Comment 29 for bug 997728

This looks much like that the data coming from acroread is already bad quality.

Sorry, I have overlooked the special case for acroread. acroread sends jobs in PostScript format AFAIK and due to DRM (Digital Rights Management) problems we do not convert this PostScript into PDF, regardless whether the printer is a PostScript printer or not. So only the pstops CUPS filter will be called. And this means that no graphical elements of the original input are turned into bitmaps (rasterized) and so if some parts of the printed output are bad quality, this bad quality must come already for the input.

So I fear we are hitting a bug in the Adobe Reader (acroread) here.