On Fri 04 Jan 2019 at 17:26:07 (+0000), Brian wrote: > On Wed 02 Jan 2019 at 22:56:22 -0500, kamaraju kusumanchi wrote: > > On Wed, Jan 2, 2019 at 9:23 PM David Wright <deb...@lionunicorn.co.uk> > > wrote: > > > On Wed 02 Jan 2019 at 14:44:14 (+0000), Brian wrote: > > > > > > > > I'm intrigued; I hadn't realised that conversion of the scanned image > > > > for some vendors' devices took place on the device itself. How do you > > > > know this happens? It is the frontend to SANE (xsane or scanimage, for > > > > example) which I've always associated with image aquisition conversion. > > > > > > It really is rather easy. You insert a USB stick into the scanner, > > > press scan, and later observe that a JPEG or PDF file has appeared > > > on the stick, as appropriate. > > > > Yes, that is precisely what I did. Stick a USB into the scanner and > > press the scan button. > > My HP Envy 4520 has no such button. There is an option for scanning to > the computer, but software is required on the computer to do that and > HPLIP does not provide it. > > Anyway, I managed to persuade the device to give me the PDF it would > have sent to a USB stick if the facility had existed (the device has > Apple's AirScan). If it matters, the PDF does not have any Creator or > Publisher information and doesn't contain any embedded or subset fonts.
It sounds as if this is sufficient to make you confident that the device is doing the conversion and not the computer: anything that decouples the two from privately passing information to one another outside the delivered file. A USB stick, or email, is just the most obvious. > Scanned at a resolution of 600: > > brian@desktop:~$ pdfimages -list out.pdf > page num type width height color comp bpc enc interp object ID x-ppi > y-ppi size ratio > -------------------------------------------------------------------------------------------- > 1 0 image 5100 6600 gray 1 8 jpeg no 1 0 600 > 600 2090K 6.4% > > ps2pdf reduces the 2090K by about 50% to 1051K. > > A different scanner device and source document, of course, and maybe > different methods of PDF production, so I wouldn't read too much into > this. Proving whether any compression applied is lossless is more difficult because pdfimages seems mute on what processes were carried out in extracting an image from the PDF. I have made the assumption that scanning compressed means that lossy compression is applied whereas scanning "uncompressed" means that lossless compression is applied. Cheers, David.
