ome
bfconvert misreports stage position when converting from metamorph tiff ND
Consider the attached ND dataset: test.zip
If we delete the nd file (to force bioformats to consider each file individually) and use showinf to display the X, Y stage position (stageX[0], stageY[0]) for each frame individually, we see that the images have been taken at two different positions ((16204.6, 444.7) and (16204.6, -1429.1)).
Yet, after converting the file with bfconvert to an ome.tiff, the PositionY of both positions (position_a and position_b) is reported as 444.7, i.e. the second Y position has been lost. This can also be seen by checking the XMLAnnotations (a bit further down in the ome metadata), which only shows X position for position #1 and Y position for position #1, but not for position 2.
Unless I am mistaken, the coordinates of the second position are thus completely lost by bfconvert.
On a similar theme, the XMLAnnotations only seem to have kept metadata for Channel #0 (491_Fast illumination), not for Channel #1 (561_Fast). (As a side point, perhaps the numbering of positions should be zero-based to be consistent with the channels.)
Thanks @anntzer for raising the issue and providing a sample fileset. My apologies for the slow response time as I have been away on holidays the past number of weeks and am just catching up with this issue now. I was able to reproduce the bug as you described using the latest Bio-Formats 7.0.0 and the provided dataset (copied to scratch/inbox/gh-4069), this does indeed look like a bug in the ND2Reader.
