Issue description
When loading a dataset that contains NaNs, the NaN values are rendered as black, rather than transparent, making they display of such datasets on top of other layers less useful.
How to recreate
Load a SEVIRI image (e.g. VIS008 channel)
Load an AVHRR granule for the same time period (e.g. channel 4)
Change driving layer to AVHRR.
As can be seen in the screen shot below, with the AVHRR granule loaded on top of the SEVIRI full disk image, all pixels outside the actual AVHRR granule are shown as black, hence hiding the SEVIRI image below. Only if the transparency of the AVHRR layer is increased, we start to see the SEVIRI image (second screen shot).
Expected behavior
NaNs shall be rendered as transparent, such that no increase in transparency of the upper layer is needed in order to see data outside the AVHRR granule.
Other remarks
This is not only a problem for LEO data, but also for GEO data not covering the full disk. E.g. if you would like to display a L2 cloud product (with NaNs for all cloud-free pixels) on top of a L15/L1C image.
Ported from EUM Gitlab issue 152
Issue description
When loading a dataset that contains NaNs, the NaN values are rendered as black, rather than transparent, making they display of such datasets on top of other layers less useful.
How to recreate
As can be seen in the screen shot below, with the AVHRR granule loaded on top of the SEVIRI full disk image, all pixels outside the actual AVHRR granule are shown as black, hence hiding the SEVIRI image below. Only if the transparency of the AVHRR layer is increased, we start to see the SEVIRI image (second screen shot).
Expected behavior
NaNs shall be rendered as transparent, such that no increase in transparency of the upper layer is needed in order to see data outside the AVHRR granule.
Other remarks
This is not only a problem for LEO data, but also for GEO data not covering the full disk. E.g. if you would like to display a L2 cloud product (with NaNs for all cloud-free pixels) on top of a L15/L1C image.
Copying the comments from EUM issue 152:
