Can you provide some problematic scans, together with your source file, so that we can have a look at what went wrong?

Sorry it took so long to get back to you. Your request came to me after I deleted everything to try again. It turns out that the problem is going from pdf to jpg. Somehow, it loses enough resolution (even though I specify a resolution of 300) so that when AMC converts from jpg to ??? (ppg? ppd? some format I didn't recognize) that it has problems with the overall resolution somehow. I tried it again two different ways (in case anyone else is interested in this):

1. I converted the pdf to postscript (using pdf2ps) and then back to pdf (using ps2pdf). This did not work. When I added the resulting pdf to AMC, the marks done on the tablet or with acrobat on a pc were lost.
2. I converted the pdf to png using convert, and then added the png files directly to the AMC project. My unrecognized scans went down from about 136 (I think) to 2. When I looked at the two scans, I saw they were defective. On one of them, the student had written close enough to one of the alignment circles that I figure it couldn't recognize it properly. On the second, the testing center stamped right over one of the alignment circles. The only problem I see with this, is that the final student exams in pdf have a black border around the page. I don't know if the png introduced this or not - I'll investigate. It is a problem because the annoted marks go partially into the black border.

Thus, it seems that the proper process is to use imagemagick's convert to convert pdf's to png's and use those directly. If you are interested in why the jpgs did not work correctly, I can try to repeat my process, but I may not be able to post the offending jpgs as they are part of actual student exams.

Thanks for listening - I'll post some specific questions about my use case in another post.

You can send me some problematic PDF, PNG, JPG scan files by e-mail if you want me to have a look at them without posting them publicly.