Visualize many views in BigDataViewer


this is a follow up from a Neubias discussion, so I’m tagging @Christian_Tischer here.

I have a dataset with many (thousands) of views (total size ca 0.5TB) which is painful to visualize in the normal BigDataViewer. There was mentioning of BDV extensions being developed that would allow to visualize such a dataset, by e.g. choosing the sources, or even do more things than just viewing (I know I’m a bit vague here…).

Could you point me to that? Thanks!


Probably this (although Tischi can probably elaborate on it):


Dear @noreenw,

Which file formats are you data in?

cc @NicoKiaru


Thanks @imagejan!

@Christian_Tischer: They come as tif from the miroscope, but so far I converted to bigdataviewer hdf5 at some point in the pipeline.


HI @noreenw,

You can download some bdv extensions on the update site

It’s a very early work (= expect things to break) done with @Christian_Tischer. It relies on BigDataViewer still so if you want to display a thousand sources you will still get into trouble. But you will be able to display more easily a subset of sources because they are sorted in a tree architecture. If your ‘xml/hdf5’ dataset has some extra attributes (Tile, Channels…) this will appear in the tree view, even more facilitating the selection.

Extra note :

  • a new UI (and inner state) is in development for bigdataviewer. Not yet there, but it should help when released.
  • maybe Paintera ( is an alternative ?
  • something else ?

And a small question : Why so many views ? Depending on the use case, I’ve heard about Composite ( which can combine many sources, but I don’t know how that works (ping @axtimwalde)


1 Like

@noreenw Here is the PR and branch if you want to give it a try:

In principle it’s ready and should be merged to master soon. We’re still making last tweaks and adapting downstream projects.


Thanks all for the replies and tips!!!

The dataset is a bit special since it comes from a custom microscope and I got the data as 40,000 small tiles. I’m grouping them already (fusing small regions) to reduce number of files by a factor 10 or more, but it’s still a lot of files and a bit atypical data.