I have a PDF file created by someone else, so I don't know precisely how it was created. When I view the PDF in Acrobat Reader, it looks OK. When I open it and view it in Inkscape, it looks OK. See http://www.sylvesterbradley.org/images/ ... kscape.png
If I modify it and save it to PDF from Inkscape, there are three images in the document that should have transparent backgrounds, but which appear in the saved PDF as black backgrounds. See
http://www.sylvesterbradley.org/images/ ... reader.png
How do I overcome this? I.e. how do I get it to save the PDF with transparent backgrounds to the images which were originally (presumably) bitmaps?

Thank you - Rowan

Hi.

The original pdf probably had the raster image with alpha values substituted with a raster image masked with another one -which doesn't get saved well from inkscape.

Since those transparencies are not fading edges, I'd try vectorising the raster images used for the masking and use them for clipping the raster objects.
The top part wouldn't need any raster parts to be clipped then, just use the vectorised mask.

I have fixed it simply by creating a new vector mask for each of the three objects, which has done the job. But I was hoping for something more automatic, in case something like this happens again with more complex shapes.

If the situation is as you have described, how do I select "the raster images used for masking", and then vectorise them?

Thanks - Rowan

I only know of the manual way of cleaning things up.

Select objects, ungroup until the status bar says that's a masked object, then release masking and trace bitmap the black&white raster image and delete it after. Not a quick solution, might worth trying the deep ungrouping feature in 0.92.