Re: Patch for sane.tex (image polarity)

From: abel deuring (a.deuring@satzbau-gmbh.de)
Date: Wed Oct 04 2000 - 03:10:38 PDT

  • Next message: Sebastien Sable: "Re: [SnapScan] Acer 300/310/610 and wine debugging"

    Henning Meier-Geinitz wrote:
    >
    > Hi,
    >
    > On Tue, Oct 03, 2000 at 02:48:56PM +0200, Oliver Rauch wrote:
    > > as far as I know there is no 1 bit color frame or image format defined.
    > > I can imagine how this would look like in 3 pass scanning mode,
    > > but there is (as far as I know) no interleaved RGB 1 bit format.
    >
    > You may be right concerning image formats (like pnm, tiff, ...). The SANE
    > standard explicitly allows this case (->byte interleaved).
    >
    > > May be we should use the 8 bit format for it.
    >
    > Definitely. If somebody really needs 1-bit color ...

    The scanners supported by the Sharp backend can produce 1 bit RGB data,
    and the backend supports the scan mode. But it expands the 1 bit data
    into 8 bits, because [x]scanimage simply copies the data received from
    the backend into the output, with a PNM header in front. And a RGB
    format with 1 bit per colour is apparently not supported by the "PNM
    programs", at least in the format delivered by the Sane programs.

    Therefore, I suggest, that the Sane standard should simply state that
    RGB data should have at least 8 bits per colour. A smaller bit depth is
    hardly of an use at a time, where you can buy a 20 GB hard disk for
    $200, so the only thing needed is a clarification of the standard.

    (BTW, what happens, if [x]scanimage receives RGB data with, say, 12 bit
    per colour? Is such a format handled properly by PNM programs or Gimp?)

    Abel

    --
    Source code, list archive, and docs: http://www.mostang.com/sane/
    To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
    



    This archive was generated by hypermail 2b29 : Wed Oct 04 2000 - 03:02:46 PDT