Re: snapscan 1212u_2 problems

From: Lawrence Glaister (lg@jfm.bc.ca)
Date: Tue Dec 26 2000 - 17:39:23 PST

  • Next message: Eugene Kanter: "xsane-0.68 configure problem fixed?"

    the color image may be a little smaller by up to several lines. The reason
    for this is that the sensor is reading the color information from slightly
    offset positions on the page and the scanner software (snapscan driver)
    attempts to reregister the lines of data. In other words, the line of red
    scan data may need to be combined with the green data from 3 scan lines
    later in the data flow. Because of this, the scanner may scan 500 * 3 lines
    but only 497 triplets of aligned data may come out of the driver.... some
    gets discarded because it doesn't have all 3 color bands available. Even at
    high scan resolutions, it looks like there should be 6 or less lines
    discarded. I did put a print statement in my driver to see what the offsets
    were and they mostly showed up as 2 or 3 lines depending on resolution. I
    have never had scan data messed up like the picture you sent me... I suspect
    there have been some changes which broke the 1212u driver more than it
    already was. If I get a chance in the next few days, I will download the
    latest snapscan driver and sane 1.0.4 and see if I can get it running again.
    Its probably worth having our discussions on the sane maillist so that more
    people are aware of the problem.
    =======================================================
    Lawrence Glaister VE7IT email: lg@jfm.bc.ca
    1462 Madrona Drive http://jfm.bc.ca
    Nanoose Bay BC Canada
    V9P 9C9
    =======================================================
    ----- Original Message -----
    From: John Coppens <jcoppens@linux2.uccor.edu.ar>
    To: Lawrence Glaister <lg@jfm.bc.ca>
    Sent: Tuesday, December 26, 2000 2:53 PM
    Subject: Re: A few questions

    > On Mon, 25 Dec 2000 20:16:08 -0800 "Lawrence Glaister" <lg@jfm.bc.ca>
    wrote/El dddd> Hi John,
    > > ok, great, we are using the same scanner. I have the "upload required"
    > > version as well. I plugged it into my ups and only have to reload every
    few
    > > months when we get a really nasty power day. I posted some fixes to the
    > > 1.0.3 snapscan version that were hacks, but got me going. I had two
    > > problems, one, the scanner software would hang in an infinite loop
    during
    > > scsi speed calibration and another where the usb status returned EAGAIN.
    I'm
    > >...
    >
    > Hi Lawrence
    >
    > After some difficult experimenting (because of 36C temperatures) I did
    > find something strange.
    >
    > I scanned an image at 50bpi, same format, in gray, b/w, and color. I then
    > checked the size of the image received from the scanner, summing the WROTE
    > statements of the debug output. I get correct numbers for b/w, and
    grayscale,
    > but the color image is off:
    >
    > Summing WROTEs: 724599 bytes
    > The image should be: 723330 bytes (570 x 1269bytes/line,
    > as reported in the same debug output)
    >
    > That's one line difference (1269 bytes).
    >
    > Now, if I only could determine where this difference comes from... Any
    > suggestions?
    >
    > John
    > --
    >

    --
    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 : Tue Dec 26 2000 - 17:32:51 PST