Problems opening Olympus oir 'line scan' files

Hi again. Having problems opening Olympus oir files that are acquired in ‘line scan’ mode.

Please see my previous post about opening Olympus oir frame scans with >=1000 images. Luckily and thanks to Bio-Formats people, this issue is being resolved with this push-request (PR).

Now it seems Bio-Formats is not opening Olympus oir files acquired in ‘line scan’ mode. In general, I am acquiring about 10,000 lines into each oir file. Regardless of the number of lines scanned, opening the resulting oir into Fiji using Bio-Formats is giving me only 512 lines?

Here is an example oir file with 10,000 lines, each line is 140 pixels:

https://www.dropbox.com/s/rb6bws9789kqxlp/20190416_b_0021.oir?dl=0

Here is the .tif output of the Olympus provided conversion software (NOT Fiji):

https://www.dropbox.com/s/pv8k58ihjqwei7v/20190416_b_0021_olympus.tif?dl=0

And here is the .tif when I drag and drop the same oir file onto Fiji

https://www.dropbox.com/s/5ktvjx07y4osii5/20190416_b_0021_fiji.tif?dl=0

If someone could try and reproduce the bug/behavior and track down how to fix it, that would be wonderful.

Thanks for all your work and support,

Robert Cudmore

Hi @cudmore,

So far I have been able to reproduce the issue as described with the latest Bio-Formats and carry out some initial debugging. Internally within the file Bio-Formats currently takes the height and width from the below properties:

    <commonimage:width>140</commonimage:width>
    <commonimage:height>512</commonimage:height>

I have not yet been able to locate the appropriate metadata which would show that it has 10,000 lines. I will have to continue debugging the issue further. I have created a Trello card to track the issue: https://trello.com/c/Kv6HwQY0/370-oir-linescan-capped-at-512-lines

Thanks for looking into this. Hope you can find the number of line scans in the header. Is it possible to read the number of lines directly from the raw image data once you determine the file contains a line scan as opposed to a frame scan?

As an update, Ive been able to spend a bit more time with this issue and it looks as though I have been able to locate the correct metadata within the file ok, so it should be possible to put a fix in place for this.

Thanks so much @dgault and that is great news. A fix would be wonderful. This and my other open PR are really limiting what I can do with the Olympus scope.

Hi @dgault. Any update or progress on tracking this down?

Hi @cudmore, Im afraid I don’t have a fix in place yet. There are 2 sets of image properties within the file and Im struggling to distinguish between them. Even then I had tried swapping to read the set with the height at 10,000. In that case the image looked sensible initially but upon inspection the first 512 rows were correct but after that the rest of the image appeared to have been shifted vertically by a significant number of pixels. I will try and dedicate more time to this on Monday morning and I will send you a private message to share the 2 sets of XML to see if you can help spot any useful differences.

Hi @dgault, any progress on fixing this? Do you have anyone at Olympus you can contact so they can improve any documentation on the format of the oir file?

Hi @cudmore, I’m afraid we have no further update on this issue. If you have a contact at Olympus that you are able to get any further information regarding the spec for line scan files or you could put us in touch with, then that would be a great help.