2000.12.18 14:33 "Re: Stripes in thumbnail", by Joris Van Damme

scanlines in the original g4 tiff(s) cannot be correctly interpretted by LibTiff, and are simply (partially) skipped, leaving a (partially) blank line. Warning and error behaviour of the g3/g4 decompressors has been

Hm. Now as you mention it... there are warnings. But as I'm an absolute newbie, I didn't put them in relation to the stripes...

I checked the original image. Seems there's no problem at all. In fact, my wrapper doesn't even emit any warnings for the first image directory. And the image is read correctly, no (partially) blank lines (columns, in fact, I noticed the original image is rotated too) whatsoever. So if you do get any warnings, I would like to take a look at them.

Seems that my original guess was wrong. If I were you, I'ld search for the problem further down the road, perhaps in the code that makes the thumbnail image. Can't help with that, though.

There is a problem in the second image directory (not that this has anything to do with problem you reported, of course). I get an error about the ImageLength tag missing.

I can elaborate on one way to get you started on the second task.

wow... it's _that_ complicated?

It might not have to be. I elaborated on one way to get this job done, but there might of course be others.

But I did forget to mention, that those thumbnails don't have to be accurate. It's enough for me - and the task it's needed for - to find that a part of the picture is green and not red. So red=255-cyan, green=255-magenta... is quite good enough.

In that case, it should be a breeze. But why (pretend to) use cmyk if it's nothing but inverted RGB? Why not simply use RGB?

> I think I can provide you with a whole bunch of G4-TIFFs which we
> have here at the newspaper.

Thanks, but no thanks, got plenty of those.

Joris