On 4.03.08, Paul Johnson wrote: > For many plots in a sub directory "PlotFigures", I have 2 versions, > one in eps version and one in pdf version.
> With Lyx-1.5.3 on Fedora Linux 8, I'm using a koma-script book format. > I've run up against this problem inserting graphics. When I insert a > pdf figure, then in the DVI/postscript output, the output is not good > because there's a big blank area at the top of the figure. In the pdf > output, the document is good (presumably because the figure was pdf, > the latex processor handles it well). On the other hand, if I insert > the eps figure in Lyx, then the DVI/postscript output is good, but the > pdf output looks bad. ... > But for working with the existing images, what is a good approach? If I remember right, PDF has 2 setting for the size of a graphic: one intended for papersize and one corresponding to the bounding box. Some programs tend to confuse these two or some users tend to set the wrong options to the converters -- in any case it sometimes happen, that the produced EPS will have a bounding box corresponding to the papersize. In my Tools>Preferences, I have a "PDF (ps2pdf)" file format and the converters PDF (ps2pdf) -> EPS pdftops -f 1 -l 1 -eps $$i $$o EPS -> PDF (ps2pdf) epstopdf --outfile=$$o $$i Mark the -eps option to ps2pdf. > In some email messages from this list, I see users ask this question > and they are advised to omit the image type from the LyX insert. LyX > will use eps for DVI output and pdf for pdf output. ... > However, when I try that in LyX, I see "figure not found" on the > screen. This advise is outdated. As of 1.5.3, LyX can do the conversion between EPS and PDF automatically, so you can use whatever output format works better. The cache feature means, that also performance problems are no longer an issue (the LyX Help is outdated in this respect as well). > In retrospect, I see I need to do more work in the R script that > manufactures the images, because I think I can stop the pdf output > from including that giant white space at the top of the pdf. But it > seems weird to have to do that in order to get an EPS output file. Maybe that your script does not set the bounding box in the pdf, but maybe it is just the pdf-eps conversion that does not use it (i.e. it is in fact a pdf-ps conversion with an inappropriately named output file). What are the problems when working with eps files only? Guenter
