Using mask upstream

[tomczak]

If I have a string of transformations and decide to build a new mask for one of them somewhere in the middle, the mask is available for the transformation downstream of it, but not for those upstream.

If for some reason I would like to use the same mask for an upstream image, the two methods I figured were a) to save the mask as a file then open this file on the side, which will then be recognised by all transformations as long as the pixel size is the same b) save settings from the mask dialogue box and recall them for in the mask upstream (but I'm not entirely sure if saving the settings save all types of masks with them).

Is there some other way of cloning the mask from a downstream image to the one upstream?

[jsachs]

How about this... (I just tried it and it seems to work)

In the downstream mask, save the mask settings in a temporary Mask settings file.
In the upstream mask, load the saved mask settings.

It a little more efficient if you then reset the downstream mask and rebuild it by selecting the upstream mask. That way the mask only gets computed once instead of twice.

[tomczak]

Thanks, that seems to be the easiest way to clone a mask upstream, plus when the downstream mask is reset and replaced by the upstream one, they will be synchronised - the upstream mask changes will percolate downstream.

On the same topic, I was trying to use Mask transformation below the downstream mask then move or copy it upstream, but I create circular reference this way as the mask in Mask transformation is still coupled to the original one, and the only way I can think of to decouple them it is saving and retrieving mask settings as above. BTW, the mask settings seem to record all the mask details made by various tools, at least those that I tried - so it's all good.

[tomczak]

There is another fairly easy way to do that, it seems:

Use Copy transformation, make its input the top image of the stream you're working on. Then copy the downstream transformation with the mask below the Copy -- the mask will be available in all transformations of the original stream.

[jsachs]

That sounds like it should work.