I got a idea for issue(lines between swaths)

2001-03-29
2001-04-26
  • nobuki kikuchi

    nobuki kikuchi - 2001-03-29

    I got a idea. Now are some approchs to the issue("lines between swaths") proceeding?
    I am Japanese.
    Sorry for my broken English.
    I am a user of Linux(debian) and using pnm2ppa on HP 720c with a thankful heart.

    TODO
    * Printing of e.g. tiger.ps (an example included with ghostscript)
      shows lines between swaths.  The Windows driver does not.  This is because
      the four passes of the color head (gMaxPass=4) are all at the same
      vertical position.   This is not the way to e.g. paint a wall using a
      paint roller - instead the successive passes should  be vertically
      staggered  so  their swath boundaries are not all at the same vertical
      position.   The Windows driver does this: the vertical position of pass 2
      is 1/4 of a swath height  below pass 1, etc.  Unfortunately, the design
      decisions in cutswath.c make this not a simple change.   Initial changes
      would be to somehow decouple the four sweeps, so they could have independent
      left and right values, directions, vertical positions, etc.

    Please let me ask about above.
    I am thinking perhaps this issue is a obstacle on actual printing color images.

    Sence pbm2ppa 0.8.6 was released, I have dreamed printing color images with 720c on Linux. Because Windows' driver of 720c did not have function of color adjustment. And unfortunately 720c's color had color shift to green. We realize that color shifts to green on human's face sensitively. We often have reported this subject to HP. But nothing has changed. Pnm2ppa has had color calibration function by respectfull the pnm2ppa team. But pnm2paa has had a issue that printing of e.g. tiger.ps (an example included with ghostscript) shows lines between swaths.
    Mr.duncan wrote "Unfortunately, the design decisions in cutswath.c make
                     ^^^^^^^^^^^^^
    this not a simple change." I think really it was unfortunate, too.
    Also Mr.duncan wrote "pnm2ppa works fine for my (non-graphics) needs" in http://sourceforge.net/forum/forum.php?thread_id=29456&forum_id=3729

    http://sourceforge.net/forum/forum.php?thread_id=51964&forum_id=3729
    But none of the original developers are still active;
    pnm2ppa now works well enough for many common printing
    tasks,

    Now are some approchs to this issue proceeding?
    I can not program. It is all I can do to prisent pnm2ppa in Japan.
    But,
    I have abseved behavior of 720c on windows98SE's driver. Of course I am thinking any anyone have known it. I hope a discussion over cutswath.c start. So I have decided to write the following.
    The best mode of windows' driver may print eight times. If power button is pushed at a printing, we can get a gradation with eight grades of color. Of course it need a timing to push power button. I suppose the driver of windows may divide a data by 8. The data divided by 8 may be printed eight times. And the eight times printings may be excuted by different groups of nozzles. The difference among nozzles may be wiped out by it. The black ink may be blowed two times.
    The normal mode of windows' driver may print six times. If power button is pushed on a printing, we get a gradation with six grades of color.
    The econofast mode of windows' driver may print one times. If power button is pushed on a printing, we get a gradation with one grade
    of color.
    The nozzles may be divided into eight or six or one groups in the vertical direction. In one time sweeping each of groups in the nozzles blow the data divided by 8 or 6 or 1. Then a paper is shited by a length corresponded to one group of the nozzules divided into eight or six or one groups. And a data of inks blowed from nozzles is shifted by one group of nozzles. The image of this way is showed by the following.

           xxxx
           xxxx  xxxx
           xxxx  xxxx  xxxx
           xxxx  xxxx  xxxx  xxxx
           xxxx  xxxx  xxxx  xxxx  xxxx
    ABCD   abcd  abcd  abcd  abcd  abcd  abcd
    EFGH         efgh  efgh  efgh  efgh  efgh  efgh
    IJKL               ijkl  ijkl  ijkl  ijkl  ijkl  ijkl
    MNOP                     mnop  mnop  mnop  mnop  mnop  mnop
    QRST                           qrst  qrst  qrst  qrst  qrst  qrst
    UVWX                                 uvwx  uvwx  uvwx  uvwx  uvwx  uvwx <1 nozzles
                                               xxxx  xxxx  xxxx  xxxx  xxxx <2 nozzles
                                                     xxxx  xxxx  xxxx  xxxx <3 nozzles
                                                           xxxx  xxxx  xxxx <4 nozzles
                                                                 xxxx  xxxx <5 nozzles
                                                                       xxxx <6 nozzles
    a = A/6
    b = B/6
    .
    .
    .
    and so forth.
    Above is the case of normal mode.

    The point is that each line or each spot is blowed by all nozzles. To tell accutately, it is blowd by not all nozzles. But each line or each spot is blowed by more different nozzles. Each nozzles must have some differences over behavior to blow inks. This way will distribute this differences to all spots evenly. This way will turn out "lines between swaths". I am thinking so. As above I can not program. But I think a reconstruction by this way may be simple.
    The origial data is divided by 6 or 8. And a paper is shited by a length corresponded to one of the nozzules divided into six or eight or one groups. And a data of inks blowed from nozzles is shifted by one group of nozzles.
    In other words a change is only changing quantity to shift a data of inks blowed from nozzles and a paper.
    They may be all changes this way need.

    How do users think about this way?

