On Sun, 28 Dec 2008 18:03:06 -0800 Glenn Linderman <[email protected]> wrote:
| On approximately 12/28/2008 5:33 PM, came the following characters from | the keyboard of Anthony Thyssen: | > On Fri, 26 Dec 2008 23:11:32 -0800 | > [email protected] wrote: | > | > | So I was doing a book layout, putting small pages on a big page. | > | | > | The idea is that page fragments, formatted to fit small page sizes | > | already, or easily cropped from something containing it, would be placed | > | in various spots on the big page. | > | | > | Some pages are to be composed of multiple pages. Note c004.tif and | > | c005.tif below... they are both contributing to the same small page, but | > | do not both contain black stuff at the same places. | > | | > | All is bitonal black and white, so it seems like -compose multiply would | > | let one merge these various pages, and it seems to work in some smaller | > | cases I had done, where I build all the pages separately and merge them, | > | but then that took lots of memory, so I thought I'd -flatten as I go.... | > | but then c004.tif doesn't get included... apparently it gets overlaid by | > | c005.tif, even the white parts. | > | | > | Does -compose only persist for one -flatten? I thought as a setting, it | > | would persist until it was changed. | > | | > The problem is that -compose is for some file formats a image meta-data | > setting, as so reading an image overrides the existing setting. | > | > I myself have had problems with this, and I think these problem will persist | > until Image magick as a formal (internal) notion of a 'global' override | > verses the image meta-data value. | > | > My own thinking is that global values should override any and all meta-data | > values unless left unset or the user purposefully unsets the global override | > using +value. | > | > Similarly if the user wants to specifically set the image meta data value the | > -set operator should be used. | > | > However IM has no true notion of a global override, verses image meta-data. | > Compose in particular is only seen by the alpha composition operator simply as | > image meta-data. | > | > Until the library is changed to a data-structure that consists of | > an 'image' with pointers to 'image-meta-data' and 'global-overrides', | > with library routines to get the appropriate settings. I doubt the confusion | > will have any resolution anytime soon. | | | Thanks. I certainly hadn't reached that level of understanding of IM | operators and settings, and had never heard of meta-data setting. | | So the safest thing for a na__ve user is to specify input file, then | settings, then operators, and not to assume that settings persist to the | next operator. That sort of bloats the command line, but if it works | consistently... that is better than being surprised, I guess. | It is a pain yes. and ideally if compose setting has been set then it should remain set regardless of if images have been read in, or parenthesis, image re-ordering or anything else is done. But currently that is not happening. You could put in a BUG report in the IM Bugs Discussion Forum, so that if -compose has been set it overrides any compose meta-data present in the input image. This is currently what is happening for -dispose and -delay for animations, and would make it more user friendly. Only a few file formats actually define a compose setting as part of the images meta-data (settings associated with the image itself), TIFF just happens to be one of them. Anthony Thyssen ( System Programmer ) <[email protected]> ----------------------------------------------------------------------------- ``Probable impossibilites are to be preferred to improbable possibilites'' -- Aristotle ----------------------------------------------------------------------------- Anthony's Home is his Castle http://www.cit.gu.edu.au/~anthony/ _______________________________________________ Magick-users mailing list [email protected] http://studio.imagemagick.org/mailman/listinfo/magick-users
