Bug: JPK: Some series' are not viewable with FS

[rleigh]

See attached image. (Virus.jpk)

Series 0 is viewable
Series 1 is broken (invalid image). May be a regression? This was fixed previously for 4.4, so maybe a related issue?

[bpindelski]

Still happens on gretzky.

[rleigh]

Series [0] is 8-bit and viewable
Series [1] is 16-bit and still showing as being an invalid image.

[jburel]

Roger: is this still happening. only gretzky?

[rleigh]

Yes. The second series is a 16-bit 3-timepoint timeseries. On gretzky, the planes in the timeseries are quite different to on howe. However, testing with bioformats develop, showinf shows the same as on gretzky/FS. showinf -fast -autoscale shows the same t=0 as howe. I'm uncertain what this is really supposed to look like, so I'm unsure if this is a reader bug or just the raw data being bad--copying Melissa.

[mlinkert]

Closing as fixed, since it looks like both series are now viewable (i.e. no "invalid image" errors).

For me, both series are the same in Insight as with showinf or Fiji. The thumbnail for series 2 looks black because the first timepoint is very dark. If you assign a color to the channel and magnify the thumbnails, then it becomes apparent that the thumbnail is correct.

Also, *.jpk files are just TIFFs, so you can open Virus.jpk in ImageMagick? (or some other program) and see that the images look more or less the same barring min/max scaling differences.