While discussing napari/napari#6695 in the PR party we noted that users might not know the implications of that setting, and in general the implications of using different data types for the labels layers. Specifically:
the bit depth determines the memory usage, and can result in an 8x difference in RAM usage
in napari in particular, uint8 and uint16 have much better performance when slicing/refreshing because of implementation details
but bit depth also limits the maximum label number
Once this page is done, ideally, we would add a link to it in the description of the new labels dtype preference.
📚 New content request
While discussing napari/napari#6695 in the PR party we noted that users might not know the implications of that setting, and in general the implications of using different data types for the labels layers. Specifically:
Once this page is done, ideally, we would add a link to it in the description of the new labels dtype preference.