertainty values from the pipeline reduced data (e.g. in case of ESPRESSO), or construct what the SNR of each pixel is directly from the data.
This is needed for proper propagation of uncertainties as well as making mock data.
When constructing it from the data, it would need to measure what the (peak - to avoid systematics, see below) SNR is in each order and what the flux is there; to calibrate the flux axis as counts. Then you also need to scale based on the flux variation with time, and then with an F(lambda,t) array, you can take the sqrt of the whole thing.
Read_e2ds should (optionally) retrieve uncertainty values from the pipeline reduced data (e.g. in case of ESPRESSO), or construct what the SNR of each pixel is directly from the data.
This is needed for proper propagation of uncertainties as well as making mock data.
When constructing it from the data, it would need to measure what the (peak - to avoid systematics, see below) SNR is in each order and what the flux is there; to calibrate the flux axis as counts. Then you also need to scale based on the flux variation with time, and then with an F(lambda,t) array, you can take the sqrt of the whole thing.