    Perhaps someone besides the original developers must excute. If there are some explanations of sorce codes it will motivate new developers. Of course I will try to study programing, too. But I start from zero. Now printer's drivers for Linux are going to be provided from some makers. So buying new printers will solve the problem. But we would do well to treasure old macines.
    Above may be nonsense. If it is so, I am sorry.

    Thank you.
    over.

     
    • nobuki kikuchi

      nobuki kikuchi - 2001-03-29

      The image of this way has crumbled.
      Again I wrote it.
             xxxx
             xxxx  xxxx
             xxxx  xxxx  xxxx
             xxxx  xxxx  xxxx  xxxx
             xxxx  xxxx  xxxx  xxxx  xxxx
      ABCD   abcd  abcd  abcd  abcd  abcd  abcd
      EFGH         efgh  efgh  efgh  efgh  efgh  efgh
      IJKL               ijkl  ijkl  ijkl  ijkl  ijkl  ijkl
      MNOP                     mnop  mnop  mnop  mnop  mnop  mnop
      QRST                           qrst  qrst  qrst  qrst  qrst  qrst
      UVWX                                 uvwx  uvwx  uvwx  uvwx  uvwx  uvwx <1
                                                 xxxx  xxxx  xxxx  xxxx  xxxx <2
                                                       xxxx  xxxx  xxxx  xxxx <3
                                                             xxxx  xxxx  xxxx <4
                                                                   xxxx  xxxx <5
                                                                         xxxx <6
      a = A/6
      b = B/6
      .
      .
      .
      and so forth.
      Above is the case of normal mode.

       
    • nobuki kikuchi

      nobuki kikuchi - 2001-03-29

      The image of this way has crumbled.
      Again I wrote it.

             xxxx
             xxxx  xxxx
             xxxx  xxxx  xxxx
             xxxx  xxxx  xxxx  xxxx
             xxxx  xxxx  xxxx  xxxx  xxxx
      ABCD   abcd  abcd  abcd  abcd  abcd  abcd
      EFGH         efgh  efgh  efgh  efgh  efgh  efgh
      IJKL               ijkl  ijkl  ijkl  ijkl  ijkl  ijkl
      MNOP                     mnop  mnop  mnop  mnop  mnop  mnop
      QRST                           qrst  qrst  qrst  qrst  qrst  qrst
      UVWX                                 uvwx  uvwx  uvwx  uvwx  uvwx  uvwx <1
                                                 xxxx  xxxx  xxxx  xxxx  xxxx <2
                                                       xxxx  xxxx  xxxx  xxxx <3
                                                             xxxx  xxxx  xxxx <4
                                                                   xxxx  xxxx <5
                                                                         xxxx <6
      a = A/6
      b = B/6
      .
      .
      .
      and so forth.
      Above is the case of normal mode.

       
    • nobuki kikuchi

      nobuki kikuchi - 2001-03-29

      Very sorry.
      Again I try.

           xxx
           xxx xxx
           xxx xxx xxx
           xxx xxx xxx xxx
           xxx xxx xxx xxx xxx
      ABC  abc abc abc abc abc abc
      DEF      def def def def def def
      GHI          ghi ghi ghi ghi ghi ghi
      JKL              jkl jkl jkl jkl jkl jkl
      MNO                  mno mno mno mno mno mno
      PQR                      pqr pqr pqr pqr pqr pqr <1
                                   xxx xxx xxx xxx xxx <2
                                       xxx xxx xxx xxx <3
                                           xxx xxx xxx <4
                                               xxx xxx <5
                                                   xxx <6
      a = A/6
      b = B/6
      .
      .
      .
      and so forth.
      Above is the case of normal mode.

       
    • nobuki kikuchi

      nobuki kikuchi - 2001-03-29

      Very very sorry.
      Space letter may be cut.
      In this time I used "_".
      "_" mean space.

      _____xxx
      _____xxx xxx
      _____xxx xxx xxx
      _____xxx xxx xxx xxx
      _____xxx xxx xxx xxx xxx
      ABC__abc abc abc abc abc abc
      DEF______def def def def def def
      GHI__________ghi ghi ghi ghi ghi ghi
      JKL______________jkl jkl jkl jkl jkl jkl
      MNO__________________mno mno mno mno mno mno
      PQR______________________pqr pqr pqr pqr pqr pqr <1
      _____________________________xxx xxx xxx xxx xxx <2
      _________________________________xxx xxx xxx xxx <3
      _____________________________________xxx xxx xxx <4
      _________________________________________xxx xxx <5
      _____________________________________________xxx <6
      a = A/6
      b = B/6
      .
      .
      .
      and so forth.
      Above is the case of normal mode.

       
    • nobuki kikuchi

      nobuki kikuchi - 2001-03-29

      In this time it has succeeded.
      If widths of fonts is changed, try to copy them on fixed-width fonts.
      And the following is wrong.
      Mr.duncan wrote "Unfortunately, ...
      ^^^^^^^^^
      The following is correct.
      Mr.duncan wrote "Unfortunately, ...
      _________________^^^^^^^^^

       
    • Duncan

      Duncan - 2001-04-26

      Dear nobuki kikuchi,
      you obviously thought a lot about this issue!

      The problem is , as you say, to stagger the printing so
      different nozzles print on the same line.

      Unfortunately, I probably dont have the
      free time to fix pnm2ppa any more at present.

      Thanks for your comments.  If I  manage to work on it,
      they could be useful.

      duncan

       

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks