Lima: Make sure Lima exposure time is meaningful
The Lima PR limagroup/Lima-tango-python!95 (merged) rework a bit the accumulation in order to expose filters and operations.
At BM18 this new accumulation mode will be used to reduce the size of the data from 32 to 16bits, using MEAN
to avoid overflows.
In such case (MEAN
), the real expo time of a lima frame is more or less the expo time of a sub frame (with more statistics), which is not the same as expo time saved as BLISS metadata.
If someone is about to play with the accumulation parameters, Flint flat-field correction will be wrong (same for daiquiri).
Do we have enough metadata to correct this intensity client side?
acc_operation
acc_expo_time
Is the read-only derivatived attributes also part of the saved parameters?
acc_expo_time
acc_nb_frames
Still we can believe that
# more or less
acc_expo_time = expo_time / (expo_time // acc_max_expo_time) if acc_max_expo_time > expo_time else expo_time
We also could imagine to create an extra dedicated metadata in order to normalize the stuff. For example a "data scale correction".