[ome-users] bfconvert multi-series LSM file

[Melissa Linkert]

Hi Niko,

> sorry for the delay. Actually I checked again and in fact you're
> right - the sliders are there when opening the file in FiJi. I
> probably got this wrong because my initial reason for converting
> them was to use the OME-TIFF as input for the XuvTools stitcher -
> which currently doesn't read my LSM files properly, but those being
> converted via FiJi to OME-TIFF work fine.
> 
> So I opened the file resulting from the bfconvert with XuvTools and
> just saw one tile - whereas the OME-TIFF exported from FiJi
> correctly showed with the many different tiles. Yes, I'm aware this
> is an issue of XuvTools (unfortunately we don't have the time to fix
> it right now), but still this means the file generated by bfconvert
> differs somehow from the one exported by FiJi even though they look
> the same when you open them in FiJi. Don't get me wrong, that's not
> a complaint, just an observation that I thought is worth sharing :)

Yes, there is a difference.  When you convert the files through Fiji,
you are concatenating all of the tiles into a single stack.  This means
that what is stored in the OME-TIFF file is treated as a single tile -
effectively, you have lost a dimension.

When you convert using bfconvert, the individual tiles are preserved as
separate stacks/series, which means that XuvTools needs to be able to detect
that there are multiple series and deal with them appropriately.  I thought
that it could already do this, but maybe not.

Regards,
-Melissa

On Mon, Sep 17, 2012 at 03:07:19PM +0200, Niko Ehrenfeuchter wrote:
> Hi Melissa,
> 
> On 13.09.2012 16:27, Melissa Linkert wrote:
> >Hi Niko,
> >
> >>>If you just need all of the series from the input file in a single
> >>>OME-TIFF, then you actually don't need to specify any options at all -
> >>>that's the default behavior.  So just this should work:
> >>>
> >>>$ bfconvert input-file.lsm output-file.ome.tiff
> >>
> >>thanks for the suggestion. That's what I tried already, but it
> >>results in one big TIFF where all the channels and z-slices are in
> >>consecutive order. At least this is what FiJi shows when I'm opening
> >>the resulting file, but I rather need to be able to distinguish the
> >>channels, Z slices and stacks (like the 5D-variant displayed by FiJi
> >>showing the sliders for C, Z and t).
> >>
> >>Is there a way to do this? Or is bfconvert just behaving strangely
> >>on my data?
> >
> >Hmm.  So there is just a single slider when you open the converted file
> >in Fiji?  I would expect that behavior if the "View stack with" option
> >was set to "Standard ImageJ" when the file is opened, or if the file was
> >converted like this:
> >
> >$ bfconvert input-file.lsm output-file.tiff
> >
> >instead of like this:
> >
> >$ bfconvert input-file.lsm output-file.ome.tiff
> >
> >(Note that the .ome.tiff extension is very much required to get an
> >OME-TIFF.)
> 
> sorry for the delay. Actually I checked again and in fact you're
> right - the sliders are there when opening the file in FiJi. I
> probably got this wrong because my initial reason for converting
> them was to use the OME-TIFF as input for the XuvTools stitcher -
> which currently doesn't read my LSM files properly, but those being
> converted via FiJi to OME-TIFF work fine.
> 
> So I opened the file resulting from the bfconvert with XuvTools and
> just saw one tile - whereas the OME-TIFF exported from FiJi
> correctly showed with the many different tiles. Yes, I'm aware this
> is an issue of XuvTools (unfortunately we don't have the time to fix
> it right now), but still this means the file generated by bfconvert
> differs somehow from the one exported by FiJi even though they look
> the same when you open them in FiJi. Don't get me wrong, that's not
> a complaint, just an observation that I thought is worth sharing :)
> 
> Cheers & thanks
> Niko
> _______________________________________________
> ome-users mailing list
> ome-users at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-users